AQ_F_0800044_20190130_CMPL_InspRpt (6)

NORTH CAROLINA DIVISION OF Washington Regional Office
AIR QUALITY Avoca LLC
NC Facility ID 0800044
Inspection Report County/FIPS:Bertie/015
Date: 04/26/2019
Facility Data Permit Data
Avoca LLC Permit 01819/T51
841 Avoca Farm Road Issued 9/4/2018
Merry Hill,NC 27957 Expires 12/31/2021
Lat: 36d 0.0030m Long: 76d 42.6550m Class/Status Title V
SIC: 2087/Flavoring Extracts And Syrups,nec Permit Status Active
NAICS: 31193/Flavoring Syrup and Concentrate Manufacturing Current Permit Application(s)TV-Minor
Contact Data Program Applicability
Facility Contact Authorized Contact Technical Contact SIP/Title V/PSD
MACT Part 63: Subpart Boiler-112j, Subpart
Brian Conner David Peele Brian Conner DDDDD, Subpart FFFF, Subpart ZZZZ
Environmental Health& President Environmental Health& NSPS: Subpart Dc
Safety Manager [phone] Safety Manager
[phone] [phone]
Compliance Data
Comments: The facility appeared to be in compliance with the permit conditions
and applicable regulations at the time of the inspection. Inspection Date 1/25/2019, 1/29/2019, 1/30/2019
Inspector's Name Betsy Huddleston
Inspector's Signature: - .- Operating Status Operating
Compliance Code Compliance-inspection
Action Code FCE
Date of Signature: ��1?G�Jq;y On-Site Inspection Result Compliance
Total Actual emissions in TONS/YEAR:
TSP S02 NOX VOC CO PM10 *HAP
2017 1.19 3.20 34.63 1268.61 25.45 0.5040 596444.19
2016 1.32 2.67 30.68 1221.52 21.50 0.5050 539989.20
2015 1.22 3.10 33.91 939.80 24.70 0.5350 414346.31
* Highest HAP Emitted(in pounds)
Five Year Violation History:
Date Letter TVUe Rule Violated Violation Resolution Date
04/20/2018 NOV 2D .0530 Prevention of Significant Deterioration 04/20/2018
04/20/2018 NOV Part 63 -NESHAP/MACT Subpart ZZZZ Stationary 04/20/2018
Reciprocating Internal Combustion Engines
02/22/2017 NOV Part 63 -NESHAP/MACT Subpart DDDDD Industrial, 02/13/2017
Commercial,and Institutional Boilers and Process
Heaters
Performed Stack Tests since last FCE:None
Directions to Facility
From WaRO take Highway 17 North and then turn right onto Highway 171 North(Old Ford). Continue to
Jamesville,and turn right onto Highway 64 East. Go through Plymouth and make a left onto Highway 45
North at Pines Elementary. Turn right onto Avoca Farm Road (SR 1502). There are signs for Avoca and
Scotch Hall (golf community across from Avoca). Follow Avoca Farm Road to the plant(on the left at dead
(SharePoint\Facilities\Bertie08\00044\agl6s\2018-2019Inspection\20190130a16.doc) Page 1

end). Check in at the guard gate, and park in front of the brick office building on the right by the water. The
guard requires visitors to read Avoca's "cGMP, Safety and Confidentiality Practices"policy(a copy is
attached to the file copy of this report). A visitor's pass is issued in the guardhouse.
Safety Concerns
• Be aware of sources of trips, slips and falls such as uneven ground,stairs,ductwork.
• Be aware of traffic in the yard and in/out of the facility. Truck traffic is very heavy during sage
season(late May/June).
• No cell phones are allowed outside of the Avoca main office building.The facility's extraction
operations involve hexane,heptane, ethyl acetate, chloroform and ethanol solvents.
• Safety glasses,hard hats and steel toe shoes are required. Because of the solvent extraction
operations, spark resistant jackets/jumpsuits will soon be required by the facility. This clothing
will be supplied by the facility.
Permit History
• T37 was issued on 8/17/2010. Two permit applications for permit renewal and 112j were
addressed in T37. CAM permit conditions were also added.
• T38 was issued on 6/03/2011 for the installation and operation of two new biomass/bio-based
solids-fired boilers (ID Nos. ES-BB 1 and ES-13132), one No. 2 fuel oil-fired rotary dryer(ID No.
ES-RD) and their associated control devices.
• T39 was issued on 1/04/2013 to correct specific condition(2.1 E.7.) for Subpart DDDDD,which
is applicable to the two biomass boilers(ID Nos. ES-BB 1 and 13132). The condition contained an
incorrect compliance date. On December 21,2012 the EPA signed a revised boiler MACT
Subpart DDDDD rule that included additional and alternative standards depending on when the
boiler was constructed.
• T40 was issued on 6/06/2013 for replacing a dryer with a larger one, and adding a new primary
reactor. The existing reactor(R-3001)became a secondary reactor. Lime injection in the
biomass boiler bagfilter was also removed from the permit. Toxics limitations were removed
from the permit by Avoca's request as allowed under Session Law 2012-91 House Bill 952.
• T41 was issued on 11/26/2013 for the following changes:
- Updated CO and NOx emission factors for the biomass boilers based on performance testing.
- Replaced the burner on the rotary dryer with a 30 MMBtu/hr burner. Propane was added as fuel.
- Updated the maximum burner rating of the two(2)biomass boilers to 24 MMBtu/hr each.
- Created a PSD avoidance permit limit of less than 40 tpy VOC from the rotary dryer.
- A limit of less than 10 tpy n-hexane from the rotary dryer was created to avoid being subject to
the 112(g)requirements listed in 2D .1112.
- Included a previous 502(B)(10) change per 15A NCAC 02Q .0523 for a molecular sieve(ID No.
ES-MSDU-1024) in the Botanical Extraction Operations.
• T42 was issued on 1/27/2014 (application 13D)for the following changes:
- Replace two underground storage tanks (ID No. ES-1001-2-1-P2)with two above ground storage
tanks(20,000 gallons capacity each,ID Nos. ES-M-125A and 125B);
- Install a new storage tank with Plant Nutrient Extraction(PNE) (ID No. ES-TK-PNE-1);
- Add a sage briquette making machine (ID No. I-Briquette)with enclosed conveyors; and
- Include existing diesel emergency generator(ID No. E104).
(SharePoint\Facilities\Bertie08\00044\agl6s\2018-2019 Inspection\20190130a16.doc) Page 2

• T43 was issued on 12/19/2014 (application 14B)to update the capacity of ES-M-125A and 125B to
19,500 gallons,and add a condenser No. CD-3002 to dryer No. D-3001 at SFG.
• T44 was issued on 3/10/2015 (applications 14D and 14E) for the following permit changes:
— Addition of a new dryer(D-1002) after the current centrifuge(C-1203)at PNE.A chilled water
process condenser and distillate tank were also installed with the centrifuge.
— PNE and EVG fume scrubbers CD-Z-9215 and CD-Z-9216 are no longer required to operate.
A new product in PNE uses heptane and ethyl acetate extraction,and the process does not need
the use of the scrubbers. MON requirements don't apply to the product that will be made using
the new dryer because it won't employ any HAPs.
— Removal of ethanol tank TK-9214 from the permit.
— New operating temperature limits for the cryogenic condensers in Botanical and Biomass areas.
The original BACT determinations considered only use of hexane in these condensers. However,
Avoca is permitted to operate these plants using other solvents. When the condensers operate at
-40'F,water-based solvents like ethanol will completely freeze in the condensers. A 17°F
cryogenic condenser temperature limit is set for water soluble solvents.
• PSD Permit T45 was issued on 1/12/2016 (application 15A)for expansion of SFG. VOC was the
only criteria pollutant affected by the expansion. This permit was intended to be the first of a three-
step expansion of production at the facility. Avoca and DAQ RCO Permits had pre-application
meetings concerning their plan to submit three separate PSD applications. The following sources
were added to the permit:
- One 12,500 gallon storage tank(ID No. T-3006)
- One 12,500 gallon process tank(ID No. T-3007)
- Two reactors(ID Nos. R-3003 and R-3004)with process condensers(ID Nos. EX-3004 and ES-
3005)
- One centrifuge(ID No. C-3002)
- One dryer with(ID No. D-3002)with process condenser(ID No. ES-3006) and a chilled water
control condenser(ID No. CD-3002)
- Optional controls—chilled water control condenser(ID No. CD-3003)and mineral oil scrubber
(ID No. CD-3004-S).
• PSD Permit T46 was issued on 7/05/2016(application 16A)for expansion of SDE operations. This
was intended to be the second step of production expansion. VOC was the only criteria pollutant
affected by the SDE expansion. The following sources were added to the permit:
- Mineral scrubber system consisting of a chilled water control condenser(ID No. CD-4002)in
series with a mineral oil scrubber(ID No. CD-4003-S) installed to control:
o One 17,900 gallon virgin solvent tank(ID No. T-4001)
o Two 6,000 gallon process tanks (ID Nos. T-4017 and T-4018)
o Three 4,200 gallon reactors (ID Nos. R-4004, R-4005, and R-4044) with process condensers
(EX-4001, EX-4002, and EX-4003)
o One 1,500 gallon reactor(ID No. R-4015)
o One centrifuge(ID No. C-4001)
o One dryer with(ID No. D-4001)with process condenser(EX-4004) and a chilled water
control condenser(ID No. CD-4001)
- Process Equipment Leaks (ID No. ES-4000-F)
- SDE-2 process wastewater stream (ID No. ES-4000-WW).
- The"Biological Equipment for Purification of Sclareolide"and"Sclareolide(SDE) Operations"
permit item lists were edited and combined. See the previous inspection report for details on the
additions/deletions.
• Renewal permit T47 was issued on 1/12/2017. Two other permit applications were rolled into the
renewal.Application 12B is a Part 2 for the two biomass boilers. Application 14A is a Part 2 for
dryer ID No.D-3001 and SFG reactor ID No. R-3002 equipped with process condenser ID No.EX-
3003. Boiler MACT(Subpart 5D) conditions were also added to the permit.
(SharePoint\Facilities\Bertie08\00044\ag16s\2018-2019Inspection\20190130al6.doe) Page 3

• Permit T48 was issued on 7/27/2017(application 17C) for an ownership change to Ashland,LLC.
There was no change to the facility name.
• Permit T49 (application 17A) states the effective date is 10/16/2017. However,the permit cover
letter and DAQ's IBEAM database cite an issuance date of 8/17/2017. The October date appears to
be a typo,as there is no indication in the permit review that any delay in effective date was
necessary. The permit contains the following changes:
- New equipment will be installed for a new chickpea oil extraction process. It will be a secondary
operating scenario (SOS)for the Biomass extraction area of the facility. Existing equipment will
also be used in the chickpea extraction. The chickpea solid material will be packaged and sent to
another facility for further processing and the oil will be discarded. The new equipment is
identified and discussed within the Regulatory Review section of this report.
- A Biomass bulk bag packaging area(ID No. I-Biomass-PKG)equipped with a bagfilter(418
square feet of filter area, ID No. ADC-1001)was added to the insignificant attachment to the
permit.
- Emergency generator E-103 was removed from the permit.
- An expanded Subpart ZZZZ permit condition was added.
- Botanical Extraction was added to the MON condition 2.2 C.I.g.
• T50 (application 18A)was issued on 4/17/2018 to revert ownership of Avoca back to Pharmachem
Laboratories LLC, and change the facility name to Avoca LLC. Pharmachem was purchased by
Ashland,but they legally determined Pharmachem is still the official parent company for the facility.
• T51 (application 18B)was issued 9/04/2018 for the addition of a new flaker operation to process raw
materials for the biomass extraction operations.The flaker operation will consist of the following
equipment:
6,000 bushel grain storage silo with loading/unloading conveyance(ID No.I-SH-2001),
— Flake conditioner(ID No. CV 5-2004),
— Flaking machine(ID No.FLK-2001),
Flake de-conditioner(ID No.ES-DEC-2001)and associated simple cyclone (ID No. CD-MHZ-
2001),
— Flake transport system(ID No.I-CVZ-2002).
• On 10/31/2018 Avoca notified DAQ of a trial run(about two weeks) of a new product at Botanical
Extraction. The biomass debagging bagfilter ID CD-1004-1-FFI was temporarily moved to Botanical
and used as part of the trial. No HAPs or toxics were reported part of the process,only particulate.
Controlled emissions were estimated at 5 lbs PM for the entire trial.
• T52 was issued 4/22/2019 and combines two applications (17B and 18C) Application 17B includes
alternative Subpart FFFF(MON MALT)parametric monitoring for the chilled water condenser(ID
No. CD-4002) and mineral oil scrubber(ID No. CD-4003-5) installed on the SDE-2 emission
sources. The CPMS limits were established during Method 18 stack testing for hexane emitted from
the scrubber on 6/06/2017. In 18C Avoca requested to replace an existing No. 2 fuel oil-fired boiler
H-103 with a new propane-fired boiler(33.475 million Btu per hour heat input,ID No. H-104).
• Applicability Determination No. 3405 was issued on 4/09/2019 granting approval to conduct a trial
run of a potential new source that would be subject to the MON(either as a Group 2 or Group 1
source). That trial would happen in early May. Toluene, acetic acid,isohexane and n-hexane
emissions would be associated with the operation.
• Application 19A was submitted on 4/15/2019 to remove the sage dryer and associated cyclone from
the permit.
(SharePoint\Facilities\Bertie08\00044\agl6s\2018-2019 Inspection\20190130al6.doc) Page 4

Facility Summary
Avoca LLC is a chemical extraction facility that serves an international market. The facility was first
constructed in 1963. The company's primary saleable product is sclareolide. It is produced from sage
plants and is used in perfumes and detergents(Bounce, Tide, Febreeze) to help fragrance last longer.
Sclareol(the initially extracted oil)comprises less than 1% by weight of the sage. They send it off-site to
another Avoca plant in Wisconsin for conversion to sclareolide. The sclareolide is returned to Merry Hill
and purified. Sclareolide is a waxy material and is a sister compound to ambergris.
Sage is planted in August and September(they contract farmers in at least eight surrounding counties).
Approximately 30,000 acres were planted for 2018 harvest. The sage lies dormant during winter. It is
harvested usually within a 45-day window starting in May. The weather needs to be very warm so that the
sage produces an optimum amount of oil. Avoca stores sage in silage tubes and operates the sclareol
extraction plant 10-11 months a year. Avoca plans to harvest about 10,000 acres this year due to Ashland
wanting to reduce their storage capacity. After extraction most of the spent sage is land applied.
The facility also produces on-demand small batch specialty items. Some of their other extraction
products at the facility include cocoa, chick pea, and synthetic vanilla flavorant. They have also
conducted cranberry juice,vitamin, nicotine and krill oil extraction.
All of the facility's production areas operate under a nitrogen blanket(all process vessels and tanks).
Inspection Observations and Regulatory Review
On 1/25/2019, 1/29/2019 and 1/3 0/20 19 1 was on-site at Avoca to perform a physical inspection of the
permitted sources/control devices and a records review dating back to February 2018. Mr. Brian Conner,
EHS Manager, and Colin White,EHS Engineer, assisted during physical inspection on 1/25/2019. Mr.
Conner was ill the following week, so Colin White assisted through the remainder of the inspection.
Permit T51 was used to conduct the permit compliance inspection. Additional records were requested in
follow-up to the inspection. Those records were provided to DAQ by email on 2/01/2019.
BOILERS 111-101,H-102,AND H-104
H-101 was operating with steam pressure at 140 psi during the inspection. H-102 was operating with
steam pressure at 140 psi. H-103 no longer exists. H-104 is constructed and was operating at 140 psi
steam pressure. There were no visible emissions observed from the boilers.
Missrgp Sop'rco Control 11gv1ee fontrPllleytca "
l ltl) Sfan SuucgDeacptptigli
ID d1 `io. De'scr� #iuu
No.
2 fuel oil-fired boilers(20.3 MMBtu/hr
max heat input rate for H-101 and H-102, and
H-101 25.2 MMBtu/hr max heat input for H-103)
H-102
H-103 and NA NA
H-104
Propane-fired boilers(33.4 MMBtu/hr max
heat input rate, ID H-104)
Regulated PollutanE Lim,,O/Standards 1pplicsble'Regulatipn"
Particulate matter 0.37 pounds per million Btu heat input 15A NCAC 2D .0503
Sulfur dioxide 2.3 pounds per million Btu heat input 15A NCAC 2D .0516
(SharePoint\Facilities\Bertie08\00044\agl6s\2018-2019 Inspection\20190130a16.doc) Page 5

Regulated Pollutant Limits/Standards Applicable Regulation
20 percent opacity: H-103 only 15A NCAC 2D .0521(d)
Visible emissions
40 percent opacity: H-101 and H-102 15A NCAC 213 .0521(c)
Hazardous Air Pollutants Case-By-Case Maximum Achievable 15A NCAC 2D .1109
Control Technology [1120) Case-by-Case MACT
HCL: 1.1E-03 pounds per million Btu heat
input 15A NCAC 02D .I I I I
Hazardous Air Pollutants Mercury: 2.0E-06 pounds per million Btu [40 CFR 63 Subpart DDDDD)
heat input
(starting May 20,2019)
2D.0503 "Particulates from Fuel Burning Indirect Heat Exchangers"
Permit Condition 2.I.A.1.
This rule applies to all of the boilers. The particulate from H-101 and H-102 each shall not exceed 0.37
lb/MMBtu heat input. I calculate the emission limit for H-104 to be 0.33 lb/MMBtu based on cumulative
99.2 lb/MMBtu heat input of the four boilers. The AP-42 emission factor for particulate emitted from
No. 2 fuel oil combustion is 0.014 lb/MMBtu(at t40,000 Btu/gal). Propane particulate emissions are
considered negligible(NC LPG combustion spreadsheet cites 0.00053 lb/MMBtu). There are no
monitoring/recordkeeping/reporting conditions for the boilers under this rule in the permit. Compliance
with 2D.0503 is assumed.
Note that the boiler NESHAP requires initial performance testing on H-101 and H-102.
2D.0516 "Sulfur Dioxide Emissions from Combustion Sources"
Permit Condition 2.I.A.2.
The sulfur dioxide emitted by the boilers shall not exceed 2.3 lbs/MMBtu.The AP-42 factor for No.2
fuel oil sulfur dioxide emission is 0.5 lb/MMBtu(sulfur content is less than 0.5%and heat content is
assumed approximately 140,000 Btu/gal). There are no monitoring/recordkeeping/reporting conditions
for the boilers under the rule in the permit. Compliance with 2D.0516 is assumed.
2D.0521 "Control of Visible Emissions"
Permit Condition 2.1.A.3.
Boilers H-101 and H-102 are limited to an opacity of 40%. Boiler H-103 is limited to 20% opacity.
The limits may not be exceeded more than once in any hour and not more than four times in any 24-hour
period.There are no permit monitoring/recordkeeping/reporting requirements for the boilers under this
rule. Visible emissions from the boilers were observed at 0% during the inspection. Compliance with
2D.0521 is indicated.
2D.1109 "112(.0 Case-by-Case Maximum Achievable Control Technology"
Clean Air Act Section 1120) - Case-by-Case MACT for Boilers&Process Heaters
Permit Condition 2.1.A.4.
The Permit requires Avoca to use best combustion practices when operating the boilers.
Monitoring/Recordkeepiing(no Reporting)
Avoca must perform an annual boiler inspection and maintenance on boilers as recommended by the
manufacturer, or as a minimum,the inspection and maintenance requirement shall include the following:
(SharePoint\Facilities\Bertie08\00044\agl6s\2018-2019 Inspection\20190130al6.doc) Page 6

