AQ_F_0100237_20130306_CMPL_Fac-Ltr

Canfor Southern Pine CANF(#)R
New South Lumber Co.,Inc.
March 6, 2013
Mr. Alan Drake ;
Division of Air Quality,Technical Services Section �. f
NC Department of Environment and Natural Resources i
1641 Mail Service Center MAR 8 2013
Raleigh,North Carolina 27699-1641
Re: New South Lumber Co.,Inc. —Graham Plant
Facility ID No 0100237
Air Permit No. 06740T17
Review of COM System Performance
Letter Dated February 12,2013
Mr. Drake,
We have received and reviewed your letter dated February 12, 2013 summarizing your review of the second semi-
annual report submitted by the New South—Graham facility on January 21, 2013. The objective of this letter is to
address the potential violation raised in your review letter.
As you know,the equipment of concern is Boiler#4(ID B-4), a 57.6 MMBtu/hr, wood-fired boiler equipped with
two (2)multiclones and an electrostatic precipitator(ESP), all in series. The boiler was installed in September 2007
and began operation in December 2007. Based on its installation/construction date,B-4 is subject to particulate
matter and visible emissions limitations under the New Source Performance Standards (NSPS), Subpart Dc. In
particular,NSPS Subpart Dc requires the facility to limit visible emissions from B-4 to<20%opacity
(40CFR60.43c(c)) and particulate matter emissions to<0.030 lb/MMBtu (40CFR60.43c(e)(1)).
NSPS Subpart Dc establishes a preferred monitoring approach that relies upon using a Continuous Opacity Monitor
(COM)to demonstrate ongoing compliance with the particulate matter and visible emissions limits
(40CFR60.47c(a)). The New South—Graham facility completed initial setup and performance certification of the
B-4 COM system on December 18, 2007. Since that time and through to June 2012, the facility has had 0.0%
monitor downtime for 16 of 18 calendar quarters, with only two (2) quarters where any portion of the COM system
malfunctioned or was unavailable during operation of the boiler. In the first incident, during the second quarter of
2010, the laser mirror associated with the COM became obscured during a period of heavy rain. Based on
instruction provided by the COM equipment supplier(Teledyne),the mirror could not be cleaned or re-aligned until
dry weather. Once ambient conditions improved, facility personnel cleaned and re-aligned the mirror and proper
operation of the COM resumed. Total monitor downtime associated with this incident was 2.38% that quarter,with
both the cause of the malfunction and the duration of the downtime attributed wholly to weather conditions. The
second incident, during the second quarter of 2011, was also caused by moisture condensing on the laser mirror in
the stack during heavy rainfall and was addressed by facility personnel as soon as ambient conditions improved.
This incident led to monitor downtime of 0.14% for the quarter. Looking at overall operation of the
Graham Plant
4408 Mt. Hermon-Rock Creek Road, Graham, North Carolina, U.S.A. 27253
Telephone [phone] Fax [phone]

Mr. Alan Drake
March 6,2013
Page 2
equipment,the New South—Graham facility had a data capture rate of 99.83%for the first 4.5 years
operation of the COM. Furthermore,the limited monitor downtime in this time period was fully
attributed to weather conditions.
In the 3rd quarter of 2012, the facility experienced an electrical malfunction that interrupted
communication between the COM system at the stack and the data acquisition system(DAS)used to
analyze the COM readings to generate 6-minute averages and record the results. The malfunction
first appeared on 07/16/2012 as an error code(-999)on the remote COM readout associated with the
data acquisition system. The malfunction initially lasted approximately 3 hours,during which time
site personnel checked and tightened the electrical connections between the COM system at the stack
and the remote COM/DAS in the control room. Normal function of the data acquisition system
resumed, and the facility believed the problem to be addressed.
Three days later,near midnight on Thursday, 07/19/2012, the error code(-999)reappeared on the
remote COM panel and information transfer to the data acquisition system was interrupted. The
COM analyzer on the stack continued to function properly and provide instantaneous readings,which
were monitored by facility personnel. The facility followed the troubleshooting procedures
established and approved in its Opacity Monitoring Plan. When the problem was not resolved by the
next afternoon,the facility contacted the COM manufacturer(Teledyne) for support. The Teledyne
support personnel agreed with the troubleshooting efforts taken by the site and suggested as a next
step complete replacement of the electrical cable running from the COM analyzer on the stack to the
remote COMMAS unit in the control room. The facility has a local supplier for this electrical cable;
however, the supplier did not have a single cable of adequate length(400 ft.) in stock to meet this
need. Splicing together multiple lengths of shorter cable would potentially decrease signal quality
and increase sensitivity of the system to moisture. A spool of cable was ordered and arrived the next
business day(Monday, 07/23/2012); however,heavy rainfall prevented installation of the new cable
until Tuesday, 07/24/2012. Proper data capture resumed upon completion of the cable replacement
effort and once again,the facility believed the problem to be addressed. However,the error code and
associated data capture problems re-appeared near midnight on Friday, 07/27/2012. The facility
resumed troubleshooting efforts,but the remote COM/DAS resumed normal operation without
intervention on Monday morning, 07/30/2012. During this period,the stack COM continued to
function properly, and the facility checked those instantaneous readings frequently to confirm visual
observations that boiler visible emissions were negligible(<4% opacity).
The COM/DAS functioned properly for the next 16 days. Then,near midnight on Thursday,
08/16/2012,the error code(-999)reappeared on the remote COM/DAS in the control room. As in
the past, the stack COM continued to function properly and provide instantaneous opacity readings.
The facility started over with troubleshooting the entire system. After checking/testing the entire
system from detector to DAS and eliminating problems in each of these areas, the facility concluded,
in consultation with the COM manufacturer, that the problem must reside inside a sealed circuitry
panel on the stack COM unit. This panel is normally kept sealed to avoid moisture intrusion to the
sensitive electronic components. No routine maintenance is recommended for components inside
this panel, and problems are not normally experienced with these parts. However, after eliminating
all other potential causes for the communication error,Teledyne support personnel instructed facility
personnel to open the sealed panel (under dry, ambient conditions) and remove the printed circuit
boards. Each circuit board was inspected and the connections were cleaned prior to re-insertion.
After cleaning and returning the circuit boards,the panel was re-sealed and the COM system