• Inspect the burner, and clean or replace any components of the burner as necessary;
• Inspect the flame pattern and make any adjustments to the burner necessary to optimize the flame
pattern; and
• Inspect the system controlling the air-to-fuel ratio, and ensure that it is correctly calibrated and
functioning properly.
The Department of Labor inspects the boilers annually. Their certification lists any repairs or
maintenance that Avoca needs to conduct on the boilers. The certificates are stored in the boiler control
room, and copies are kept in the environmental files.
Avoca has a contractor that performs tune-ups that meet the above requirements. Semi-annual tune-ups
were performed on all three fuel oil boilers. The most recent on was completed 10/19/2018. CO, 02 and
CO2 were measured multiple times during the tune-ups.
The boilers appear to be in good condition. Avoca uses Mechanic's Mate system for creating work orders
for maintenance and repair,but will be shifting to using Maximo in the next year.
Avoca appears to be in compliance with 2D.1109. Note that this rule will give way on 5/20/2019 to the
NESHAP Subpart DDDDD requirements.
2D.1111 "Maximum Achievable Control Technology" 40CFR,Part 63,Subpart DDDDD
Permit Condition 2.I.A.S.
Avoca opted to delay Boiler MACT applicability by obtaining a permit under the Clean Air Act Section
1120)Case-by-Case MACT. The requirements of this permit condition do not become effective until
5/20/2019. Avoca has begun working with Trinity Consultants to meet the requirements of the
regulation.
There are no monitoring or recordkeeping requirements in 5D for the new propane-fired boiler H-104
other than annual tune-up. The first tune-up will be due 13 months following startup,which was
1/15/2019. Therefore, it's due 2/15/2020. An initial notification was also due for H-104 within 15 days
of startup. Avoca did not send an official notification until 2/25/2019. This is a reportable deviation that
WaRO has decided to not enforce. The initial notification was satisfied with the permit application for
the boiler. Trinity and I both agree that no NOCS is required for this boiler. However,they will owe
annual reports associated with the tune-ups. 63.7550(b)(1-4)sets up the first report to be due 1/31/2021
(February 2020 due date for tune-up).
Avoca must meet the following emissions limits for H-101 and H-102(H-103 is gone):
The emissions must not exceed the following emission limits,
except during startup and shutdown
HCI 1.1E-03 pounds per million Btu of heat input
Mercua 2.0E-06 pounds per million Btu heat input
CO 130 ppm by volume on a dry basis corrected to 3 percent oxygen
Filterable PM 7.9E-03 lb per million BTU heat input;or
(or TSM) (6.2E-05 lb per million BTU heat input)
Initial Compliance Demonstration
• Complete an initial tune-up on or before 5/20/2019.
• Complete a one-time energy assessment on or before 5/20/2019.
• Perform performance testing or fuel sampling for each of the above pollutants.
Avoca plans to perform stack testing of all the pollutants. A protocol(2019-092ST) was submitted on
312112019. They plan to test H-1Oland H-102 using Methods 1-5, 10, 26A and 30B. Testing is
tentatively scheduled for late May/early June.
• Establish operating limits according to Table 7 of Subpart 5D and conduct CMS performance
evaluations.
(SharePoint\Facilities\Bertie08\00044\ag16s\2018-2019 Inspection\20190130a16.doc) Page 7

Avoca, Trinity and L interpret the regulation to read as long as the boilers burn ultra-low sulfur fuel
and the initial performance tests demonstrate compliance with the limits, Avoca won't be required to
conduct further stack testing/fuel sampling or monitoring other than fuel type,fuel usage, and suer
certifications.
• Submit a Notification of Compliance Status(NOCS)before the close of business on the 60"day
following full completion of the initial compliance demonstration items above.
Avoca has 180 days after 5/20/2019 to complete stack testing, submit the test results and submit the
NOCS. The stack test report(s) are due 60 days after completion of the tests and the test results must be
submitted to EPA CDX.
I confirmed the above requirements with Gary Saunders (SSCB Supervisor) on 4/09/2019. It was noted
that Part 70 may compel DAQ to require testing for compliance against the 5D limits once per permit
term.
Monitoring and Recordkeeping
• Boiler tune-ups must be conducted annually;
The tune-ups that Avoca performs for compliance with 112j also meet the 5D requirements.
• Keep records of monthly fuel use;
• An oxygen analyzer system does not have to be installed if the boilers test in compliance with the
emission limits;
• Load monitoring is not required if the boilers test in compliance with the emission limits;
• Records of the occurrence and duration of each malfunction of each boiler and records of actions
taken to minimize emissions including corrective actions to restore to normal operation;
• Records of the date,time, and duration of each startup and shutdown,and type and amount of fuels
used during each startup and shutdown.
Reporting Requirements
Avoca must submit semi-annual summary reports beginning 1/30/2020. The permit condition outlines
the required content of the reports and they must be submitted to EPA CEDRI.
ENGINES/GENERATORS
ES-PkGenl was operating on 1/29/2019 due to power outage. All of Merry Hill area had lost power for
several hours. The plant remained in operation. I observed VE from the generator at approximately 5%.
E,nilsgiou Source �nii$�ion Source Descrnption` Corifrnillevice Controlxlevip8 77777
1b No. . : :' 1D No,; De§a tion
ES-PkGenl One No 2 fuel oil-fired limited use emergency One catalytic oxidizer
generator(2,935 kilowatt maximum rated power CD-CatOxl (695°F minimum inlet
output) temperature)
E101 One No.2 fuel oil-fired emergency generator N/A N/A
(587 horsepower maximum rated power output)
E102 One No.2 fuel oil-fired emergency generator N/A N/A
(760 horsepower maximum rated power output)
E104 One No.2 fuel oil-fired emergency generator(401 N/A N/A
horsepower maximum rated power output)
FP One No.2 fuel oil-fired emergency fire water pump N/A N/A
(285 horsepower maximum rated power output)
(SharePoint\Facilities\Bertie08\00044\agl6s\2018-2019 Inspection\20190130a16.doc) Page 8

Regulated Pollutant Umits/Standards Appl cnbie Regulation
Sulfur dioxide 2.3 pounds per million Btu heat input 15A NCAC 2D .0516
Visible emissions 20 percent opacity 15A NCAC 2D .0521(d)
Hazardous Air Pollutants Maximum Achievable Control Technology 15A NCAC 2D .1 I I I
[40 CFR Part 63,Subpart ZZZZ]
Nitrogen Oxides(NOx) Emit less than 40 tons of NOx per consecutive 15A NCAC 2Q .0317 to avoid
12-month period(PkGenl only) 15A NCAC 2D .0530
2D.0516 "Sulfur Dioxide Emissions from Combustion Sources"
Permit Condition 2.I.B.I. and 2.I.F.1.
The sulfur dioxide emitted by the generators shall not exceed 2.3 pounds per million Btu heat input. The
SO2 AP-42 emission factor from large engines (>500 hp) is 0.5 lb/MMBtu. These engines fire ultra-low
sulfur diesel. There are no monitoring/recordkeeping/reporting conditions under this rule in the permit
for the engines. Compliance is indicated.
2D.052I "Control of Visible Emissions"
Permit Conditions 2.I.B.2. and 2.I.F.2.
The engines are all limited to an opacity of 20%. The limits may not be exceeded more than once in any
hour and not more than four times in any 24-hour period. There are no monitoring/recordkeeping/reporting
requirements in the permit for the engines under this rule. Gregory Poole conducts quarterly maintenance
checks on the engines, and Avoca maintenance staff perform weekly maintenance/testing. The engines
are compliant with 2D.0521.
15A NCAC 2D.1111 "Maximum Achievable Control Technology"
40CFR 63 Subpart ZZZZ"Stationary Reciprocating Internal Combustion Engines NESHAP"
Permit Conditions 2.1.B.3. and 2.1.E3.
Avoca originally chose to categorize PkGenl as an existing limited use engine, which means it may
operate for emergency or peak shaving less than 100 hours per consecutive 12-month period in
accordance with 40 CFR §63.6590(b)(1)(ii). Avoca has been operating this engine for emergency reasons
(no peak shaving). Permit T52 will change PkGenl to an emergency engine and eliminate Permit
Condition 2.1.B.3. The other engines are only for emergency use.
There is another engine E106 associated with a county water pump to keep pressure to the county line if
process water is disrupted. Avoca does own this engine. Mr. Conner identified it as a portable unit,but
it's my understanding it has been kept in place for water system emergency use for several years. It was
installed in 2005. The manufacture date is 10/25/2005. The engine is 105 hp in size. This engine is
considered an insignificant source, and should be added to the permit insignificant list or permit list(if
requested by facility)with the next permit modification.
Operating, Maintenance and Monitoring Requirements
The following requirements apply to E104 and FP (and E106 if considered stationary):
• Change oil and filter every 500 hours of operation or annually,whichever comes first.
• Inspect air cleaner every 1,000 hours of operation or annually, whichever comes first.
• Inspect all hoses and belts every 500 hours of operation or annually, whichever comes first.
• Minimize the engine's time spent at idle during startup and minimize the engine's startup time to a
period needed for appropriate and safe loading of the engine,not to exceed 30 minutes.
• There is an option to participate in the oil analysis program to avoid annual oil/filter replacements.
(SharePoint\Facilities\Bertie08\00044\ag16s\2018-2019 Inspection\20190130a16.doc) Page 9

• Operate and maintain the engines according to the manufacturer's emission-related written
instructions or develop a maintenance plan.
• Install a non-resettable hour meter.
All of the engines must meet the following requirements:
• An emergency engine may be operated up to 100 hours per year for maintenance checks and
readiness testing. The regulation requires documentation of how many hours are spent for emergency
operation vs. hours for maintenance and testing(date, start time and end time of the engine operation
for these purposes). Up to 50 of those hours may be used for non-emergency operations, as long as
it's not for sale to grid/peak shaving.
Avoca operates their engines according to the manufacturer manuals. Avoca boiler maintenance staff
perform weekly maintenance checks(about 30 minutes) on all of the engines. Gregory Poole is
contracted to perform quarterly maintenance checks and full annual inspections on all of the engines. The
most recent inspection and maintenance for Pk-Gen and FP that satisfies the above requirements was
conducted on 4/27/2018. The most recent inspection and maintenance conducted for E101, E102, E104
that satisfies the above requirements was 4/24/2018. There was an inspection and maintenance record
observed for E106 that demonstrated compliance for 2018. I did not write down the date.
The permit limits PkGenl to 100 hours operation per rolling 12-months. Outside of maintenance hours,
Avoca only uses PkGEn 1 for emergency purposes. Since last inspection WaRO has agreed to treat
PkGen as an emergency generator and no longer a limited-use engine.
Recordkeepiing Requirements
The following requirements only apply to E104 and FP (and E106 if considered stationary):
• Records of the occurrence and duration of each malfunction, and actions taken to resolve the
malfunction.
• Records of all required maintenance.
• Records of the hours of operation of the engine that is recorded through the non-resettable hour meter.
The owner or operator must document how many hours are spent for emergency operation, including
what classified the operation as emergency and how many hours are spent for non-emergency
operation.
There were no malfunctions recorded for the engines for CY2018. Gregory Poole performs most of their
manufacturer required maintenance. Those records are on file at Avoca or accessible at Gregory Poole.
Avoca also creates SAP maintenance logs.
All of Avoca's engines have hour meters, including E 106. I reviewed hour records for all engines except
E106. Hours are recorded monthly, and the records discern between emergency and maintenance/testing
hours. E101,E102,E104 and FP did not operate for emergency purpose in 2018 and did not exceed 30
maintenance hours for any given 12-month period since February 2018. Pk-Gen operated 111.5
emergency hours in 2018. 82 of those hours were during the rare, extreme subzero weather event in
January of 2018. Copies of the hours records are attached to the file copy of this report.
During the inspection I read the hour meter on each engine:
PkGenl 1,354 hrs E101 832 Ins
E102. 1,940.2 Ins E104 304.6 hrs
FP 348 Ins
(SharePoint\Facilities\Bertie08\00044\ag16s\2018-2019Inspection\20190130a16.doc) Page 10

Reporting Requirements
The following semi-annual reporting requirements apply to E104 and FP.
• Company name and address, responsible official certification/signature, date of report and beginning
and ending dates of the reporting period;
• If a malfunction or instance of noncompliance occurred during the reporting period,the report must
include the number, duration, cause, and mitigation for each malfunction/instance;
• If there are no instances of noncompliance with any emission or operating limitations, a statement
that there were no instances of noncompliance with the emission or operating limitations during the
reporting period;
Reports have been submitted on time and are complete.
2Q.0317 "Avoidance Conditions"and 2D.0530 "Prevention of Significant Deterioration"
Permit Condition 2.I.B.4.
Generator ES-PkGenl must emit less than 40 tons of nitrogen oxides per consecutive 12-month period.
To stay under 40 tpy,the engine must not exceed 1,500 hours in a consecutive 12-month period.
The permit conditions require monthly monitoring of hours of operation and semi-annual reporting of the
monthly and rolling annual hours. I reviewed their records and they were complete. The semi-annual
reports have been submitted on time. ES-PkGenl is in compliance with 2D.0317.
ROTOCEL,RECOVERY,SFG, SDE-1 AND SDE-2 OPERATIONS
l anlssjdn Source Control Device t pntrbl Device
F,mission Santee Description
TYj)1Sa. ID No, Rescri tion
Rotocel Operations (Operating during inspection)
Rotocel extractor, desolventizer,and solvent CD-31209 One chilled water(final)
separation/recovery condenser venting to
ES-1001-2-1-P1 CD-31209 is actually a grouping of three CD-1001-2-S-1 One packed tower scrubber
condensers(series). Scrubber inlet gas temp
gauge reading was 60°F. (8 gals/min minimummineral oil injection rate
CD-1001-2-C-1 One chilled water
ES-M-125A and Two storage and recycle process tanks condenser venting to
M-125B ASTs located behind the Rotocel process
(formerly tanks along woodline,near fuel oil tanks. CD-1001-2-S-1 One packed tower scrubber
ES-1001-2-1-P2) (8 gals/min minimum
mineral oil injection rate)
ES-1001-2-1-F Process equipment leaks NA NA
ES-1001-2-1-WW Rotocel Operations wastewater stream
No hexane expected in this W W stream. NA NA
Produces approximately 25,000 gals/day.
Recovery Operations (Operating.During Inspection)
Arcon process tank M-1 CD-1001-1-3 One chilled water
This is the hexane/sclareol recovery tank. (located above the condenser venting to
ES-1001-1-1-P1 This tank,the stripper, receiver M-21 and tank-black
process tank 1001-1-1-P3 emission lines insulation)
combine to a single line that ties in just after One packed tower scrubber
condenser 31209(located above M-1). CD-1001-2-S-1 (8 gals/min minimum
mineral oil injection rate)
(SharePoint\Facilities\Bertie08\00044\ag16s\2018-2019Inspection\2019013006.doc) Page 11

Emission Source Emission Source Description Control Device Control Device
ID No. ID No. Descei lion
ES-1001-1-1-P2 Stripper T-5 and receiver M-21 CD-1001-1-T5B One chilled water
Stripper reboils/vaporizes the hexane. M-21 (inside building) condenser venting to
vents to T-5.
Seven process tanks of various capacities and One packed tower scrubber
one fixed roof methanol storage tank M-25 CD-1001-2-S-1 ' (8 gals/min minimum
(7,050 gallon capacity) mineral oil injection rate)
One heavy sage oil tank(waste sage by-
product M-57,which is not in use)feeds back
to Rotocel,one hexane/MeOH tank(M-8),
and two McOH work tanks(M36A and
ES-1001-1-1-P3 M36B)located together,and two
hexane/MeOH storage tanks (used to wash
carbon cell)M-35 and M-41 are on the
opposite side of the building.M2,M36AB
and M8 tanks gas to the headspace of the MI
tank that ties into the scrubber line.There is
also an 8t1 McOH tank M-126 that we
identified for methanol storage(from use in
carbon cell cleanouts).It vents directly to
atmosphere.
ES-1001-1-1-F Process equipment leaks NA NA
ES-1001-1-1-WW Recovery Operations wastewater stream NA NA
Concrete Operations(NOT OPERATING)
ES-1001-1-2-P Six process tanks of various capacities CD-1001-1-2 chilled water condenser
located behind the M-36 tanks
HB-1 One steam-heated hot box NA NA
HB-2 One steam-heated hot box NA NA
HB-3 One steam-heated hot box NA NA
HB-4 One steam-heated hot box NA NA
ES-1001-1-2-F Process equipment leaks NA NA
ES-1001-1-2-WW Concrete Operations wastewater stream NA NA
Sclareol Recrystallization(SFG) Operations(Operating during inspection)
T-3001 One process tank(6,700 gallons capacity) Optional controls Optional controls
he tane CD-3003 Chilled water condenser
T-3002 thin 3005 Four process tanks(2,538 gallons capacity each)
he tane) CD-3004-S Mineral oil scrubber
T-3006 One storage tank(12,500 gallons capacity) never installed
New he tane tank outside of SDE-2 bldg.
T-3007 One process tank(12,500 gallons capacity)
New he tane tank outside of SDE-2 bldg.
C-3001 and 3002 Two centrifuges C-3002 is a new source.
C-3001 vents to the dryer vent.
R-3002 One reactor equipped with a chilled water
in old bldg. process condenser(EX-3003)located right
above the reactor.
R-3003 One reactor equipped with a chilled water
process condenser(EX-3004)Condenser is
physically labelled EX-3003,but is EX-3005
on control room panel
R-3004 One reactor equipped with a chilled water
process condenser(EX-3005)Condenser is
physically and electronically labeled EX-3004
(SharePoint\Facilities\Bertie08\00044\ag16s\2018-2019Inspection\20190130a16.doc) Page 12

Emission Source ID No. tion'
Emission Source Description Control Device Control Device
ID Na. Dessri lion
R-3001 One reactor equipped with two chilled water CD-3001 Chilled water condenser
in old bld 2rocess condensers(EX-3001 and EX-3002)
D-3001 One steam-heated dryer equipped with a chilled Optional controls Optional controls
in old bldg. water process condenser(EX-3002-this is a CD-3003 Chilled water condenser
small unit)
CD-3004-S Mineral oil scrubber
never installed
D-3002 One steam-heated dryer equipped with a chilled CD-3002 Chilled water condenser
water process condenser(EX-3006)The
condenser ID is actually EX3105. Two shell- Optional controls Optional controls
in-tube exchangers in series. CD-3003 Chilled water condenser
CD-3004-S Mineral oil scrubber
never installed
ES-1003-10-F Process equipment leaks NA NA
ES-1003-10-WW SFG Process wastewater stream NA NA
Sclareolide(SDE-1) Operations (not operating during inspection)
• Hexane storage tank(M-2)
• Process tank(M-4)with chilled water process
condenser(T-6)and Process tank(M-4A)
with chilled water process condenser(T-6A)
• Receiver tank(M-39)for M-4/T-6
ES-1001-1-3-P • Process tank(M-44)with two chilled water NA NA
MACT FFFF process condensers(T-13 and T-15)
• Receiver tank(M-16)for T-13 and T-15
• Crystallizer tank(M-15)(next to M-44)
• Tank for cooled hexane for centrifuge(M-11)
• Receiver tank for dryer condenser(TK-1210)
• Process tanks(M-17 and M-17A)
G-17 One centrifuge NA NA
D-1202 One steam-heated dryer with condenser NA NA
ES-1001-1-3-Filters Filters NA NA
ES-1001-3-F SDE-1 process equipment leaks NA NA
ES-1001-1-3-WW SDE-lwastewater stream NA NA
Sclareolide(SDE-2) Operations (Operating during inspection)
T-4001 One 17,900 gallon virgin solvent tank
T-4017 and T-4018 Two 6,000 gallon process tanks
R-4004 One 4,200 gallon reactor with process condenser
(EX-4001) Chilled water condenser
R-4005 One 4,200 gallon reactor with process condenser CD-4002
(EX-4002)
R-4044 One 4,200 gallon reactor with process condenser CD-4003-S
(EX-4003) Mineral oil scrubber
R-4015 One 1,500 gallon reactor(no associated process
condenser)
C-4001 One centrifuge
CD-4001 Chilled water condenser
D-4001 One dryer with process condenser(EX-4004) CD-4002 Chilled water condenser
CD-4003-S Mineral oil scrubber
ES-4000-F SDE-2 process equipment leaks NA NA
ES-4000-WW SDE-2 process wastewater stream NA NA
(SharePoint\Facilities\Bertie08\00044\agl6s\2018-2019 Inspection\20190130al6.doc) Page 13

Regulated Limits/Standards Applicable Regulation`
Pollutant
Best Available Control Technology
Permit Conditions 2.2 B.1. and 2.2 B.2. 15A NCAC 2D .0530
Excluding two solvent recycle process tanks
(ES-M-125A and M-125B)
Volatile organic compounds Compliance Assurance Monitoring
Rotocel and Recovery Operations only 15A NCAC 2D .0614
Permit Condition 2.1.C.1.
Work practice standards 15A NCAC 2D .0958
Permit Condition 2.2 A.1.
Hazardous Air Pollutants Maximum Achievable Control Technology 15A NCAC 2D .1111
Permit Condition 2.2 C.1. [40 CFR Part 63,Subpart FFFF]
Odorous emissions Permit Condition 2.2 A.2. 15A NCAC 2D .1806
Toxic Air Pollutants Permit Condition 2.2 A.1. 15A NCAC 2D .0705/.0711
15A NCAC 2D .1100
2D.0958, 2D.1100, 2D.1806, and 2D.1111 are discussed in the Multi-Source Applicable Regulations and
Permit Conditions Section 2.2. later in this report.
There are four processes involved in producing sclareolide at Avoca;Rotocel,Recovery, SFG
(recrystallization), and SDE (sclareolide purification). In the Rotocel area sage is carried from storage to
a bay early each morning and is angered to a belt that carries it to an extractor. Hexane isomer is added
to the extractor to strip out sclareol from the sage(holds approximately 10,000 gallons).
Avoca has two suppliers of hexane; one sells a blend with less than 1.01/o n-hexane by wt., and the other
sells a blend of less than 2.6%n-hexane by wt. For each delivery they calculate the percent n-hexane in
the blended tanks using information in the delivery certifications. As of 1/24/2019 the iso-hexane blend
in their tanks contain 2.39%by weight n-hexane. The most recent shipment contains 2.01%n-hexane.
Over the last year n-hexane has been about 2.5%by wt. in the tanks. Note as long as they can maintain
less than 5% by wt. n-hexane there should be no issue with toxics limits.
The spent sage goes to the desolventizer,which steams the hexane from the sage. Hexane that is
vaporized is condensed(CD-31209)and collected in the solvent separation/recovery tank(M-1).
Condensed water and hexane separate in this tank(hexane floats). The recovered hexane is stored in the
two recycle process tanks(M-125A and 13).
The sage exiting the desolventizer is very hot and still contains some hexane. Most of the hexane flashes
fugitively out of the sage when it is removed from the desolventizer. Per Avoca's emissions inventory,
about 759.89 tons of iso-hexanes were emitted from Rotocel and Recovery in CY2017, and 674.45 tons
of iso-hexanes were emitted from the same areas in CY2016. Of the CY2017 total, 107.7 tons were
emitted from the final process vent at scrubber CD-1001-2-S-1.
The sclareol leaves the Rotocel process and goes to the Recovery process. Hexane is flashed off in the
stripper and the sclareol is mixed in a receiving tank(M-21)with methanol to further strip out the hexane.
The resulting purified oil settles to the bottom of the tank. It is then sent through a contro-mixer into
cans. The contro-mixer vents to a small condenser. There is negligible hexane and very little McOH
(ppb) in the sclareol(Avoca tests frequently for iso-hexane content). Avoca reported McOH emissions
from Recovery at 290.95 tons in CY2017 (as compared to 264.8 tons in CY2016).
The liquid mix in M-21 goes to M-8 and settles out the hexane(top)from the methanol (bottom). The
hexane fraction goes to M-1, and the McOH fraction is stored in any of the McOH tanks.
(SharePoint\Facilities\Bertie08\00044\agl6s\2018-2019 Inspection\20190130aA.doc) Page 14

The heavy oil tank M-57 contains a sludge removed from the final oil product. There should be
negligible solvent in it. Avoca currently isn't using this tank, and instead is piping the sludge back into
the Rotocel extractor. It can be shipped off-site as waste to a cement plant for use. If there is any solvent
left in it, it is recycled through the desolventizer.
In the past,Avoca has imported a small volume of thicker sage oil from overseas that is a"concrete wax." It
is piped directly into the Rotocel extractor, but if too solid, it can be heated first in the concrete operation hot
boxes. There are no emissions expected. The concrete operation did not run in CY2018.
The sclareol oil cans are carried to the SFG(sclareol recrystallization)process. The oil is emptied into a
reactor along with high purity heptane. Ethyl acetate is added to help with the reaction. There is a second
reactor that is used to re-process oil that doesn't complete the conversion in the first reactor. After every four
batches,there is usually a re-crop batch. After about four re-crop batches there is a third crop batch.
After the reactor,the sclareol goes to a centrifuge and dryer. The sclareol comes out of the dryer as a very
pure white powder(95% pure).
The SFG process expansion equipment began operation in CY2017. The new SFG reactors run the first and
second crops, and the old SFG equipment processes third crops. Note that scrubber CD-3004-S has not been
constructed.
From SFG the sclareol is bagged and shipped to an Avoca plant in Wisconsin. The sclareol is converted to
sclareolide there and is returned to Merry Hill Avoca for final purification in the SIDE (Sclareolide)process.
SDE-1 sources have not operated since the startup of SDE-2 in CY2017.
The sclareolide is placed in tanks M-4 and M-4A along with iso-hexane (approx.38-40%
n-hexane). The mixture goes through filtration to M-44,M-15, G-17 and D-1202. The final product is still a
white powder. Every batch is tested for percent hexane. The results are indicative of how well the
condensers are working.
2D.0614 "Compliance Assurance Monitoring"(ROTOCEL AND RECOVERY)
Permit Condition 2.1.C.1.
CAM applies to the Rotocel and Recovery sources(the processes can potentially emit over 100 tpy
uncontrolled VOC emissions).
Control Requirements/Parameter Ranges
• VOC emissions from the Rotocel extractor, desolventizer, and solvent separation/recovery (ES-1001-
2-1-P),two solvent recycle process tanks (M-125A and M-125B),Arcon process tank M-1 (ES-1001-
1-1-P1), stripper T-5 and receiver M-21 (ES-1001-1-1-P2), and eight process/storage tanks(ES-1001-
1-1-P3) shall be controlled by the associated packed tower scrubber(CD-1001-2-5-1). Avoca must
maintain a daily average mineral oil temperature at the inlet of scrubber (1001-2-5-1)of less than or
equal to 100 °F whenever the associated sources are operational.
Monitorine
• Monitor the inlet mineral oil temperature of the packed tower scrubber at least once each day.
• Perform periodic inspections and maintenance as recommended by the equipment manufacturer.At a
minimum,the inspection and maintenance shall include the following(as per permit condition
2.2.13.1.j.):
i. An annual inspection of spray nozzles and packing materials, chemical feed system (if so
equipped), and perform maintenance and repair when necessary to assure proper operation of the
packed tower scrubber
ii. An annual inspection, cleaning, and calibration of all associated instrumentation.
iii. Whenever the packing is replaced, inspect for nozzle plugging and settling of the packing.
(SharePoint\Facilities\Bertie08\00044\ag16s\2018-2019Inspection\20190130al6.doc) Page 15

• Install,maintain,operate, and calibrate the temperature gauge in accordance with manufacturer's
recommendations.
• Operation of the mineral oil packed tower scrubber CD-1001-2-5-1 with an inlet mineral oil
temperature that exceeds 100°F is considered an excursion. Avoca is required to take appropriate
action to correct an excursion as soon as practicable. If the packed tower scrubber CD-1001-2-S-1
operates under excursion for more than 5% of the Rotocel and Recovery operational time during a
consecutive 6-month period,then Avoca must develop a Quality Improvement Plan (QIP) in
accordance with 40 CFR§64.8.
Recordkeeping
Avoca is required to maintain all monitoring, inspections,maintenance and calibrations in a logbook:
• The date and time of each recorded action;
• The results of the monitoring,noting any excursions along with any actions taken to correct the inlet
mineral oil temperature of packed tower scrubber CD-1001-2-S-1;
• The results of any inspections or maintenance performed on mineral oil packed tower scrubber CD-
1001-2-5-1 or the associated temperature and flow rate gauges.
I reviewed the daily scrubber mineral oil temperature readings for CY2018 through 1/23/2019. A copy of
the 2018 daily records was obtained during the inspection, and it is attached to the file copy of this report.
There were no excursions. The records were complete. Oil temperature daily measurements never
exceeded 82°F. Copies of the records are provided by Avoca in their semi-annual reports. The data
matches that reported semi-annually.
An internal inspection was completed on 5/23/2018. On that date the nozzles and packing were replaced,
the mineral oil was changed out,and the flowmeter was calibrated. A copy of the 5/23/2019 inspection
record is attached to the file copy of this report. The scrubber underwent teardown on 8/22/2018, and the
flowmeter was replaced at that time. The most recent internal inspection was conducted 1/21/2019,but
the maintenance record had not yet been entered into the environmental department's inspection log.
The scrubber(and condensers) is also inspected internally monthly. I reviewed the logs for CY2018
through January 2019. The Rotocel operators have an hourly checklist for walkby visual inspection of the
scrubber.
All temperature gauges are checked and the flowmeter is calibrated during the annual tear down and
inspection of the scrubber. They replaced the temperature gauges and calibrated the flowmeter during the
annual inspection on 5/23/2018. They inspect the temperature gauge when performing quarterly
calibrations for all instruments. If the gauge is mis-reading, it's replaced.
Reporting
Avoca is required to submit a summary semi-annual report of the monitoring and recordkeeping
activities. The reports have been on time,reviewed and deemed complete(through January 2019 report).
Avoca appears to be in compliance with the CAM requirements for Rotocel and Recovery sources.
2D.0530 "Prevention of Significant Deterioration"(SFG)
Permit Condition 2.LG.1.
Avoca submitted a PSD application for the SFG expansion(T45 issued). SFG is limited to no more than
217.4 tons VOC emitted per consecutive 12-month period.
(SharePoint\Facilities\Bertie08\00044\ag16s\2018-2019Inspection\20190130a16.doc) Page 16

MonitorinWRecordkeeping
• VOC emissions must be calculated and recorded at the end of each month. VOC emissions must be
calculated by multiplying the total amount of each VOC-containing material consumed during the
month by the VOC content of the material.
• Re orting
Avoca must submit semi-annual reports containing the following data:
i. the monthly VOC emissions for each of the previous 17 months; and
ii. the yearly VOC emissions for each consecutive 12-month period ending on each month of the
previous six-month period.
I reviewed the records for CY2018. Avoca tracks heptane usage daily with their batches, and Mr. Conner
compiles the daily data to a monthly spreadsheet. The monthly and rolling 12-month data were complete.
The annual rolling totals did not exceed 140.19 tons.
Their semi-annual reports have been submitted on time and have been reviewed. Avoca is in compliance
with the permit condition.
2D.0530 "Prevention of Significant Deterioration"(SDE-2)
Permit Condition 2.1.1.1.
Avoca submitted a PSD application for the SDE-2 expansion(T46 issued). SDE-2 is limited to no more
than 354.4 tons of VOC emitted per consecutive 12-month period.
Mon itoring/Recordkeepine
• VOC emissions must be calculated and recorded at the end of each month. VOC emissions must be
calculated by multiplying the total amount of each VOC-containing material consumed during the
month by the VOC content of the material.
Re orting
• Avoca must submit semi-annual reports containing the following data:
i. the monthly VOC emissions for each of the previous 17 months; and
ii. the yearly VOC emissions for each consecutive 12-month period ending on each month of the
previous six-month period.
I reviewed the records for CY2018. Avoca tracks iso-hexane usage daily with the SDE-2 batches, and Mr.
Conner compiles the daily data to a monthly spreadsheet,which is attached to file copy of this report. The
monthly and rolling 12-month data were complete. The rolling total did not exceed 61.91 tons for the
calendar year.
Their semi-annual reports have been submitted on time and have been reviewed. The data in the most recent
semi-annual report matches the copy I obtained during the inspection. Avoca is in compliance with the
permit condition.
(SharePoint\Facilities\Bertie08\00044\agl6s\2018-2019Inspection\20190130a16.doc) Page 17

BOTANICAL AND BIOMASS EXTRACTION OPERATIONS
Botanical Extraction Operations (.Not operating during inspection)
Emission SourcCptissiod.Source Description Control Device Control bevies
iD No: ID No. DescriptionImmersion extractor Z-1001,desolventizer CD-1001-11-EX1002 One chilled water
Z-1002, day tank 90024(6,000 gallon capacity), condenser venting to
first-stage evaporator EX-1012,second-stage
ES-1001-11-P evaporator EX-1013,distillation column EX- CD-1001-11-EX1003 One cryogenic (nitrogen)
90008,and nine process tanks of various condenser system
capacities(TKS 1002, 1003,1007 separator,
1009,1010,CIP,T1001 and two unlabeled
product tanks
ES-1001-11-F Process equipment leaks NA NA
One plant material grinder(1,011 pounds per One bagfilter(244 square
MHZ-1002 hour nominal feed rate)Not used.Only non- CD-1003-4-1 feet of filter area)
MACT source at Botanical.
ES-1001-I1-WW Botanical extraction wastewater stream NA NA
Biomass Extraction Operations (Not operating during inspection)
One cartridge filter(6.7:1
ES-1004-1 Biomass extraction debagg ng CD-1004-1-FFl max air-to-cloth ratio)
ES-1004-2-F Process equipment leaks NA NA
ES-1004-2-WW Biomass extraction operation wastewater stream NA NA
PRIMARY OPERATING SCENARIO: CD-1004-2EX1002 One chilled water
Immersion extractor Z-41001,desolventizer condenser venting to
Z-41002,day tank 490025 (9,953 gallon
capacity),iso-hexane storage tank 490024 CD-1004-2EX1003 One cryogenic (nitrogen)
(13,536 gallon capacity),first stage evaporator condenser system
ES-1004-2-P EX-41012,second stage evaporator EX-41013,
distillation column EX-490008(not installed),
and nine process tanks of various capacities
(TKS 1001,1004, 1005, 1006,1007, 1009,
1011,and 1012)There are two day tanks
labeled TK-1024 and 1025 that are the same
as 490024 and 490025.
SECONDARY OPERATING SCENARIO: CD-1004-2EX1002 One chilled water
Immersion extractor Z-41001, condenser
day tank 490025(9,953 gallon capacity),
storage tank 490024(13,536 gallon capacity), CD-1004-2EX1003 One cryogenic (nitrogen)
first stage evaporator EX-41012, condenser system
second stage evaporator EX-41013,
distillation column EX-490008,and
nine process tanks of various capacities
NEW EQUIPMENT: Tray Dryer(RV-1002)
equipped with bagfilter(DC-1001)vented to
a vent condenser(HX-1001)and a solvent
knockout pot(TK-1002)
Biomass silo loadout This was not used in One bagfilter (254 square
ES-1004-2Silo 2017. Chick pea trial product was loaded CD-1004-2-FF2 feet of filter area)
directly to truck.
Molecular sieve(does the same solvent/water
ES-MSDU-1024 separation job as what a distillation column N/A N/A
would have done
(SharePoint\Facilities\Bertie08\00044\agl6s\2018-2019 Inspection\20190130a16.doc) Page 18

The Botanical Extraction area was not operating during the inspection. Mr. White and I came to a conclusion
that the CIP tank is actually tank No. 1010, and there are three unlabeled tanks rather than two (two are inside
of the building and one is next to TK1009). The CIP tank contains water and caustic, so really should not
even be considered for permitting.
Botanical Extraction operated from 12/05/2018 through 12/19/2018 for a trial that used hexane and methanol
for extraction. It operated 3/29/2018 through 4/09/2018 for cocoa production(acetone solvent), and
8/18/2018 through 8/24/2018 for tobacco flavoring production(heptane solvent).
Biomass Extraction was not in operation during the inspection. Biomass operated 2/13/2018 through
3/20/2018 and 6/25/2018 through 7/13/2018 processing chick peas using ethanol. The chick peas are the
product, not the extracted material. The equipment for the secondary operating scenario is installed,but not
yet operating. This equipment will be used for chick pea extraction.
Note that during the physical inspection Mr. White and I could not find tank TK-1002.
The Botanical and Biomass Extraction equipment are very similar,but the Biomass equipment can run
bulkier materials. There are actually two cryogenic condensers in each system. When Avoca uses
hexane, it tends to freeze completely and plug the condensers, so they have a second condenser to use
while the frozen condenser is thawed with steam injection. The coolant is nitrogen at-2127 out of the
tank. The condensers are not freezing when employing ethanol or other extraction solvents.
Mr. White said that Avoca plans to run Botanical for cocoa and Biomass for chick pea within the next
month.
Regulated Rollutant Limlts/Standards Appllcable Regulatioq
Particulate matter Plant material grinder,biomass extraction debagging,
and biomass silo loadout Permit Condition 2.I.D.1, 15A NCAC 2D.0515
Visible emissions 20 percent opacity Permit Condition 2.1.D.2. 15A NCAC 2D .0521(d)
Best Available Control Technology
Volatile organic compounds See Sections 2.2 B.1. and 2.2 B.2, 15A NCAC 2D .0530
Compliance Assurance Monitoring 15A NCAC 2D .0614
Permit Condition 2.1.D.3.
Odorous emissions Permit Condition 2.2 A.3. 15A NCAC 2D .1806
Hazardous Air Pollutants Maximum Achievable Control Technology 15A NCAC 2D .1111
Permit Condition 2.2 C.1. (Botanical only) 40 CFR Part 63, Subpart FFFF
Toxic Air Pollutants LPe:rmftt�Condition .2.A.2 15A NCAC 2Q .0705/.0711
15A NCAC 2D .1100
2D.I100, 2D.1806, and 2D.III1 are discussed in the Multi-Source Applicable Regulations and Permit
Conditions Section later in this report.
2D.0515 "Particulates from Miscellaneous Industrial Processes"
Permit Condition 2.I.D.1.
The debagger, silo and plant grinder have not operated since the date of last DAQ inspection. The
debugger is not used for the chick pea process. The new tray dryer has not yet started operation.
Particulate matter is limited by the following equation:
E=4.1OxP111
Where: E=allowable emission rate in pounds per hour; and
P=process weight in tons per hour
(SharePoint\Facilities\Bertie08\00044\agl6s\2018-2019Inspection\20190130a16.doc) Page 19