Mr. Alan Drake
March 6,2013
Page 3
restarted. Proper operation of the COM/DAS resumed on Thursday, 08/30/2012. No further
problems with the remote COMMAS have been observed since this time. Monitor downtime for the
4th quarter of 2012 was 0.0%.
The DAS was unable to collect valid opacity data for approximately 570 hours in the Yd quarter of
2012 due to this exceptional electronic communication malfunction. The COM itself continued to
work properly throughout the malfunction,providing instantaneous visible emissions readings at the
stack. Facility personnel monitored these instantaneous readings periodically during each period of
lost communication. Based on those instantaneous readings and frequent visual observations of the
emission point, boiler stack opacity remained below 5% throughout this time period. In addition,the
boiler control panel is equipped with numerous alarms to notify operators of the most-likely upset
conditions on the boiler or the ESP. None of these alarms were activated during this period of
improper data communication between the COM system and the data acquisition system. There is no
indication of boiler or ESP malfunction,nor any other reason to suspect an emissions exceedance
during this time period. As such, we believe there was no harm to the environment associated with
the electrical malfunction and lost data capture.
The facility has an outstanding history of compliance and successful operation of this monitoring
system. Even with this extended malfunction,the facility has achieved 98.4%data capture from the
COM in the 5 years it has operated. We hope the Department will consider this excellent
performance record in its evaluation of this particular incident.
We also believe it is important to look at the situation in the context of the specific regulation. As
stated earlier,use of a COM is certainly the preferred method of compliance monitoring under NSPS
Subpart De. However, the regulation does allow certain boiler operators to conduct Method 9 visible
emissions tests in lieu of installation and use of a COM(40CFR60.47c(a) and(f))—including those
with wood-fired boilers equipped with ESP particulate matter control. Numerous conditions affect
the frequency of required Method 9 observations; however,under no circumstance does the
regulation require Method 9 observations any more frequently than once every 45 days. While we
understand that our current permit is not written to make use of this alternative method of monitoring
compliance, we believe it provides a valuable point of comparison when considering the impact of
the DAS malfunction discussed above. The entire period of concern, starting from the first error
message on 07/16/12 to final resolution of the problem on 08/30/12,took place in 45 days. During
that time,numerous periods of normal COM/DAS function occurred allowing normal data capture
and instantaneous COM readings were available throughout.
We have instituted several procedural changes in response to this incident. In the future,if the DAS
loses contact with the COM, facility operators will maintain manual,hand-written logs of the
instantaneous COM readings throughout the period of DAS malfunction. Readings will be taken on
an hourly basis during boiler operation until normal DAS function is restored. In addition, facility
personnel will seek input from the COM manufacturer for any COMMAS malfunction that is not
resolved within 24 hours. We have also arranged to have a member of our boiler supervisory staff
attend smoke school. We believe this will enhance our ability to monitor overall boiler operations,
regardless of COM/DAS performance.
We strive to maintain compliance with regulatory obligations at all times, and attempt to go beyond
the minimum whenever possible. In keeping with this philosophy, facility management encourages a

Mr. Alan Drake
March 6, 2013
Page 4
proactive schedule of preventive maintenance for all equipment at the mill and provides extensive
training to both operations and maintenance personnel to enhance operation of the facility. We
believe our response to this malfunction was reasonable, especially in the context of our performance
history with the equipment in question and the alternative compliance monitoring options allowed in
the applicable regulation. In addition,the incident did not impair performance of the boiler or
control device or otherwise cause harm to the environment.
Please let us know if you have any additional questions concerning this incident or the actions we
have taken to prevent a recurrence. We would be happy to meet with you and discuss the situation in
person. Thanks again for your time.
Sincerely,
c
Ishop
General Manager
Attachment
c: Margaret Love/WSRO
Michael Pjetraj/RCO
Debbie Sand/New South
Kathy Ferry/KJF Consulting