To demonstrate compliance,Avoca is required to perform inspections and maintenance as recommended
by the manufacturer. At a minimum the bagfilters and system ductwork must be externally inspected
monthly and internally inspected annually. I reviewed the bagfilter monthly external and annual internal
inspection records for the baghouses. Mr. Conner kept records noting non-operation each month. The
most recent dates of internal inspection are 6/22/2018 for debagging bagfilter and 6/27/2018 for the
storage silo bagfilter.
A semi-annual summary report of inspection and maintenance is required for these baghouses. The
reports have been complete and submitted on time. Compliance with 2D.0515 is indicated.
2D.0521 "Control of Visible Emissions"
Permit Condition 2.I.D.2.
This permit condition applies to the Biomass debagger and silo. Each is limited to an opacity of 20%
under this rule. The limits may not be exceeded more than once in any hour and not more than four times in
any 24-hour period. To ensure compliance, Avoca is required to observe visible emissions from the
debagger and silo baghouses on a semi-annual basis. Neither of these sources have operated since the
date of DAQ's last inspection. Mr. Conner still performed walk-bys every month anyway for both
sources.
Avoca is required to submit a semi-annual summary report for VE observations of the operations. The
reports have been complete and submitted on time. The Botanical and Biomass operations appear to
be compliant with 2D.0521.
2D.0614 "Compliance Assurance Monitoring"
Permit Condition 2.I.D.3.
CAM applies to the Botanical ES-1001-I I-P and Biomass ES-1004-2-P sources because the processes
can potentially emit over 100 tpy uncontrolled VOC emissions.
Control Requirements/Parameter Ranges
• VOC emissions from the subject botanical extraction operations sources must be controlled by
associated cryogenic condenser system ID No. CD-1001-11-EX1003.VOC emissions from the
biomass extraction operations sources must be controlled by cryogenic condenser system ID No. CD-
1004-2EX1003. The 12-hour average outlet temperature of both of the condensers must be less than
or equal to -40 °F whenever the associated sources are using non-water soluble solvents(like hexane).
The 12-hour average must be maintained at less than or equal to 17 OF when using water soluble
solvents (like ethanol).
Monitoring
• Avoca is required to monitor the outlet temperatures of the cryogenic condenser systems at least once
each hour(when operating), and calculate the average outlet temperature for the consecutive 12-hour
period ending with that hour.
I reviewed the hourly measurements for both sets of Botanical and Biomass cryogenic condensers
during the few weeks they operated in CY2018. Copies of the 12-hr average records are attached to
the file copy of this report. Hourly records for Botanical in December 2018 when they used hexane
for a trial extraction are also attached to the file copy of this report. There were no 12-hr average
exceedances. Note that acetone is de-listed, so no monitoring is required when it is employed.
• Per condition 2.2.13.l.h. periodic inspections and maintenance as recommended by the equipment
manufacturer are required. In addition,Avoca must perform an annual inspection of each condenser
system, including the following:
(SharePoint\Facilities\Bertie08\00044\agl6s\2018-2019 Inspection\20190130a16.doc) Page 20

i. inspect and maintain the structural integrity of each condenser, including inspection for leakage
of coolant and, if the system is under positive gauge pressure,leakage of the contaminated gas
stream. In order to monitor leakage of the coolant,the condensate shall be inspected for the
presence of coolant.
ii. inspect/maintain the structural integrity of ductwork/piping leading to/from each condenser.
The Botanical condensers were most recently structurally inspected per the permit condition on
11/07/2018 in preparation for the trial. The Biomass condensers were most recently inspected on
6/22/2108 prior to chick pea operation.
Mr. Conner also conducts monthly external inspections even when the condensers are not in
operation. He performed inspections 3/15/2018 and 7/12/2018 during chick pea processing,and
12/18/2018 while the trial was running.
• Per condition 2.2.13.Li. Avoca must install,maintain, operate, and calibrate, in accordance with
manufacturer's recommendations, a sensor to continuously measure the outlet temperature of each
condenser. Each sensor must be installed in an accessible location and be maintained such that it is in
proper working order at all times. Note that condition 2.I.D.3. incorrectly references permit
condition 2.2.B.Li. as 2.2.B.1 j.
The sensors are checked quarterly. During my inspection in 20181 spoke with maintenance
personnel and he said that temperature gauges are replaced immediately after discovery of failure, and
they are replaced annually when the condenser internal inspections are completed. He noted that the
control system is programmed to shut down the process as soon as a cryo-condenser temperature
average goes out of range.
No temperature monitoring was conducted during the cocoa run in March due to use of acetone,
which was de-listed as a VOC.
• Operation of the condenser systems with a 12-hour average outlet temperature that exceeds a limit is
considered an excursion. Avoca is required to take appropriate action to correct an excursion as soon
as practicable.Further, if excursions occur for more than 5 percent of the operational time during a
consecutive 6-month period,then Avoca must develop a Quality Improvement Plan (QIP) in
accordance with 40 CFR§64.8.
There were no excursion periods recorded for CY2018.
Avoca has a cryogenic condenser SSM plan.
Recordkeenine
• The results of monitoring, inspections,maintenance and calibrations must be maintained in a log that
contains
ii. the date and time of each recorded action;
ii. The results of the monitoring,noting any excursions along with any actions taken to correct the
outlet temperature of cryogenic condenser systems;
iii. The results of any inspections or maintenance performed on cryogenic condenser systems or the
associated temperature gauges; and
iv. Any variance from manufacturer's recommendations, if any, and corrections made.
The temperature measurement and inspection records appeared to be complete.
Reporting
Avoca is required to submit a summary semi-annual report of the monitoring and recordkeeping
activities. They appear to be complete and on time. The condenser inspection log is included in the
reports.
(SharePoint\Facilities\Bertie08\00044\ag16s\2018-2019 Inspection\20190130al6.doc) Page 21

Avoca appears to be in compliance with the CAM requirements for Botanical and Biomass Extraction
sources.
PNE AND EVG OPERATIONS
Plant Nutrient Extraction (PNE) Operations (Not operating during inspection)
Emission Soflrce, Control Device
. ml §inn rgnttr.Ce DeseYiptiBn Cbntr,6 Dnvlf esoriptioli-
EX2203 One chilled water condenser
venting to
CD-Z-9215
One water spray fume scrubber
D31214 One product extract reactor (0.5 gals/min minimum water
CD-Z-9216 injection rate)venting to
One water spray fume scrubber
(0.5 gals/min minimum water
injection rate)
EX2205 One chilled water condenser
venting to
CD-Z-9215 One water spray fume scrubber
D31211 One waste solids separator vessel (0.5 gals/min minimum water
(1,333 gallon capacity) injection rate)venting to
CD-Z-9216
One water spray fume scrubber
(0.5 gals/min minimum water
injection rate)
Solvent Process Tank(9,500
ES-TK-PNE-1 gallons capacity)This tank is NA NA
not installed.
ES-1003-2-1-F Process equipment leaks NA NA
ES-1003-2-1-WW PNE process wastewater stream NA NA
One water spray fume scrubber
ES-1003-2-1-P Seven process tanks of various (0.5 gals/min minimum water
capacities CD-Z-9215 injection rate)venting to
C-31203 One centrifuge One water spray fume scrubber
D-1002 One dryer with a process chilled CD-Z-9216 (0.5 gals/min minimum water
water condenser injection rate)
Ethyl Vanillin Glucoside(EVG) Operations (Operating during inspection)
Emission Source i Control Devlec
II?No. knilsshinSourcebeseription ID)!1b. Control bevice'De§cription
D-2202 One reactor CD-Z-9215 One water spray fume scrubber
D-1215 One reactor (0.5 gals/min minimum water
injection rate)venting to
D-1218 One reactor
One steam-heated dryer(labeled One water spray fume scrubber
D-1201 as D-1002 on the equipment) (0.5 gals/min minimum water
CD-Z-9216 I injection rate)
ES-1003-2-2-F Process equipment leaks NA NA
ES-1003-2-2-WW EVG Operation wastewater stream NA NA
(SharePoint\Facilities\Bertie08\00044\agl6s\2018-2019 Inspection\20190130a16.doc) Page 22

PNE and EVG are in the same building. In CY2018 PNE equipment was used for the following products:
SFG sclareol 1/02-10/2018, 4/18/2018 through 6/11/2018, 7/18/2018through 8/23/2018, 10/12/2018 through
11/22/2018.
Nicotine blend 3/28-29/2018, 10/11/2018
Tobacco absolute 8/24/2018through 9/05/2018
SFG uses heptane extraction. Nicotine blending does not involve any solvents. Tobacco absolute employs
acetone extraction.
The EVG process uses chloroform to extract ethyl vanillin glucoside,which provides vanilla flavoring in
cigarettes (popular in Japan and China). EVG may also produce other products using ethanol or heptane
extraction,but has only produced vanillin glucoside in the last year.
Both EVG and PNE processes run a primary extraction and then a re-crop stage.The scrubbers,reactors
(D-2202, D-1215 and D-1218) and dryer D-1201 can actually be shared between EVG and PNE.
I performed physical inspection of the PNE and EVG sources. There did not appear to be any change to the
equipment and ductwork. No physical problems were identified for the sources.
Itegglated Pollutant liimits/Standards Applicable Regulation
Odorous emissions Odorous emissions must be controlled ISA NCAC 2D.1806
Permit Condition 2.2.A.2.
Toxic Air Pollutants Permit Condition 2.2 A.1. 15A NCAC 2Q .0705/.0711
15A NCAC 2D .1100
Hazardous Air Pollutants(HAP) Maximum Achievable Control Technology 15A NCAC 2D.I I I I
Permit Condition 2.2.C.1. [40 CFR Part 63,Subpart FFFF]
2D.1100, 2D.1806, and 2D.1111 are discussed in the Multi-Source Applicable Regulations and Permit
Conditions Section later in this report.
There are no individual conditions in section 2.1 of the permit for PNE or EVG.
Note that the PNE and EVG scrubbers are in series. Operation of water spray fume scrubber CD-Z-9215 and
water spray fume scrubber CD-Z-9216 is only required for control of emissions from the PNE and EVG
operations if ethyl acetate,hexane, chloroform or other HAP is being used as a solvent. Ethyl acetate is
generally used as a hexane sweetener to enhance extraction.
While not required by the permit,Avoca performs internal inspections on the scrubbers. The most recent were
completed on 3/15/2018 (teardown). The most recent maintenance was conducted on 11/20/2018.
BIOMASS BOILERS AND ROTARY DRYER
EmtssioA Spurce c Centrp(llevIce
ID'Nu.` unssibti 5 fui' c Desc iptiau C'entr4l Peviee Deseviliti4At
IV,
Nb.
ES-BBI and ES-BB2 Two biomass bio-based solids-fired
One simple cyclone
NSPS De CD-BB IC and (114 inches in diameter)in series
MACT DDDDD boilers (24 million Btu per hour CD-BB 1BH with one bagfilter(8,900 square
maximum heat input rate each) feet of filter area)
ES-RD One No.2 fuel oil/propane-fired One simple cyclone
rotary dryer(30.0 million Btu per CD-BB IC and (114 inches in diameter)in series
hour maximum heat input rate) CD-BB1BH with one bagflter(8,900 square
feet of filter area)
(SharePoint\Facilities\Bertie08\00044\agl6s\2018-2019 Inspection\20190130a16.doc) Page 23

The south boiler(BB 1)re-bricking project was completed 6/06/2018.
The dryer is no longer operated,because burning sage in the boilers was too difficult(the sage
combustion produced too much particulate and plugged the baghouse frequently). Application 19A was
recently submitted to have the dryer and cyclone removed from the permit. The cyclone's purpose was
more for sage product movement to the boiler than for emissions control.
Avoca shuts down the boilers for maintenance and tube cleanouts at least twice a year. The most recent
cleanout was the week previous to this inspection. Extended boiler maintenance shutdowns were on
5/21/2018 through 6/06/2018, 8/11/2018 through 9/18/2018 and 1/08/2019through 1/19/2019. When
operating,the boilers soot-blow every 10 minutes. The soot blowers are inserted into the boilers like a
tube tray, so it is very easy to perform maintenance on them.
Reg Bated ,Liduts/Standards Applleabie iregulotion
pollutant
Particulate matter Boilers(ID Nos.ES-BBI and B132): 15A NCAC 02D.0503
0.33 pounds per million Btu heat input
Rotary dryer(ID No. ES-RD): E=4.10(P)o.6r
Particulate matter 15A NCAC 02D.0515
Sulfur dioxide 2.3 pounds per million Btu heat input 15A NCAC 02D.0516
Visible emissions 20 percent opacity when averaged over a six-minute period 15A NCAC 02D.0521
N/A Initial notification requirements, 15A NCAC 02D.0524
Notification of boiler size and fuel combusted,and NSPS—Subpart De
Record and maintain amount of each fuel combusted during each
calendar month.
PM,HCI,CO,Hg, Per Subpart 15A NCAC 02Q .I I I
MACT Subpart DEEM
Volatile Organic Emissions of VOCs shall be less than 40 tons per consecutive 15A NCAC 02Q.0317
Com ounds 12-month period Avoidance of 2D.0530)
Hazardous Air Rotary Dryer: n-hexane emissions shall be less than 10 tons per 15A NCAC 02Q.0317
Pollutants year Avoidance of 21).It 12)
2D.0503 "Particulates from Fuel Burning Indirect Heat Exchangers"
Permit Condition ZI.E.1.
Particulate matter from the combustion of sage or wood shall not exceed 0.33 Ib/MMBtu for each boiler.
There is a more restrictive MACT particulate limit of 0.03 Ib/MMBtu on the boilers. The boilers have
been tested(M5 filterable)three times since startup. On 6/26/2012 the combined emission was 0.103
lb/MMBtu. It met the 213.0503 limit, but not the MACT limit. On 2/21/2013 the re-test result was
0.0012 lb/MMBtu. The most recent test was conducted on 12/14/2016, and the result was 0.0015
lb/MMBtu. All of the test reports have been reviewed and accepted by SSCB.
The permit requires Avoca to perform inspections and maintenance on the bagfrlter and cyclone as
recommended by the manufacturer. At a minimum Avoca must conduct monthly external inspections of
ductwork,the cyclone, and the bagfilter. Avoca must also perform an annual internal inspection of the
bagfrlter. The monthly inspection records were complete. The baghouse and cyclone monthly
inspections are on the same log. The baghouse is internally inspected monthly. The most recent internal
inspection was completed during the 1/8-19/2019 boiler shutdown.All of the bags were changed out on
6/03/2018.
The baghouse sensors determine when a purge is needed. It will automatically purge at 8 hours of
operation if it has not already purged earlier. Operators purge manually at least once per day as well.
(SharePoint\Facilities\Bertie08\00044\agl6s\2018-2019 Inspection\20190130a16.doc) Page 24

Avoca also checks and records the pressure drop across the baghouse daily.
Semi-annual summary reports to DAQ have been complete and on time. The biomass boilers appear to
be in compliance with 2D.0503.
2D.0515 `Particulates From Miscellaneous Industrial Processes"
Permit Condition 2.I.E.2.
Emissions of particulate matter from the rotary dryer shall not exceed an allowable emission rate as
calculated by the following equation:
E=4.1OxP111
Where, E=allowable emission rate in pounds per hour
P=process weight in tons per hour
The calculated limit does not include fuel oil use. The dryer can process a maximum of about 3 tons/hr.
The limit is calculated at 8.6 Ibs/hr. The total boiler and dryer emissions were last tested on
2/21/2013 at 0.0012 lb/MMBtu(48 MMBtu/hr)=0.06 lb/hr.
Avoca must maintain production records. The dryer is no longer operated. The dryer is
in compliance with 2D.0515.
2D.0516 "Sulfur Dioxide Emissions from Combustion Sources"
Permit Condition 2.I.E.3.
The SO2 emitted by the boilers shall not exceed 2.3 lb/MMBtu heat input. The boilers combust only
wood chip. Sage is no longer burned. The SO2 AP-42 emission factor for wood combustion is 0.025
lb/MMBtu. There are no monitoring/recordkeeping/reporting conditions under this rule in the permit for
the boilers. Compliance with 2D.0516 is indicated.
2D.0521 "Control of Visible Emissions"
Permit Condition 2.I.E.4.
While the boilers are NSPS and MACT applicable,those regulations do not include visible emissions
limits for the boilers. Therefore,the boilers and dryer are limited to an opacity of 20% under this rule.
The limits may not be exceeded more than once in any hour and not more than four times in any 24-hour
period. No visible emissions(0%)were observed from the boilers during the inspection.
Avoca is required to observe boiler opacity once a day. "Normal" is any VE below 20%opacity. Both
Brian and Colin are VE certified readers. The daily observation must be made each day of the calendar
year. I reviewed the observation records for CY2018 through January 2019. They are complete,and
match those included in the semi-annual reports. No problems with VE were recorded. The semi-annual
summary reports have been complete and on time. Compliance with 2D.0521 is indicated.
2D.0524 "New Source Performance Standards," ASPS 40 CFR Part 60,Subpart Dc
Permit Condition 2.I.E.5.
Subpart De applies to the boilers because they were constructed after 6/09/1989 and are larger than 10
MMBtu/hr. However,there are no emission limits for these wood-fired boilers because they are rated at
less than 30 MMBtu/hr. Avoca was required to submit notices of construction and startup. The
construction notice was submitted on 5/31/2011. The startup notices were submitted on 10/17/2011 and
11/07/2011.
Avoca is required to keep daily fuel consumption records. They track usage by using the total wood
received each day and the hours of operation. The CY2018 -January 2019 records appeared to be
complete. Avoca appears to be compliant with 2D.0524.
(SharePoint\Facilities\Bertie08\00044\agl6s\2018-2019 Inspection\2019013 Oat 6.doc) Page 25

2D.1111 "National Emission Standard for Hazardous Air Pollutants
40 CFR Part 63,Subpart DDDDD—Boiler and Process Heaters"
Permit Condition 2.I.E.6.
The boilers are considered new under the regulations because they were constructed after 6/04/2010. The
applicable emissions limits are as follows:
Limits,except during startup
Pollutant and shutdown Required Test Methods
Units in all subcategories HCI 2.2E-02 lb per MMBtu of heat For M26A,collect a minimum of 1
designed to bum solid fuel. input dscm per run;for M26 collect a
minimum of 120 liters per run.
Mercury 8.0E-07 lb per MMBtu of heat For M29,collect a minimum of 4
input dscm per run;for M30A or M30B,
collect a minimum sample as
specified in the method; for ASTM
D6784 collect a minimum of 4 dscm.
Stokers/sloped grate/others CO 620 ppm by volume on a dry MI0, 1 hr minimum sampling time.
designed to bum wet biomass basis corrected to 3 percent
fuel oxygen,3-ran average
Filterable PM 3.0E-02 lb per MMBtu of heat M1-5,collect a minimum of 2 dscm
(or TSM) input;or(2.6E-05 lb per per run.
MMl3m of heat input)
Testing
An initial performance test was required for the above pollutants. The initial performance tests were
conducted 6/26/2012. Avoca chooses to perform stack testing for all of the pollutants rather than fuel
sampling analysis for HCI and mercury. The boilers were in compliance, except for particulate(0.103
Ib/MMBtu). A NRE was issued to Avoca on 10/31/2012, and a civil penalty of$4,549.00 was assessed.
Due to problems with the bagfilter and bag leak detection system,the retest was not conducted until
2/21/2013. The results were compliant with the standard and approved by DAQ SSCB on 7/14/2013.
Testing was performed on 12/19/2013. The stack test observation report is in the WARO file.
it
PM 0.0014lb/mmBtu 0.03lb/mmBtu 5
CO 19 ppmv @ 3%02 620 ppmv 3%02 3
HCI 9.0E-5lb/mmBtu 0.022lb/mmBtu 0.4
Hg 2.95 E-7 lb/mmBtu 8.0 E-71b/mm 3tu 36.9
Those test results have been reviewed and accepted by SSCB. Because the results are all less than 75% of
the standards, like they were also for the 2/21/2013 tests,then Avoca was allowed by the regulation to
schedule their next performance tests three years from the 12/19/2013 test date. Therefore, the next tests
were conducted on 12/14/2016. The test observation report is in the WARD file. No problems were
observed with the test methods. The test results were submitted to DAQ postmarked 2/13/2017,and were
submitted to EPA ERT on 2/14/2017. 40 CFR 63.7575 (f)requires the test results and a NOCS to be
submitted within 60 days after test completion, and the postmark date to DAQ was 61 days following the
12/14/2016 test date. A NOV was issued on 2/22/2017 to Avoca for late submittal. SSCB has reviewed
and approved the 12/14/2016 test results. The next tests are not due until December 2019.
(SharePoint\Facilities\Bertie08\00044\agl6s\2018-2019 Inspection\20190130a16.doc) Page 26

PM 0.0015lb/mmBm 0.03lb/mmBtu 5
CO 214 pmv 3%02 620 ppmv 3%02 34.5
HCl 6.8E-5lb/mmBtu 0.022lb/mmBtu 0.3
H 2.9 E-7 Ib/mmBtu 8.0 E-7 Ib/mmBtu 36.3
Avoca must monitor the oxygen trim system for each boiler during the stack tests so that the lowest
hourly average oxygen concentration measured during the tests on each boiler is established as the
operating limit for the boiler. Avoca must also monitor the steam production rate during testing so that
the highest hourly average is established as an operating limit for the boilers. Avoca did establish the
percent oxygen and steam rate limits with their performance tests. The data was reviewed and accepted
by SSCB. See the discussion on page 30 for details about continuous monitoring requirements for
oxygen and steam.
Monitorin Recordkeepine
Tune-ups
Because the boilers employ an oxygen trim system, they did not have to do a tune-up until October 2016,
which was five years after startup. The following actions must be completed for the tune-up:
• Inspect the burner, and clean or replace any components of the burner as necessary.
• Inspect the flame pattern, as applicable, and adjust the burner as necessary to optimize the flame
pattern. The adjustment should be consistent with the manufacturer's specifications.
• Inspect the system controlling the air-to-fuel ratio, as applicable, and ensure that it is correctly
calibrated and functioning properly.
• Optimize total emissions of CO. This optimization should be consistent with the manufacturer's
specifications.
• Measure the concentrations in the effluent stream of CO in parts per million, by volume, and oxygen
in volume percent,before and after the adjustments are made (measurements may be either on a dry
or wet basis,as long as it is the same basis before and after the adjustments are made).Measurements
may be taken using a portable CO analyzer.
The tune-ups were completed in-house by Jared Findley on 9/01/2016 and 10/27/2016. Copies of the
tune-up records are stored in the WaRO file.
Avoca performs boiler inspection and tube cleanouts at least twice a year.
Start-up and Shut-down
The regulation requires Avoca to use natural gas, synthetic natural gas,propane, distillate oil, syngas,
ultra-low sulfur diesel,fuel oil-soaked rags, kerosene,hydrogen, paper,cardboard,refinery gas, or
liquefied petroleum gas as the fuel for startup. Avoca uses LPG, and sometimes uses a fuel oil soaked rag
to initially start the flame.
Avoca's SOP requires the baghouse to always be in operation before the boilers begin start-up, and it
continues to operate through a shut-down. It's easier to do a start-up with the baghouse fans running to
encourage building the flame.
The oxygen trim system does record data during startups and shutdowns. Steam production is also
continuously recorded.
(SharePoint\Facilities\Bertie08\00044\ag16s\2018-2019 Inspection\20190130a16.doc) Page 27

Baghouse Leak Detection System
The MACT requires bagfilters to have a bag leak detection system(BLDS). Jared Findley(816-863-
0590)manages the biomass boiler system, and is the contact for discussing all maintenance on the BLDS.
Avoca's system has a conductivity probe in the bagfrlter stack. When the alarm level condition triggers,
the unit shows an alarm on the touchscreen. The system measures the conductivity every second, and
retains the data(the data can be downloaded to a jump drive).
It is my understanding the alarm level is set to mirror the particulate MACT emission limit. They use the
emissions and conductivity measured during the most recent particulate 5D stack test,and ratio the
conductivity up by the difference between the measured particulate and the particulate limit
(approximately 20%). Therefore, if the alarm is triggered,then it could be assumed that the boilers are
exceeding the MACT limit. While excess emissions from the baghouse because of malfunction of the
bags wouldn't necessarily constitute a permit/regulation violation, it could still pose a compliance
problem for the facility in terms of credible evidence of an emissions violation,and would be a reportable
deviation for their annual compliance certification. I shared this observation by email to Mr.Findley on
3/23/2018, and Mr. Conner on 4/11/2018.
For day to day operation it is my understanding that the conductivity is represented on channel range of 0-
2000 on the control panel,with 10-100 being the normal rate during boiler operation. The alarm is set at
802. The actual signal range of the equipment is 4-20 ma.
Their leak detection system was initially certified by the manufacturer on 1/24/2013. The manufacturer
installed the system on the boiler stack for Avoca.
Two conductivity alarms occurred in CY2018 (none so far in January 2019). On 1/07/2018 the alarm
triggered during boiler shut-down. There was no actual conductivity reading,and the alarm did not last
more than a few seconds. The other alarm occurred briefly during last year's DAQ permit compliance
inspection on 1/31/2018. The reason for the alarm is unknown.Readings were reviewed during the
inspection following the alarm.
The temperature set point for the bagfilter is 200°F. There is a low temperature alarm at 125°F. 1
Avoca is required to conduct a performance evaluation of the bag leak detection system in accordance
with their BLDS monitoring plan and consistent with the guidance provided in EPA-454/R-98-015
(incorporated by reference(§63.14)). The guidance document is saved on the WaRO LAN at
\Bertie08\00044\AO16s\2017-2018 Inspection\EPA BLDS Guidance Document.pdf, and a hard copy is in
the WaRO file. I've read through the BLDS requirements in the MACT, as well as the guidance
document. I called Jared Findley on 3/23/2018 to discuss the BLDS regulatory requirements, and
followed up the conversation with an email. A copy of Mr. Findley's response to the email is attached to
the file copy of this report(dated 4/09/2018). A discussion of the questions and his responses is provided
below:
• Do you have a monitoring plan other than the manufacturer operating manual for the BLDS to share
with me? The guidance document requires a performance evaluation of the bag leak detection system
in accordance with your monitoring plan(incorporated by reference, see §63,14). If you have a
monitoring plan(or maintenance manual), does it recommend periodic inspection or preventative
maintenance activities? Is there any maintenance/inspection performed on the system other than just
periodic external observations and the calibrations with stack testing?
Avoca email response: "The current monitoring plan consists of routine supervision by the boiler
operators to verify that the unit is operating and reading(values are changing and the probe is
functioning). At least 4 times daily boiler operators review all panels in the boiler room to check for
errors or any alarms. The probe normally has an external visual inspection once quarterly from the
stack access platform. The probe is normally cleaned during stack testing and data from the stack
(SharePoint\Facilities\Bertie08\00044\agl6s\2018-2019 Inspection\20190130a16.doc) Page 28

testing is used for calibration on alarm limits. During boiler shutdown periods the unit is evaluated to
verify that it returns to a Zero reading. The unit is also back-checked by environmental personnel
daily when they do opacity checks on the system for any particulate entering the air. I have attached
a standard maintenance brochure for the Goyen EMP7."
I confirmed during this inspection that operators check the alarm conductivity readout as part of their
rounds(as often as every 15 minutes,per boiler operator). The keep daily records of their
observations. They are supposed to alert Brian Conner of any alarms. Environmental personnel also .
check the panel periodically for alarm data(as a backup).
A copy of the brochure and operator's manual (which includes a maintenance section) is in the
WaRO file and is saved to SharePoint under the\Bertie08\00081\aq 16s\2017-2018 Inspection\
directory.
• Please consider the following sections on pages 18-19 of the EPA guidance, and provide any
associated information, or an explanation as to why it is not conducted:
- Monthly response testing
- Drift Checks
- QA procedures for inspection and cleaning
- Annual instrumental set up (calibration)
- Recordkeeping of all of the above
Mr.Findley said that they do not perform regular response testing,drift checks or cleaning activities.
The probe is pulled out for inspection only prior to stack testing. There is high air flow in the stack,
and he's never seen any accumulation on the probe.
The Goyen operator's manual contains the following maintenance/inspection recommendations:
"It is desirable to periodically remove, inspect and clean the inserted parks of the Active Mead.
Maintenance frequency should be determined based on the material characteristics. After initial
installation the Active Flead should be removed and inspected weekly then monthly to gain a working
knowledge of'the interval that would be required for maintenance.
Built-in self-check circuitry is enabled every time the unit is powered up, ensuring the units-
integrity, It is recommended that this is done at least once a month for installations that only require
plant condition monitoring, systems that are calibrated should conduct this at an interval not
exceeding once a week. The procedure for this test is to remove the sensing head from the stack,
disconnect the probe shaft,power down for a period of 30 seconds to enable the built in self-check:
While the device is conducting the internal se f check the mA output will be at.Snz'I, this will last
about 30 seconds. If the result of the internal self=check,fails then the instrument will fail to 3.8mA
and hold this value until rectified.
As these devices are typically mounted outdoors inspection of cabling, moisture ingress and
general condition of the monitor is vital.
Things to lookfor are.•
•Moisture ingress into the body of the monitor.
• Worn offrayed cabling.
•Any burn marks on termination.
• Tightening of all screws (terminals, lid,probe shaft etc)
•Particulate build-up on the sensor front end and probe shaft
Since the 2017/2018 inspection and above correspondence,Avoca has been pulling the probe on a
monthly basis to inspect per the above manufacturer recommendation. E&I staff do the inspections
and create PMS records to document. During the inspections they do a check to ensure the probe can
(SharePoint\Facilities\Bertie08\00044\agl6s\2018-2019 Inspection\20190130al6.doc) Page 29

zero out. Their inspection procedure includes disconnecting, cleaning,re-powering, observation of
the self-check per Goyen's specifications.
• The bag leak detection system certified by the manufacturer to be capable of detecting PM emissions
at concentrations of 10 mg/m3 or less.
Mr. Findley confirmed that the BLDS measures at concentrations as low as 0.01 mg/m3. The
manufacturer's spec sheet is in the WaRO file and is saved to SharePoint.
Corrective action must begin within 1 hour of a bag leak detection system alert. The bagfilter must be
maintained so that periods which would cause an alert are no more than 5 percent of the operating time
during a 6-month period. Avoca is required to keep records of the date,time,and duration of each alert,
the time corrective action was initiated and completed, and a brief description of the cause of the alert and
the corrective action taken. In calculating this operating time percentage, if inspection of the fabric filter
demonstrates that no corrective action is required,no alert time is counted.If corrective action is required,
each alert shall be counted as a minimum of 1 hour. If it takes longer than 1 hour to initiate corrective
action,the alert time shall be counted as the actual amount of time taken to initiate corrective action.
Steam Production and Oxygen Monitoring
During the inspection on 1/29/2019 the following boiler operating data was observed:
02% CO ppm
South Boiler BB 1 16.3 50
North Boiler BB2 11.2 59
Combined steam was 14,870 lbs/hr. The CO monitor is a maintenance monitor.
To comply with the CO emission limitation,Avoca operates an oxygen trim system. The trim system is
made by B3 Systems, Model-Maple Systems HM15070TH. The 02%must be set no lower than the
lowest hourly average oxygen concentration measured during the most recent CO performance test. The
12/14/2016 stack test established new minimums of 10.5% for both of the boilers. It is my understanding
that the boilers must meet these minimums on a rolling 30-day average basis,per revision to the Boiler
MACT on 1/31/2013.
The increment of measure does not appear to be specified in the regulation. Mr. White corrected my
understanding of data averaging for the oxygen monitors. Avoca's system averages oxygen every 5
minutes. It then averages to daily and rolling 30-days. For CY2018 the 30-day rolling average records
indicate no instances where the rolling average went below the minimum percentages.
The regulation requires the facility to maintain the operating load of each unit such that it does not exceed
110 percent of the highest hourly average operating load recorded during the most recent performance
test. Avoca monitors steam load and tracks the 30-day rolling average steam production. It is my
understanding that the 30-day average isn't explicitly established in the rule, but it is the generally
understood among industry to apply it. The system records 5-minute averages, and then averages them
for each day and 30-days. Avoca back calculates daily and monthly wood fuel feed tonnages based on
the steam load. Avoca established a maximum of 25,735 lbs/hr from the 12/14/2016 stack test. For
CY2018 the 30-day rolling average records indicate no instances where the rolling average went above
this limit.
40CFR 63.7505 requires Avoca to have a site specific monitoring plan for the oxygen analyzer and the
steam monitor. The monitoring systems have a manufacturer's manual that operators follow. Avoca has
written a basic boiler SSM plan that discusses how the monitors are to operate and what to do when limits
are exceeded. Copies of each document are saved to the WaRO LAN at\Bertie08\00044\AO16s\2017-
2018 Inspeetion\Nanodac IUser Guide v9.1)df and Biomass Boiler SSM Plan.
(SharePoint\Facilities\Bertie08\00044\agl6s\2018-2019 Inspection\20190130a16doc) Page 30

The oxygen analyzers and steam meter are checked for accuracy quarterly(most recent 11/06/2018). The
oxygen meters are recalibrated if needed. A probe is used in the stack to verify oxygen and stack
temperature. The steam meter is verified daily with output to the plant. The steam meter is factory
calibrated, so must be replaced as needed instead of calibrated. The records are attached to the file copy
of this report and are saved to/Bettie08/00044/aq l6s/2018-2019 hrspection/2019 Biomass Boiler Steam
Meter Calibration and Maintenauco xlsx .
The oxygen and steam data are saved to the control panel, but are not accessible for viewing without
downloading the data to a thumb drive.
Notifications/Reports
The Initial Notification was submitted 5/28/2013.
The Notification of Compliance Status was submitted to DAQ on 2/17/2014 within 60 days of the
12/19/2013 test.
Avoca must submit semi-annual reports for Subpart DDDDD. The reports include a listing of current
emissions and operating limits, operating times for the period, dates of BLDS certification, list of new
fuels, any malfunctions and deviations(dates,times, descriptions). Avoca includes all oxygen and steam
30-day rolling averages in the semi-annual reports,as well as analyzer downtime. I have reviewed
Avoca's semi-annual MACT reports for CY2018. They were submitted on time and appear to be
complete.
I believe Avoca is performing adequate inspection,maintenance and calibration activity on the bag leak
detection system. Improvements have been made to satisfy EPA guidance and the Goyen(manufacturer)
operation manual. Avoca appears to be currently in compliance with Subpart DDDDD.
15A NCAC 2Q.0317 "Avoidance Conditions"
2D.0530 "Prevention of Significant Deterioration"
Permit Condition 2.I.E.7. (formerly condition 2.2.D.1.)
The dryer is limited to less than 40 tons of VOC emitted per consecutive 12-month period. The following
must be monitored if sage is sent to the dryer:
• Measure the sage input on a daily basis,
• Measure the hexane content of the sage exiting the process on a weekly basis. The sampling shall
include collecting sage material as it exits the desolventizer before it enters the rotary dryer every
hour for a 24-hour period each week.
• Calculate daily VOC emissions by using the daily sage input and the hexane content
(lb hexane/lb sage) determined weekly for the sage drying until the next weekly result is obtained.
The dryer is no longer operated, and Avoca has submitted and application to remove it from the permit.
Avoca must submit a semi-annual summary report of the monthly VOC emissions for the previous 17
months and the emissions for each of the 12-month periods over the previous 17 months.
The reports have been submitted on time and are complete. The dryer is in compliance with 2D.0530.
2Q. 0317 "Avoidance Conditions,Limitation to Avoid 2D.1112"
112(g) Case by Case Maximum Achievable Control Technology
Permit Condition 2.I.E.8.
In order to avoid applicability of 112(g),n-hexane emissions from the rotary dryer(ID No. ES-RD)shall
be less than 10 tons per year.
To demonstrate compliance,Avoca must monitor the following:
(SharePoint\Facilities\Bertie08\00044\agl6s\2018-2019 Inspection\20190130al6.doe) Page 31

• Quantity of n-hexane in pounds used in the Rotocel operations each month and for the 12-month
period ending on that month.
• Records of purchase orders and invoices of materials containing n-hexane that are used in the Rotocel
operations.
• Utilize the calculations specified in Permit Condition No. 2.1.E.8.b. (measure the hexane content of
the sage exiting the process on a weekly basis by collecting sage material as it exits the desolventizer
before it enters the rotary dryer every hour for a 24-hour period each week. Then calculate daily
VOC emissions by using the daily sage input and the hexane content).
The dryer is no longer operated and Avoca has submitted an application to remove it from the permit.
I'm not sure why Avoca is required to track hexane usage and emissions from the Rotocel operation to
show that the sage dryer emits less than 10 tpy HAP. Avoca does track Rotocel usage and emissions on a
daily basis.
Avoca is required to submit a semi-annual summary report of the monthly n-hexane emissions from the
dryer for the previous 17 months. Emissions for each of the 12-month periods over the previous 17
months must also be reported. Even though there are no emissions to report from the dryer,the permit
required reports have been complete and submitted on time. The dryer is in compliance with 2Q.0317.
FLAKER DECONDITIONER(ID No. ES-DEC-2001)
The flaker deconditioner is planned for use with the chick pea process at Biomass. It has not started
operation yet.
Regulated Eollutant Limits/Standards Applicable Regulation
Particulate matter Flaker deconditioner: E=4.1 OP" 15A NCAC 02D .0515
Visible emissions 20 percent opacity 15A NCAC 02D .0521(d)
21).0515 "Particulates From Miscellaneous Industrial Processes"
Permit Condition 2.I.J.I.
The deconditioner emissions are controlled by cyclone. The permit condition requires monthly visual
external inspection of the cyclone and an annual inspection of structural integrity. The results of
inspection must be recorded in a logbook. Submittal of a semi-annual summary report of the inspection
activities is required.
2D.0521 "Control of Visible Emissions"
Permit Condition 2.1.J.2.
The deconditioner has an opacity limit of 20%. Avoca is required to observe the deconditioner opacity
once per month and record observations in a logbook. Semi-annual summary reports are required.
INSIGNIFICANT ACTIVITIES
Emission Sonrce Control Device Control Device
Emission Source Aescription'
;1D No, ID No. Description
Wastewater treatment plant equalization tank No. 1
IWWTP-ETl (65,500 gallon capacity)
N/A N/A
Wastewater treatment plant equalization tank No.2
IWWTP-ET2 (65,500 gallon capacity)
N/A N/A
Wastewater treatment plant aeration tank No.2
IWWTP-AT2 (63,500 gallon capacity) NIA N/A
(SharePoint\Facilities\Bertie08\00044\ag16s\2018-2019 Inspection\2019013006.doc) Page 32

Emission Source Control Device Control Device
ID No, kmissiott Source Description- ID No. Descri loon
IW WTP-AT3 Wastewater treatment plant aeration tank No.3 N/A N/A
(63,500 gallon capacity)
IWWTP-CLR Wastewater treatment plant clarifier N/A N/A
(1,310 gallon capacity)
ITK9238 No.2 fuel oil storage tank(50,000 gallon capacity) N/A N/A
ITK9239 No.2 fuel oil storage tank(50,000 gallon capacity) N/A N/A
ITK102 No.2 fuel oil storage tank(495 gallon capacity) N/A N/A
for generator
ITK103 No.2 fuel oil storage tank(495 gallon capacity) N/A N/A
for generator
ITKFP No.2 fuel oil storage tank(270 gallon capacity) N/A N/A
for generator
One water spray fame
CD-7-9215 scrubber(0.5 gals/min
ECS process:batch preparation of ethylenediamine/ minimum water injection
copper sulfate solution rate)venting to
IECS This is located at PNE.Has not operated in
several years. One water spray fume
CD-Z-9216 scrubber(0.5 gals/min
minimum water injection
rate)
IES-PV Propane vaporizer not in use N/A N/A
I-Briquette Sage Briquetting Machine DNE N/A N/A
IIB-5 through Five steam heated hot boxes N/A N/A
FIB-5
I-SFG-PKG SFG packaging area equipped with a bagfilter N/A N/A
I-Biomass-PKG Biomass Bulk bag packaging area ADC-1001 Bagfilter(418 square feet
of filter area)
I-CVS-2002 Flaker truck unloading N/A N/A
I-SH-2001 Flaker storage silo N/A N/A
We performed a walk-by of the insignificant sources and no physical issues were noted.
Avoca has three small cooling towers that are not included in the insignificant list. One is at the Biomass
area, one is at SFG and one is at Rotocel.
We observed a small dust collector at SDE-2. A duct suctions off the dryer to the filter. There is a
collection drum off of the filter. This appears to be an insignificant source.
The flaker equipment are constructed, but have not yet been operated.
MULTI-SOURCE APPLICABLE REGULATIONS AND PERMIT CONDITIONS
2D.1100 "Control of Toxics Air Pollutants"
Permit Condition 2.2.A.I.
The permit condition simply states for any non-NESHAP source, any increase in toxic air pollutants must
be evaluated.Avoca's toxics limitations were removed from permit T40. As allowed under Session Law
2012-91 House Bill 952,the facility is not required to evaluate toxics from MACT sources(the burden
belongs to DAQ). All of the modeled toxics sources are MACT FFFF,DDDDD, 112j or ZZZZ affected.
Changes in toxics emissions due to the modifications made under T40 and T41 were modeled by Avoca
and reviewed by Alexander Zarnowski(5/01/2013)and Tom Anderson(5/16/2013). The DAQ permit
(SharePoint\Facilities\Bertie08\00044\agl6s\2018-2019 Inspection\20190130al6.doc) Page 33

reviews stated that the modifications did not pose unacceptable toxics health risks.Modifications in T42
through T52 did not require re-modeling by DAQ Permits.
Avoca toxics that require review include n-hexane,hexane isomers, chloroform, and ethyl acetate. It is
my understanding combustion toxics from the boilers and engines were modeled based on maximum
operating rate of the sources.
Per the 2017 emissions inventory, 652.19 tons of iso-hexanes(including n-hexane)were emitted
fugitively from Rotocel and Recovery operations. 107.7 tons were emitted from the final process vent
at scrubber CD-1001-2-S-1. 62.3 tons of iso-hexane(including n-hexane)were emitted from SDE-1 and
SDE-2. Note that Alexander's modeling review memo indicates that fugitives were included in his
modeling review.
Application 16A was submitted 1/20/2016 is for expansion of SDE,which could double the n-
hexane and iso-hexane emissions from that area. Avoca performed a netting analysis of hexane
emissions,taking into account the SDE expansion and switch to heptane on SFG. Emissions of n-
hexane and iso-hexane used in the 2013 air modeling were compared with potential emissions from
this modification. The results of the emission comparison are shown in the table below(approved by
Betty Gatano in her T46 permit review). The overall potential emissions of n-hexane decreased by
19.3% and emissions of iso-hexane emissions decreased by 2.4%when compared with the emissions
used in the 2013 modeling. Therefore,the addition of the new SDE-2 operations was determined to not
present an unacceptable risk to human health.
Comparison of Facility-Wide Emissions
Pollutant Averaging Emissions Used in 2013 Air Emissions after Modifications
Period Modelin
Facility-wide % of AAL Facility-wide % Decrease
emissions (lb/hr) Emissions lb/hr in Emissions
n-hexane 24-hour 117.4 92% 94.7 19.3%
Iso-hexane 1-hour 1707.3 19% 1666.8 2.4%
Notes:
• Potential emissions of n-hexane from the new SDE-2 operations are estimated as 32.4 lb/hr,and potential
emissions of iso-hexane were estimated as 48.5 lb/hr. The emissions take into account the control efficiency of the
condenser and mineral oil scrubber used on process vents in the SDE-2 operations.
• Emissions of n-hexane from the SFG operations used in the 2013 modeling were 11.6 Ib/hr. Emissions of iso-
hexane were 421.5 lb/hr. Only heptane is now used in the SFG operations,and n-hexane and iso-hexane are no
longer emitted from the SFG operations.
• For the existing Rotocel/Recovery Operations, emissions of n-hexane will decrease by 43.2 lb/hr due to a change
in the n—hexane to iso-hexane ratio used in the Rotocel/Recovery operations. Emissions of iso-hexane will
increase by 334.0 pounds/hr when compared to the emissions used in the 2013 air modeling.
•Emissions of n-hexane and iso-hexane from all other sources at Avoca remain the same as used in the 2013
modeling.
2D.1806 "Control and Prohibition of Odorous Emissions"
Permit Condition 2.2.A.2.
Avoca may not operate the facility without implementing management practices or installing and
operating odor control equipment sufficient to prevent odorous emissions from the facility from causing
or contributing to objectionable odors beyond the facility's boundary. I did not smell any odors outside of
the property boundaries. There was a slight hexane odor near the Rotocel and Recovery sources.
There are no odor complaints for Avoca in IBEAM Complaints database. Mr.White said they have
received no odor complaints since the date of the last DAQ inspection.
(SharePoint\Facilities\Bertie08\00044\agl6s\2018-2019 Inspection\20190130a16.doc) Page 34

2D.0530 "Prevention of Significant Deterioration"
Permit Condition 2.2.B.
The following Best Available Control Technology(BACT) limits shall not be exceeded:
Emission Source Pollutuut BACT Emission,Limits Control Technology
Rotocel extractor, 47.31 tons per Condenser CD-31209 and
desolventizer, and solvent 10.8 pounds consecutive
separation/recovery VOC per hour 12-month packed tower scrubber
(ID No. ES-1001-2-1-P1) period CD-1001-2-5-1
Rotocel equipment leaks VOC N/A Leak detection and repair
(ID No. ES-1001-2-1-F) (Fugitive) (LDAR)
Rotocel wastewater stream Fixed roofs on
(1D No. ES-1001-2-1-WW) VOC N/A wastewater treatment
tanks
When Rotocel is Operating: Condenser CD-1001-1-3
0.80 pounds per hour and 3.50 and packed tower
tons per consecutive 12-month scrubber CD-1001-2-5-1
period
Recovery Arcon tank M-1 When Rotocel is NOT Condenser CD-1001-1-3
(ID No. ES-1001-1-1-P1) VOC Operating and Recovery
Process is processing
Concrete:
8.76 pounds per hour and 0.63
tons per consecutive 12-month
period
When Rotocel is Operating: Condenser CD-1001-1-
0.85 pounds per hour and 3.72 T5B and packed tower
tons per consecutive 12-month scrubber CD-1001-2-5-1
period
Recovery stripper T-5 and When Rotocel is NOT Condenser CD-100 1-I-
receiver M-21 VOC(ID No. ES-1001-1-1-P2) Operating and Recovery TSB
Process is processing
Concrete:
4.89 pounds per hour and 1.99
tons per consecutive 12-month
period
Recovery process/storage
tanks VOC N/A Fixed roofs
(ID No. ES-1001-1-1-P3)
Recovery equipment leaks VOC N/A LDAR
(ID No. ES-1001-1-1-F) (Fugitive)
95%mass removal from Fixed roofs on
Recovery wastewater stream VOC wastewater stream consisting wastewater treatment
(ID No. ES-1001-1-1-WW) of methanol-wash tanks and biological
treatment
(SharePoint\Facilities\Bertie08\00044\agl6s\2018-2019 Inspection\2019013006.doc) Page 35

Emission Source Pollutant BACT Emission!Limits Control Technology
Botanical extraction
immersion extractor, 61.76 tons per
desolventizer, first and 14.1 pounds consecutive Condenser CD-1001-11-
second stage evaporators, VOC EX1002 and condenser
distillation column, day tank, per hour 12-month CD-1001-11-EX1003
and multiple process tanks period
(ID No. ES-1001-11-P)
Botanical extraction
equipment leaks (ID No. ES- VOC N/A LDAR
1001-11-F) (Fugitive)
Botanical extraction Fixed roofs on
wastewater stream(ID No. VOC N/A wastewater treatment
ES-1001-11-WW) tanks
Biomass extraction
immersion extractor,
desolventizer, day tank, iso- 61.8 tons per Condenser CD-1004-
hexane storage tank, first and VOCs 14.1 pounds consecutive 2EX1002 and condenser
second stage evaporators, per hour 12-month CD-1004-2EX1003
distillation column, and period CD-
multiple process tanks (ID
No. ES-1004-2-P)
Biomass extraction
equipment leaks(ID No. ES- VOCs N/A MAR
1004-2-F)
Biomass extraction Fixed roofs on
wastewater stream(ID No. V OCs N/A wastewater treatment
ES-1004-2-WW) tanks
No. 2 fuel-oil fired boilers
(ID Nos. H-101, H-102, VOC 0.2 pounds per 1,000 gallons Combustion control
H-103)
Testing
Condition 2.2.B.Lb gives the state power to require emissions testing against the BACT limits, and
requires any testing to be in accordance with 15A NCAC 2D .2601 and General Condition JJ found in
Section 3 of the permit. DAQ has not required any testing.
Monitoring/Recordkeepin Reportine
• The recycle tanks may not be filled by truck unless the Rotocel is in operation. These tanks are now
being used as holding tanks. No fresh hexane is ever added. Avoca only fills delivery hexane tanks
at Rotocel.
• The Permittee shall limit the operation of the recovery stripper T-5 and receiver M-21 (ID No. ES-
1001-1-1-P2)while the extraction(Rotocel)process (ES-1001-2-1-P1) is not operating to no greater
than 34 days in any consecutive 12-month period. The only time this would happen is when Avoca
starts up Concrete operations. They have not employed the Concrete operation since the date of last
DAQ inspection.
(ShaiePoint\Facilities\Bertie08\00044\ag16s\2018-2019 Inspection\20190130a16.doc) Page 36

The permit requires Avoca to record the hours when Rotocel isn't operating. Mr. Conner keeps a log
of gallons delivered to the recycle tanks and confirmation that Rotocel sources and the scrubber are in
operation. I reviewed the records for CY2018-January 2019 and did not see any issues.
• The total number of days during which the liquid flow into the recovery Arcon tank M-1 (ID No. ES-
1001-1-1-P1)exceeds the liquid flow out of the Arcon tank(i.e., days when the liquid level in the
tank rises)while Rotocel is not operating is limited to no greater than 6 days in any consecutive 12-
month period. The only time this happens is when Avoca starts up the Concrete operation. The
permit requires Avoca to record the hours of flow difference when Rotocel isn't operating. The
Concrete process has not operated since the date of last DAQ inspection.
• Avoca is required to implement a leak detection and repair(LDAR) program for the Rotocel,
Recovery,Botanical and Biomass irrespective of whether there is the presence of HAP. The
BACT is established for VOC emissions. The LDAR requirements in the permit follow 40 CFR Part
63, Subpart LIU. Avoca's summary of Subpart UU requirements are attached to the file copy of the
2014 inspection report.
We did not physically verify all of the LDAR points during the inspection. LDAR sniff tests for
valves,pumps and connectors are usually conducted during the summer around sage harvest. I asked
Colin to put a reminder on his calendar to contact me when he's ready to conduct testing so I may
observe.
Equipment Identification
Brian Conner established the list of LDAR points by visually identifying all affected equipment in
each process. The processes are inspected every year to confirm all points are included. Mill-wide
P&ID diagram updates have been completed. Ashland conducted a PSM audit that only resulted in
finding nomenclature errors on the diagrams.
Avoca is currently working on a project to change all of the LDAR point nomenclature to match
Ashland's specifications. Ashland requires all control devices and sources to be physically labeled
with IDs that match the air permit. The company also requires all process lines to be labeled(water,
steam,vent). I observed that the labeling is in progress throughout the plant.
There are no hard to reach or unsafe LDAR test points.
Instrument and Sensory Monitoring for Leaks
Avoca is required to conduct instrument monitoring per Method 21 of 40 CFR Part 60, appendix A
for
i. Valves in gas/vapor or light liquid service;
ii. Pumps in light liquid service;
iii. Connectors in gas/vapor or light liquid service; and
iv. Pressure relief devices in gas/vapor service.
The instrument detector type is not specified in Method 21,but it must meet the specifications and
performance criteria contained in Section 6.0 of Method 21. To qualify the detector must:
• respond to the compound being tested.
• be capable of measuring the leak definition concentration specified in the regulation.
• be readable to f2.5 percent of the specified leak definition concentration.
• be equipped with an electrically driven pump to ensure that a sample is provided to the detector at
a constant flow rate.
(SharePoint\Facilities\Bertie08\00044\ag16s\2018-2019 Inspection\20190130al6.doc) Page 37

• be equipped with a probe or probe extension or sampling not to exceed 1/4 in outside diameter,
with a single end opening for admission of sample.
• be intrinsically safe for operation in explosive atmospheres as defined by the National Electrical
Code by the National Fire Prevention Association.
Brian Conner purchased a Mini-Rae detector in July 2016. It does appear to satisfy the above
requirements. It is essentially a LEL meter-industry standard for LDAR. It did not come with an
extended probe, but Colin White confirmed he is able to comfortably read all of the test points with
the existing probe. During the CY2016 DAQ permit compliance inspection they demonstrated how
the detector operates. The probe appeared to meet size requirements. The meter is designed for an
internal pump to sample at a constant flow rate.
The regulation requires the instrument to calibrated annually. They send the detector to Geotech once
per year to clean and calibrate. The detector is calibrated to measure hexane(it's my understanding
that includes background, so no background adjustment). There is a pre-programmed conversion
factor for isobutylene to hexane. A copy of the most recent calibration record(12/26/2018)is
attached to the file copy of this report. From the record, it appears they have calibrated it with a span
of 100 ppm. It had previously concerned me that this span may not be adequate for detection of leaks
up to the reportable thresholds in the regulation(i.e. 500 ppm and 1,000 ppm). The meter is actually
capable of detecting up to 15,000 ppm.Avoca provided a link to a spec sheet for the detector
(hltp_/iwylw,gec techenv_com/pdf/air_dua1ity/miuirae_ppt7rae3OOOnpd). On the second page in the
top right corner,there is a section titled "Sensor Specifications for VOCs",which describes the
accuracy of the meter. Geotech told Mr. White that a 100 ppm cal gas is based off the manufacturer's
recommendation. When the meter is calibrated to 100 ppm span gas, it will be accurate within 0.1
ppm when reading up to I000ppm. Readings over 1000 ppm when calibrated to 100 ppm span gas
the meter is within Ippm accuracy.
Avoca also has a 100 ppm hexane calibration kit that Mr. White uses prior to beginning any LDAR
sampling. For a non-hexane solvent in Botanical or Biomass,Avoca applies correction factors that
are provided by the manufacturer.
Leaks must be repaired within 15 calendar days of discovery. The first attempt at repair must be
made within 5 days of discovery. Avoca must record the dates of first attempt to repair a leak,the
date of successful repair, and the maximum instrument reading immediately after repair of a leak.
Any equipment shutdowns between leak discovery and repair must be recorded. Avoca's policy is to
fix any leak immediately. Mr. Conner or Mr. White performs daily walktbroughs of the facility,and
operators perform hourly checks on all processes. If anyone senses something may be leaking,they
will perform immediate repair(these leaks are recorded in maintenance logs),or first check for a leak
with the detector.
There are fixed process hexane LEL meters installed(and operating 24/7)in the production buildings
that are also helpful for leak detection.
Per Specific Condition 2.2.13.2.1r.ii.E.I. Avoca has created a delay of repair procedure for Rotocel
and Recovery. A copy of the procedure is attached to the file copy of the 2014 inspection report.
They have not had to employ this procedure.
The sections below are associated with Avoca's LDAR Inspection Summary Table,which is provided
to WaRO semi-annually as part of Subpart UU and MON MACT reporting. It documents the most
recent LDAR testing dates,results, and next test dates.
The LDAR requirements in Condition 2.2.13. for Subpart UU are referenced in Permit
Condition 2.2.C. as applicable to MON NESHAP affected processes (i.e.those that meet the in-
service HAP definitions). Therefore,the SDE-1, SDE-2 and EVG processes are also evaluated
in the LDAR subsections below.
(SharePoint\Facilities\Bertie08\00044\agl6s\2018-2019 Inspection\20190130a16.doc) Page 38

Valves in Gas/Vapor/Light Liquid Service
A valve leak is any reading>500 ppm. All valves were tested in 2016 for the Botanical,Rotocel and
Recovery, SDE-1 and EVG processes. The number of leaks constituted less than 0.25%of the total
valves,which means they may test at a frequency of every two years. Botanical valves were tested on
8/22/2018. Rotocel and Recovery valves were tested 5/30/2018 (copies are attached to the file copy
of this report). The most recent leak testing at SDE-1 is still 6/22/2016 because it has not been in
operation since SDE-2 startup. To be compliant leak testing must be completed upon the next startup
in HAP service. EVG was tested 4/02/2018.
SDE-2 was tested on 5/15/2017. Biomass was last tested on 11/06/2017.
All of the most recent valve leak tests resulted in less than 0.25% of the valves with leaks<500 ppm.
The Method requires repaired valve leaks to be monitored at least once within the first three months
after its repair.
Pumps in Light Liquid Service
Avoca must use the detector to monitor pumps on a monthly basis. A pump leak is any reading>
1,000 ppm. Repair is not required unless the reading is>2,000 ppm.
I reviewed the monthly pump LDAR records for CY2018 through January 2019. There were no
missing records, and they appeared to be complete. Botanical ran in December and August of 2018,
and pump LDAR tests were done during operation. Biomass operated in February and June 2018,
and pump LDAR tests were completed while operating. SDE-1 has not operated since SDE-2 startup,
so the last pump LDAR was completed 7/20/2017.
The December 2018 and January 2019 pump LDAR for Rotocel, and the January 2019 pump LDAR
for Recovery are attached to the file copy of this report.
If, when calculated on a 6-month rolling average for the percent leaking pumps, at least the greater of
either 10% of the pumps in a process unit or three pumps in a process unit leak, Avoca must
implement a quality improvement program for pumps. Avoca has never triggered the QIP
requirement.
Each pump must be checked by visual inspection each calendar week for indications of liquids
dripping from the pump seal. Avoca is required to document each inspection.If there are indications
of liquids dripping from the pump seal at the time inspection,Avoca must monitor the pump. If the
instrument reading indicates a reading of>2,000 ppm it shall be repaired within 15 days(first
attempt within 5 days) or eliminate the visual indications of liquids dripping.
Brian and Colin do the weekly pump leak inspections. The inspection checklist is set up to cover all
areas of the plant. An example 1/23/2019 inspection is attached to the file copy of this report. I
reviewed all weekly inspections for CY2018 through January 2019. All of the records appeared to be
complete. The gaps between inspection dates appeared to be consistent.
Avoca operators perform hourly visuals of the pumps. There is a check point on their operator logs.
The operators record any repairs on the daily log. If maintenance is called, a work order records the
repair.
Connectors in Gas and Vapor Service and in Light Liquid Service
A connector leak is any reading>500 ppm. The Method requires repaired connectors to be monitored
at least once within the first three months after its repair. Avoca's previous sniff tests on all
connectors at the facility resulted in less than 0.25% leaks,which means they may test 50% at a
(SharePoint\Facilities\Bertie08\00044\ag16s\2018-2019 Inspection\20190130al6.doc) Page 39

frequency of every four years (100%within 8 years). The Botanical, Rotocel and Recovery, SDE-1
and EVG processes were tested in 2016. Biomass was tested November 2015, but operated in
November 2017;therefore,LDAR was conducted on 11/06/2017 during operation(no leaks
detected). SDE-2 was tested on 5/16/2017.
Pressure Relief Devices in Gas and Vapor Service
A pressure relief loss is any reading>500 ppm. After each pressure release the pressure relief device
shall be returned to a condition indicated by an instrument reading of less than 500 ppm within five
calendar days after release. The device must be monitored no later than five calendar days after the
restoration to confirm<_500 ppm above background. Any pressure relief device that is equipped with
a rupture disk upstream of the pressure relief device is exempt from monitoring,provided Avoca
installs a replacement rupture disk upstream of the pressure relief device no later than five calendar
days after each pressure release. The most recent pressure release I've seen in the records occurred in
2010 at Rotocel and Recovery. Operations staff are required to notify Mr. Conner of any rupture.
Summary of LDAR Recordkeeping Requirements
i. Either tag equipment or keep equipment identification through written documentation(Avoca
has P&IDs and checklists). Avoca has metal tags on standby for use.
ii.. Identify and keep a written plan for any equipment that is designated as unsafe-or difficult-to-
monitor. Avoca has no unsafe or difficult points.
iii. Keep records for leak repair and records for delay of repair.
iv. For valves, maintain the monitoring schedule for each process unit.
V. For pumps,maintain documentation of pump visual inspections.
vi. For connectors,maintain the monitoring schedule for each process.
vii. Keep records of the dates and results of monitoring following a pressure release.
viii. For a pump QIP program,the Permittee shall maintain individual pump records, QIP
documentation and QIP records(Avoca has not triggered QIP).
Based on my records review during the inspection,Avoca appears to be appropriately keeping all of
the required LDAR Subpart UU records.
LDAR Reporting Requirements
Avoca is required to submit a semi-annual summary report of monitoring and recordkeeping
activities. The report shall address in summary format for valves,connectors,pumps and relief
devices the number of components for which leaks were detected,the percent leakers and the total
number of components monitored. The report shall also include the number of leaking components
that were not repaired, and for valves and connectors, identify the number of components that are
determined to be non-repairable. Where any delay of repair is utilized,report that delay of repair has
occurred and report the number of instances of delay of repair. A copy of the most recent report is
attached to the file copy of this report.
The reports have been complete and have been submitted on time. I have reviewed the reports
submitted through January 2019. No problems have been discovered on the reports.
• Avoca must determine the mass removal efficiency of the on-site biological wastewater treatment
plant(i.e. Wastewater Treatment Plant Aeration Tank No. 1; ID No. W WTP-ATI) for volatile
organic compounds (VOC)on a weekly basis, when the wastewater stream consisting of methanol-
wash from the Recovery operations (ID No. ES-1001-1-1-WW) is discharged to it. The mass
removal efficiency must be more than 95%.
Avoca's wastewater treatment plant is a thermophylic (hot) system. Any VOCs that volatilize out of
the wastewater are emitted in the first aeration tank and retention basin. Avoca calculates the removal
(SharePoint\Facilities\Bertie08\00044\agl6s\2018-2019 Inspection\20190130a16.doc) Page 40

efficiency every day. The Botanical, Sclaerolide, and Recovery operations connect to one wastewater
line with a flowmeter on it. Rotocel has its own line and flowmeter. They use the WATER9 model
to calculate the removal efficiency. They sample the inlet to the aeration tank for methanol. They
also measure daily flow through the tank. The daily records for CY2018 were reviewed. No daily
efficiencies were below 95%.
Avoca must perform periodic I&M on the condensers as recommended by the equipment
manufacturer. In addition,Avoca must perform an annual inspection of each condenser system,
including the following:
i. inspect and maintain the structural integrity of each condenser,including inspection for leakage
of coolant and, if the system is under positive gauge pressure,leakage of the contaminated gas
stream. In order to monitor leakage of the coolant,the condensate shall be inspected for the
presence of coolant.
ii. inspect and maintain the structural integrity of ductwork and piping leading to and coming from
each condenser.
I reviewed the inspection records for CY2018. For the cryogenic condensers,Brian conducts monthly
external inspections(looking for leaks) and annual internal inspections. The Botanical condensers
were most recently structurally inspected per the permit condition on 11/07/2018 in preparation for
the trial. The Biomass condensers were most recently inspected on 6/22/2108 prior to chick pea
operation. Mr. Conner also conducts monthly external inspections even when the condensers are not
in operation. He performed inspections 3/15/2018 and 7/12/2018 during chick pea processing,and
12/18/2018 while the trial was running. CD-31209, CD1001-1-3 and CD10001-1-T5B were
inspected quarterly.
• Avoca must install, maintain, operate, and calibrate, in accordance with manufacturer's
recommendations, a sensor to continuously measure the outlet temperature of each condenser. The
temperatures output from the sensors must be continuously monitored and hourly values used to
determine the 24-hour average temperature at the condenser outlets. The outlet temperatures of the
condensers must be below the limits in the table below.
Emtssign Source Condenser ID No. Required Outlet Temperature'
Recovery Arcon tank M-1 CD-1001-1-3 45 °F, 24-hour average,when
(ID No.ES-1001-1-1-P1) source ES-1001-2-1-P1 is not
Recovery stripper T-5 and operating and the tanks are used
receiver M-21 (ID No. ES-1001- CD-1001-1-T5B for concrete processing in the
1-1-P2) recovery operation
-40 T, 24-hour average for non
Botanical extraction operations CD-1001-11-EX1003 water soluble solvents, and
(ID No. ES-1001-11-P) 17 T, 24-hour average for water
soluble solvents
-40°F, 24-hour average for non
Biomass extraction operations CD-1004-2EX1003 water soluble solvents,and
(ID No. ES-1004-2-P) 17°F, 24-hour average for water
soluble solvents
I have interpreted the condition as only applying to the condensers listed in the above table.
Concrete has not been operated for several years, so no temperature monitoring was necessary on the
Arcon tank and stripper/receiver condensers. There are two condensers per system at Botanical and
Biomass, so one can operate while the other is thawed if hexane(non-water soluble) is used as the
(SharePoint\Facilities\Bertie08\00044\agl6s\2018-2019 Inspection\2019013 Oat 6.doc) Page 41

solvent. I reviewed the hourly measurements for both sets of Botanical and Biomass cryogenic
condensers during the few weeks they operated in CY2018. No 24-hr averages exceeded 17°F.
The temperature gauges are checked every quarter, and if they are not properly operating,they are
replaced.
Avoca has a startup, shutdown and malfunction plan for the cryogenic condensers. A copy is
attached to the file copy of the DAQ 2014 inspection report.
Note that there is a 12-hour rolling average temperature measurement requirement for the cryogenic
condensers under CAM permit conditions as well.
• Avoca must perform periodic inspections and maintenance on the Rotocel and Recovery scrubber as
recommended by the equipment manufacturer. In addition to the manufacturer's inspection and
maintenance recommendations, or if there is no manufacturer's inspection and maintenance
recommendations, as a minimum,the following must be performed:
i. Annual inspection of spray nozzles and packing materials, chemical feed system(if so equipped),
and perform maintenance and repair when necessary to assure proper operation of the packed
tower scrubber;
ii. Annual inspection, cleaning, and calibration of all associated instrumentation.
iii. Additionally, whenever the packing is replaced, inspect for nozzle plugging and settling of the
packing.
The scrubber undergoes an annual teardown and full inspection. This was done most recently on
5/23/2018 prior to sage harvest. The packing was replaced. The mineral oil is changed out with the
annual internal inspection. The scrubber is also inspected monthly. The operators have an hourly
checklist for visual inspection of the scrubber.
Avoca has a startup, shutdown and malfunction plan for the mineral oil scrubber. A copy of the plan
is attached to the file copy of the DAQ 2014 inspection report.
• Install,maintain,operate, and calibrate a scrubbing liquid flow meter,a scrubbing liquid inlet
temperature sensor,and an emission stream inlet temperature sensor for packed tower scrubber CD-
1001-2-S-1.
i. The scrubbing liquid injection rate must be maintained at or above 8 gpm;
ii. The scrubbing liquid inlet temperature must be<_105 °F;
iii. The emission stream inlet temperature must be<90 °F;
iv. The flow meter and temperature sensors must continuously operate.
The above must be measured and recorded once a day, if Rotocel and Recovery are in operation.
I have reviewed the daily records for CY2018 through January 2019, and the scrubber appeared to
meet the limitations. The daily records are included in Avoca's semi-annual reports.
The temperature sensor is replaced annually with the teardown. The flowmeters are calibrated
quarterly.
• Record V OC emissions for each calendar month, and for the consecutive 12-month period ending
with each calendar month for Rotocel and Recovery operations, Botanical Extraction operations, and
Biomass Extraction operations.
It is my understanding that the hourly VOC emissions limits in this permit condition are averages
calculated by dividing the monthly process total by the hours of operation for the month.
V OC emissions are primarily determined through purchasing records,daily measurement of solvent
recovery(tank levels) inventory, and tracking batches/day (formulation). Avoca determines
(SharePoint\Facilities\Bertie08\00044\ag16s\2018-2019 Inspection\20190130al6doc) Page 42

emissions on a daily basis. I reviewed these daily records. Brian Conner keeps a worksheet that
totalizes each area VOC emissions by the month, including fugitives.There is another summary
spreadsheet that is submitted to DAQ semi-annually that satisfies the 12-month rolling average
reporting requirement for this permit condition.
Avoca determines Rotocel overall emissions (point and fugitive)by taking daily inventory of the
gallons hexane used(taking into account any shipments of hexane)and subtracting out the amount of
hexane that was recovered(which is accomplished by daily measurement of the change in volume of
the recovery tanks). Avoca receives new rail shipments of hexane about 3 times a month. Each load
is about 6,500 gallons.
Historically I was informed by RCO Permits that fugitives are not included in the limits and
recordkeeping required by PSD Permit Condition 2.2.B.I. Brian Conner is of the same
understanding. I had recommended to RCO Permits that DAQ revisit the original permit application
when completing the permit renewal review to make sure fugitives did not need to be included in that
PSD limits. Kevin Godwin's T47 permit renewal review states that fugitives were considered in the
PSD applicability determinations made for this permit. However,there were no modifications made
to the permit condition limits.
Per Avoca's 2017 emissions inventory, about 652.19 tons of iso-hexanes(including n-hexane)were
emitted from Rotocel and Recovery (10% increase from 2016). Of that total, 107.7 tons were emitted
from the final process vent at scrubber CD-1001-2-S-1,Arcon tank and Recovery T-5 stripper,
compared to 92.2 tons in 2016. Fugitive loss of hexane occurs when the sage is exiting the
desolventizer. There is still a significant amount of hexanes left in the sage and it flashes when the
desolventizer is opened to remove the sage.
Methanol is also emitted from Rotocel and Recovery processes, so the CY2017 total VOC (point and
fugitive) emitted from Rotocel and Recovery was 1,050.85 tons (10.7% increase from CY2016).
The CY2017 total VOC from Botanical was 0.06 ton(including fugitives). The CY2017 total VOC
for Biomass was 0.01 ton.
I have reviewed the VOC hourly averages and rolling 12-month emissions for 2017 and 2018 against
the limits(restated below-concrete limits aren't included because it hasn't operated in several years)
using the January 2019 semi-annual report. The values stated in bold in the table are the highest 12-
month rolling totals and hourly averages during each of those years. Again, fugitives are not
included in the totals provided below, so the emission rates are only from point sources.
Emission Source Pollutant $,4GT Eud§slon Lions Cogt)rol Techn4lagy
10.8 Ibs/hr 47.31 tons per
Rotocel extractor, consecutive 12-
(CY2017-5.3 Condenser CD-31209 and
desolventizer,and solvent month CD-
separation/recovery VOC Ibs/hr, period packed tower scrubber CD-
se
P ry CY2018-5.3 (CY2017-29.29 1001-2-5-1
(ID No. ES-1001-2-1-P1) Ibs/hr) tpy, CY2018-
22.13 tpy)
When Rotocel is Operating: Condenser CD-1001-1-3
E(IDecON:voo.
Arcon tank M-1 0.80 Ib/hr and 3.50 tons per and packed tower scrubber
S-1001-1-1-P1) VOC consecutive 12-month period CD-1001-2-S-1
(CY2017-0.53 lb/hr and 2.54 tpy,
CY2018-0.52 Ib/hr and 2.22 tpy)
(SharePoint\Facilities\Bertie08\00044\agl6s\2018-2019 Inspection\20190130al6.doc) Page 43

Emission Source Pollutant 'BACT Emission Limits Control Technology
When Rotocel is Operating: Condenser CD-1001-1-T5B
Recovery stripper T-5 and 0.85 Ib/hr and 3.72 tons per and packed tower scrubber
receiverM-21 VOC consecutive 12-month period CD-1001-2-S-1
(ID No. ES-1001-1-1-P2) (CY2017-0.4 lb/hr and 2.25 tpy)
(CY2018—0.4 lb/hr and 1.65 tpy)
Recovery process/storage These tanks ultimately tie into the
tanks VOC condenser/packed tower scrubber Fixed roofs
(ID No. ES-1001-1-I-P3) system
Botanical extraction 61.76 tons per
immersion extractor, 14.1 Ibs/hr consecutive 12-
desolventizer, first and (CY2017-0.16 month period Condenser CD-1001-11-
second stage evaporators, VOC lb/hr, (CY2017-0.55 EX1002 and condenser CD-
distillation column, day tank, CY2018-13.64 tpy, CY2018- 1001-11-EX1003
and multiple process tanks Ibs/hr) 3.3 tpy)
(ID No.ES-1001-11-P)
Biomass extraction
immersion extractor, 14.1 61.8 tons per
desolventizer, day tank, iso- 17-0.09 consecutive 12-
(CY2017- Condenser CD-1004-
hexane storage tank, first and
VOCs Ib/hr, month period 2EX1002 and condenser
second stage evaporators, CY2018-0.86 (CY2017-0.05 CD-1004-2EX1003
distillation column, and lb/hr) tpy, CY2018-
multiple process tanks 0.43 tpy)
(ID No. ES-1004-2-P)
No. 2 fuel-oil fired boilers 0.2 pounds per 1,000 gallons (AP-
(ID Nos. H-101,H-102,H- VOC 42) Combustion control
103)
The hourly average emissions and 12-month rolling average appear to be less than the limits. Note that
the reported inventory VOC emissions for CY2017 for botanical and biomass are different from that
provided in Table 13 of the semi-annual reports b/c the inventory value is process and fugitive emissions
based on accounting data, and the other is a calculated point source estimate based on hours operation and
actual production rate.
Outside of the LDAR reporting requirements discussed above,Avoca is required to submit semi-annual
reports of the monthly VOC emissions from Rotocel/Recovery,Botanical and Biomass for each of the
previous 17 months and the cumulative total VOC emissions from each for each of the consecutive 12-
month periods ending during the reporting period. The reports have been submitted on time, and they
have been complete.
There is a concern that the hexane and methanol emissions from tanks M-25,M-36A/B and M-8 (ID No.
ES-1001-1-1-P3),which vent to Arcon Tank M-1, might need to be included with the M-1 tank emissions
in comparison to its BACT limits. Permit condition 2.2.B. BACT limits were created in permit revision
T28. Avoca submitted application 0800044.04A requesting BACT limits for Botanical and Rotocel and
Recovery in place of the existing PSD limit of facility-wide 225 tons VOCs per year. I found the permit
review in the file,but WaRO does not have a copy of the application. The permit review does not
acknowledge any control devices for the process tanks. Therefore, I assume that their emissions do not
need to be included with the Arcon Tank, as the permit engineer assumed the process tank emissions
small enough to declare no controls or limits for BACT.
Avoca appears to be in compliance with PSD permit conditions 2.2.B.1 and 2.2.B.2.
(SharePoint\Facilities\Bertie08\00044\agl6s\2018-2019 Inspection\20190130a16.doc) Page 44

2D.1111 "Maximum Achievable Control Technology"and
40 CFR Part 63,Subpart FFFF"Miscellaneous Organic NESHAP"
Permit Condition 2.2.C.
The permit states that MON MACT applies to the EVG,PNE,Rotocel and Recovery, SDE-1 and SDE-2,
Botanical, and Biomass operations, as well as the Wastewater Treatment Plant Aeration Tank No. 1 and
the hexane storage/recycle tanks. The Concrete process is also included,but Avoca has not operated it in
years. SFG is no longer a MON process since it permanently no longer uses HAPs (heptanes only).
Botanical and Biomass rarely operate with HAP solvents. Mr. Conner and I understand the MON rule
only applies when HAPs solvents are in use.
All of the above operations except SDE-2 and the W WTP are permitted as Group 2 processes. Note that
the BACT limits and monitoring requirements in permit condition 2.23. were established partly because
they were permitted as MON processes. SDE-2 and the Wastewater Treatment Plant Aeration Tank(ID
WWTP-ATI) are considered Group 1 sources. A Group 1 wastewater stream consists of process
wastewater that contains compounds listed in Tables 8 and 9 of the subpart(methanol).
Emissions Limits and Monitoring Requirements
• The Permittee must comply with the applicable control requirements found in 40 CFR 63.2455
through 63.2490 for the affected sources. This is accomplished through compliance with the BACT
requirements for all plant processing areas, which are discussed earlier in this report.
• For the continuous processes with in-HAP service (i.e. Rotocel and Recovery,Botanical,EVG, SDE-
1, SDE-2; Biomass did not use HAPs in 2018) leak testing on the affected sources must be conducted
according to the leak detection and repair(LDAR)program found in Section 2.2 B.2, of the permit,
and discussed earlier in this report.
See the LDAR discussion within the Section 2.2.13.2. regulatory review above. I reviewed the
CY2018 records,and Avoca appears to have performed monthly MON LDAR testing on all pumps
and LDAR leak detection testing on schedule according the requirements of Subpart UU.
• The Permittee does not have to create a SSM Plan for Group 2 emission points. However,there is a
SSM Plan required for wastewater to WWTP-ATI (63.2485 and Table 7)and for SDE-2. All of
Avoca's wastewater is collected and sent to the W WTP, and a copy of the W WTP plan is attached to
the file copy of the 2016 WaRO inspection report. The SSM plan for the SDE-2 Mineral Oil
Scrubber is attached to the file copy of the 2018 WaRO inspection report report and is saved to
\Bertie08\00044\AQ1.6s\20180131 SDE-2 Scrubber SSM Plan doc. The Rotocel and Recovery plant,
as well as the Botanical and Biomass plant have startup, shutdown and malfunction plans as well.
Copies of the plans are attached to the file copy of the 2014 WaRO inspection report.
• Avoca is combining organic HAP emissions from different emission types in the SDE-2 operations
(e.g., storage tanks and batch process vents). In accordance with 40 CFR 63.2450(c)(2)(i), SDE-2
must comply with the requirements for Group 1 batch process vents in Table 2 of 40 CFR Part 63,
Subpart FFFF and 40 CFR 63.2460 for the combined streams, including applicable testing,
monitoring, recordkeeping,and reporting:
i. Table 2 in the regulation requires a reduction of organic HAP emissions from the sum of all vents
by>95%by weight by venting emissions from a sufficient number of the vents through one or
more closed-vent systems to the chilled water condenser(ID No. CD-4002) in series with the
mineral oil scrubber(ID No. CD-4003-S)
ii. An initial performance test and NOCS are required within 150 days of initial startup to
demonstrate SDE-2 meets the reduction minimum.
SDE-2 started up operation on 2/21/2017. On 6/06/2017 Brent Hall (DAQ SSCB) and I observed
EPA Method 18 stack testing for hexane from the SDE-2 condenser ID CD-4002 and mineral oil
(SharePoint\Facilities\Bertie08\00044\agl6s\2018-2019 Inspection\20190130al6.doc) Page 45

scrubber ID CD-4003-S. The test observation report is saved online as
\Bertie08\00044\aq 16s\20170606a 16 and stk.doc. The stack test report and NOCS were
submitted to DAQ on 7/07/2017. The test result was 98.9%removal efficiency. SSCB has
reviewed and approved the test results(memo dated 8/16/2017).
iii. Operating limits for the chilled water condenser and mineral oil scrubber (ID No. CD-4003-5)
must be established with the performance testing. The permit during the inspection quotes the
regulation that Avoca must continuously monitor the mineral oil scrubber temperature and
specific gravity of the mineral oil, and continuously monitor the condenser exit temperature.
Because flow through the chilled water condenser could be intermittent, Avoca must install,
calibrate, and operate a flow indicator at the inlet or outlet of the condenser to identify periods of
no flow. However, Avoca did not believe the parametric monitoring parameters set by the
regulation and the permit were appropriate for their equipment (they couldn't find a specific
gravity monitor in existence that would work for them). A request for alternative monitoring was
submitted to DAQ on 3/13/2017, and approved by SSCB (Don Van der Vaart) by letter to Avoca
on 6/05/2017:
— Inlet vapor temperature to the absorber(maximum 105°F)
— Inlet oil temperature to the absorber(maximum 105°F)- DAQ incorrectly referenced this
as oil temp to the scrubber in the letter
— Oil flow rate through the scrubber system(minimum 10 gpm)
— Oil temperature into the stripper(minimum 200°F)
Avoca submitted permit application 17B for the alternative parametric monitoring and permit T52
was issued 4/22/2019 containing the limits. Avoca established parametric monitoring limits
during the 6/06/2017 performance testing. They attempted to operate controls at reduced
efficiency in order to produce the requested 105°F maximums. However,Avoca could not
physically force the inlet air temperature and inlet oil temperature to the absorber any higher than
70-80% of the maximums. Raleigh Permits accepted the test results,and no additional testing
requirements were added in T52.
The oil flow through the scrubber during the 6/06/2017 test ranged 9.88 gpm-10.06 gpm. The
heated oil temperature to the stripper during the test was 209.3°F-210.1°F During this inspection
the absorber inlet vapor temperature was 54.6°F,the absorber inlet oil temperature was 55.8°F,
the oil flow rate was 12 gpm and the stripper oil temperature was 225°F.
While the regulation requires the CPMS to record data at least every 15 minutes,Avoca is
monitoring every five minutes. The permit requires Avoca to record the 24-hour(daily) average
of each parameter. These records appeared to be complete for CY2018-January 2019 (copy of
the data is attached to the file copy of this report).
CPMS instruments must be calibrated annually. Avoca conducts quarterly temperature gauge
checks and flowmeter calibrations.
See Betty Gatano's T46 permit application review for a thorough breakdown of the requirements for
SDE-2 to comply with the MON.
• For the batch process vents associated with the SDE-1 and EVG Avoca must comply with the
requirements of 40 CFR§63.2460 and Table 2 of 40 CFR Part 63,Subpart FFFF.To maintain Group
2 classification for these emission sources the organic HAP emissions must be less than 10,000
pounds per consecutive 365-day period from each source. Avoca must monitor the organic HAP
emissions from each of these emission sources,monthly, as follows:
(SharePoint\Facilities\Bertie08\00044\agl6s\2018-2019 Inspection\20190130a16.doc) Page 46

EVG:
Organic HAP= 1.0 pounds x B
( batch ) ]
Where:
B= The number of batches processed in the EVG Operations
The daily 12-month rolling total HAP emissions did not exceed 550 lbs in CY2018.
SDE-1:
� ounds) 1r / ounds) / ounds) ounds)OrganicHAP= 2.27 pxBtfI+L4.801 p JxB,e$ + 3.661 pIxB.eC +L4.19(p IxB .tan 111 l batch L ` batch l J batch )
Where:
B,ez = The number of regular batches
B,e� = The number of recrop batches
Btc = The number of third crop batches
Btf = The number of hexane tank(M-2) fills.
Avoca's MON logs for SDE-1 are properly using the new equation.Note that this equation applies to
SDE-2 as well. SDE-1 has not operated since startup of SDE-2.
• For tanks,Avoca must comply with the requirements of 40 CFR§63.2470 and Table 4 of Subpart
FFFF. The tanks do meet requirements for Group 2.
• Heat exchangers must meet 40 CFR §63.2490 and Table 10 of Subpart FFFF. There are four cooling
towers at Avoca. The regulation requires preparation and implementation of a monitoring plan that
documents the procedures that will be used to detect leaks of process fluids into cooling water. The
plan must require monitoring of one or more surrogate indicators or monitoring of one or more
process parameters or other conditions that indicate a leak.
Avoca's plan is attached to the file copy of the 2013 inspection report. Brian Conner collects
quarterly cooling water samples and sends them to SGS Environmental. The samples are analyzed
for n-hexane,ethyl acetate, chloroform, and methanol, according to the area the samples were taken.
If>1 ppm of these compounds is detected, a leak is assumed. The lab's detection limits are all less
than 1 ppb. The records indicate there have been no leaks since the last inspection. Brian said that
they have never had a sample come back indicating a leak. I reviewed CY2018 records for three
quarters (dated 3/28/2018, 6/27/2018, 9/25/2018). The fourth quarter results had not yet been stored
in the hard copy file, and Mr. White did not know where to retrieve the information.
The regulation requires a leak to be repaired no later than 45 calendar days receiving sample results.
Once a leak has been repaired,Avoca must confirm by sampling that the heat exchange system has
been repaired within 7 calendar days.
Recordkeeping Requirements
• Create and retain a record of each time a safety device is opened to avoid unsafe conditions. There
are no records, as this has not occurred.
• Create and retain the following records on each affected operation:
i. A description of the process and the type of process equipment used;
ii. An identification of related process vents (including associated emissions episodes),wastewater
points of determination, and storage tanks;
iii. The applicable control requirements pursuant to 40 CFR Part 63, Subpart FFFF, including the
level of required control, and for vents,the level of control for each vent;
iv. The control device or treatment process used, as applicable, including a description of operating
and/or testing conditions for any associated control device;
(SharePoint\Facilities\Bertie08\00044\agl6s\2018-2019 Inspection\20190130a16.doc) Page 47

v. The process vents,wastewater POD,transfer racks, and storage tanks(including those from other
processes)that are simultaneously routed to the control device or treatment process;
vi. The applicable monitoring requirements of this subpart and any parametric level that assures
compliance for all emissions routed to the control device or treatment process;
vii. Calculations and engineering analyses required to demonstrate compliance.
Avoca appears to have records that meet the above requirements. The requirements above are met
through LDAR records, P&ID diagrams, and recordkeeping associated with other permit
requirements outside of MON, as well as other specific permit conditions associated with the MON.
Their wastewater is collected in a fully closed piping system.
• The mass removal efficiency of the on-site biological wastewater treatment plant(i.e. Wastewater
Treatment Plant Aeration Tank No. l; ID No. W WTP-ATI)for methanol is tracked under the PSD
permit condition 2.2.B. on a daily basis,when the wastewater stream consisting of methanol-wash
from the Recovery operations(ID No. ES-1001-1-1-WW)is discharged to it. The mass removal
efficiency must be more than 95%. Avoca uses the EPA WATER9 model to calculate removal
efficiency and methanol emissions from the tank. I reviewed the CY2018-January 2019 records and
did not discover any issues.
• Create and retain a schedule or log of operating scenarios for the batch operations updated each time a
different operating scenario is put into effect. Avoca records the daily batches processed at MON
affected areas(currently Rotocel/Recovery,EVG, and SDE-2). SDE-1 hasn't operated since startup
of SDE-2. I viewed their records for CY2018 and did not discover any issues.
• Recordkeeping requirements for SDE-2 are summarized below.
i. The parametric monitors must record data at least every 15 minutes. Record the daily averages of
each parametric monitor.
ii. The daily average records must be retained for 5 years.
iii. Periods of startup, shutdown, and malfunction may not be excluded.
iv. Records of monitoring system calibration checks and the maintenance performed must be
maintained.
I reviewed the daily CPMS averages for CY2018 -January 2019. No issues were discovered. A
copy of the daily average spreadsheet is attached to the file copy of this report. The beginning of
each data day starts at 12AM. The parametric monitors record data every five seconds. SSM appears
to be included in their data.
CPMS calibration records are in their file.
Note that periods of no flow may not be used in daily averages, and they may not be used in fulfilling
a minimum data availability requirement. I saw no evidence that Avoca is violating this condition.
Avoca is not employing supplemental gases at SDE-2, so permit condition(u)(iv)(f)does not apply.
• For each affected MPCU with a Group 2 process vent,retain the following records:
i. A record of the day each batch was completed;
ii. A record of whether each batch operated was considered a standard batch;
iii. The estimated uncontrolled and controlled emissions for each batch that is non-standard;
iv. Records of the daily,rolling 12-month summations of emissions, or alternative records that
correlate to the emissions (e.g., number of batches), calculated no less frequently than monthly.
I have reviewed the CY2018 daily batch records,daily HAP emissions, monthly emissions and
rolling 12-month HAP emissions records for EVG. Avoca appears to be properly keeping
these records. Rotocel and Recovery are a continuous process, and the monthly and rolling 12-month
(SharePoint\Facilities\Bertie08\00044\agl6s\2018-2019 Inspection\20190130al6.doc) Page 48

HAP emissions are tracked as part of the PSD, permit condition 2.2.B. recordkeeping
requirements. All of this data is provided in their semi-annual reports.
• For the process equipment leaks from the affected sources, Avoca must retain each applicable record
required by 40 CFR Part 63, Subpart UU. Avoca must comply with the recordkeeping requirements
of the LDAR program found in Sections 2.2 B.2.ii. through jj. of the permit. The records have been
reviewed for CY2018 — January 2019 and Avoca appears to be in compliance. See the discussion
above for permit condition 2.2.13.
• For each affected Group 2 wastewater stream,retain the following records:
i. MPCU identification and description;
ii. Stream identification code; ,
iii. Concentration of compounds listed in Table 8 and Table 9 of 40 CFR Part 63, Subpart FFFF (in
ppmw), including documentation of the methodology used to determine concentration;
iv. Stream flow rate(in liters/min).
The wastewater treatment system is fully closed pipe. Methanol is the only listed compound
employed at Avoca. The methanol content of the wastewater is sampled and analyzed daily. The data
is recorded in a spreadsheet, along with the generated gallons wastewater/day, average liters/sec flow
of the wastewater through the system for each day, and the daily calculated mass removal efficiency
of methanol in the system. I reviewed the records for CY2018 and did not discover any issues.
• For each affected heat exchanger system, retain the following records:
i. Monitoring data indicating a leak,the date when the leak was detected, and if demonstrated not to
be a leak,the basis for that determination;
ii. Records of any leaks detected by procedures other than those provided in the written heat
exchanger monitoring plan, including the date the leak was discovered;
iii. The dates of efforts to repair leaks;
iv. The method or procedure used to confirm repair of a leak and the date repair was confirmed.
There have been no leaks ever detected for the heat exchangers at Avoca. The records appear
complete. The towers are tested quarterly.
Reporting
• The Permittee must submit a Notice of Compliance Status Report for the Botanical,Biomass,
Concrete and/or the PNE operations prior to operation in organic HAP service (using a fluid that
contains at least 5 percent by weight HAP).
• The Permittee shall submit a Notification of Compliance Status (NOCS) Report for the SDE-2
operations no later than 150 days after startup. Avoca submitted this report on 7/07/2017.
• For any process vents that change from Group 2 to Group 1,the Permittee shall comply with the
notification requirements of 40 CFR§63.2460(b)(6) and 40 CFR§63.2520(e)(10). This has yet to
occur at Avoca.
• The Permittee shall submit a semiannual compliance report of the following:
i. Company name and address,with responsible official signature certifying the accuracy of the
content of the report;
ii. Date of report and beginning and ending dates of the reporting period;
iii. If there are no deviations from any emission limit, operating limit or work practice standard
specified in this subpart, include a statement that there were no deviations from the emission
limits,operating limits, or work practice standards during the reporting period;
(SharePoint\Facilities\Bertie08\00044\ag16s\2018-2019 Inspection\20190130al6.doc) Page 49

iv. For each deviation from an emission limit, operating limit, and work practice standard, include
the following information:
(A) The total operating time of the affected source during the reporting period; and,
(B) Information on the number, duration, and cause of deviations(including unknown cause,
if applicable), as applicable, and the corrective action taken.
v. Identification each new operating scenario which has been operated since the time period covered
by the last compliance report and has not been submitted in the previous compliance report.For
the purposes of this paragraph, a revised operating scenario for an existing process is considered
to be a new operating scenario;
vi. For the equipment listed below,report in a summary format by equipment type,the number of
components for which leaks were detected and for valves,pumps and connectors show the
percent leakers, and the total number of components monitored. Also include the number of
leaking components that were not repaired as required, and for valves and connectors, identify the
number of components that are determined to be non-repairable as described in 40 CFR
§63.1025(c)(3).
(A) Valves in gas and vapor service and in light liquid service;
(B) Pumps in light liquid service;
(C) Connectors in gas and vapor service and in light liquid service; and,
(D) Agitators in gas and vapor service and in light liquid service.
vii. Where any delay of repair for leaks is utilized, report that delay of repair has occurred and report
the number of instances of delay of repair.
vii. For pressure relief devices,report the results of all leak monitoring to show compliance
conducted within the semiannual reporting period.
viii.Report, if applicable,the initiation of a monthly leak monitoring program for valves.
ix. For each affected heat exchanger system for which the Permittee invokes the delay of repair,
include the following information:
(A) The presence of the leak and the date that the leak was detected.
(B) Whether or not the leak has been repaired.
(C) The reason(s)for delay of repair.
(D) If the leak is repaired,the owner or operator shall report the date the leak was
successfully repaired.
(E) If the leak remains unrepaired,the expected date of repair.
Avoca's has submitted the MON semi-annual reports on time. Brian Conner has been certified by David
Peele as an authorized signatory for Avoca. I have reviewed the reports through January 2019, and each
of the above requirements are addressed. I have found no issues within the reports.
Avoca appears to be compliant with Subpart FFFF.
Other Applicable Permit Conditions
General Condition 3.R "Compliance Certification"
General Condition 3.X. "Annual Emissions Inventory Requirements"
I have reviewed and approved the CY2018 ACC which was submitted on 2/27/2019. The only issue
noted(which I shared by email with Brian Conner)was on page 7 where only permit T49 is cited for
Conditions 2.1.D.2. a-e. I think it should be T49, 50 and 51. This is a very minor observation and
doesn't require Avoca to submit an addendum.
(SharePoint\Facilities\Bertie08\00044\agl6s\2018-2019 Inspection\20190130a16.doc) Page 50

I have reviewed and approved the CY2017 emissions inventory,which Avoca submitted to DAQ on
AERO 4/24/2018. The certification was received 5/07/2018. My review notes are in the WaRO file and
saved to SharePoint at/Bertie08/00044/CY2017 Avoca Inventory Review Notes doc/ The electronic
copy of the emission calculations are found at/I3e_r__t_i_eOB/QQ�44/( Y2017.Llectr<3nic_C�.py of Elnissi�ns
Inventorv.xls.
5-yr Compliance History
• NOD 3/05/2014
Boilers ES-BB 1 and ES-13132 January-June 2013 oxygen analyzer and steam meter downtimes were
in excess of the thresholds considered acceptable in the NCCEP.
• NOV 2/22/2017
40 CFR 63.7575 (f)
Late submittal of MACT 5D stack test report.
• NOV 8/20/2018
40 CFR Part 63, Subpart ZZZZ—"NESHAPS for Stationary RICE", and
15A NCAC 2D .0530 "Prevention of Significant Deterioration"
Permit Specific Conditions 2.I1.3.f.i. and 2.2.B.2.r.
Inspection records review resulted in discovery that annual maintenance activities on two
emergency engines had not been completed. Additionally,the February 2017 monthly leak
detection test on the Botanical process area pumps were not conducted.
Final Comments and Conclusion
The facility appeared to be in compliance with the permit conditions and applicable regulations at the
time of the inspection.
(SharePoint\Facilities\Bertie08\00044\agl6s\2018-2019 Inspection\20190130a16.doc) Page 51