June 28, 2007

Mr. Gerardo C. Rios

Chief, Permits Office

United States Environmental Protection Agency

Region IX

75 Hawthorne Street

San Francisco, CA 94105-3901

Dear Mr. Rios:

This letter amends the revised Prevention of Significant Deterioration
(PSD) application submitted to your office on December 11, 2006 for the
Big West of California, LLC Bakersfield Refinery Clean Fuels Project
(CFP).  There are several amendments to be made to the application,
specifically:

The refinery has committed to meeting a fuel gas sulfur content limit of
40 ppmv total sulfur (as H2S) on a 4-hour average basis;

Basic information is presented about the additional equipment that will
be necessary to achieve this fuel gas sulfur level;

The location of the existing Mild Hydrocracker and its two heaters 14-H1
and 14-H2 was incorrectly modeled;

Additional supporting information is presented regarding the cost
effectiveness of installing selective catalytic reduction (SCR) on the
two CFP heaters that are less than 50 MMBtu/hr (VGO Feed Heater, 47
MMBtu/hr; and VGO-HDS Fractionator Feed Heater, 35 MMBtu/hr); 

An analysis of the proposed NSPS Subpart Ja and how it affects the
project combustion units and FCCU, which indicates that a short-term NOx
emissions limit will be required for the FCCU, additional CEMS or
parametric monitoring will be required for PM emissions from the FCCU,
and additional NOx CEMS monitoring will be required for the VGO-HDS
heaters; and

Additional clarifications on the design and operation of the ground
flare that will serve CFP units.

Each of these amendments and additional ground flare data are described
in detail below.

Fuel Gas Sulfur Limit

As you are aware, Big West has worked extensively with our process
design engineers to achieve a fuel gas sulfur limit lower than 100 ppmv.
 As presented in the July 12, 2006 letter from J. Scott Lewis of Linde
BOC Process Plants LLC (provided as an attachment to an email dated July
24, 2006 from Ev Ashworth), the contractor for the amine treatment
system was only willing to guarantee that 50 ppmv H2S could be achieved.
However, as the design has progressed, we were able to develop detailed
fuel gas balance scenarios to ascertain the expected fuel gas sulfur
content in normal and “worst-case” operating situations. Big West
now anticipates that with the addition of a caustic scrubber to remove
the potentially high level of sulfur in the fuel gas from Area 3 that
may, in some circumstances, be introduced into the Area 4 fuel gas
system, Big West can meet a limit of 40 ppmv total sulfur (as H2S)
averaged over a 4-hour period at all times.

This modification will not adversely affect the PSD permit application.
Modeling will not be performed again, as the current modeling is
conservative in its inclusion of 100 ppmv sulfur in the combusted fuel
gas.  To meet the 40 ppmv limit, the facility will need to treat fuel
gas supplied to Area 2 from the Area 3 Delayed Coker Gas Amine Treater,
to reduce non-H2S sulfur compounds in the total fuel gas burned in Areas
2 and 4.  The new treatment unit will use caustic to extract these
sulfur compounds; the sulfur compounds will be converted to disulfides
and returned to a hydrotreater for conversion to H2S.

The addition of this caustic scrubbing unit does not directly affect the
PSD permit application, as the only emissions from the unit will be
volatile organic compounds (VOC), which is not within the scope of the
PSD permit. The refinery will submit a revised application for a Permit
to Construct with the San Joaquin Valley Air Pollution Control District
to support construction of the revised Area 4 fuel gas treatment unit. 

A revised Table 5-3, that reflects the lower SO2 emissions resulting
from combustion of fuel gas subject to a 40 ppmv sulfur content limit,
is presented below.

Table 5-3: Big West Clean Fuels Project Source Emission Ratesa

Source

ID	

Model ID	NOx

(g/s)	SO2b

(g/s)	CO

(g/s)	PM10

(g/s)

VGO Feed Heater (47 MMbtu/hr)	vgohtr	0.1438	0.0833 

0.0333	0.2189	0.0441

VGO HDS Fractionator Feed Heater (35 MMBtu/hr)	vgofrhtr	0.1071	0.0620

0.0248	0.1630	0.0329

Hydrogen Plant Reformer	h2reform	0.4904	1.1364

0.4545	0.5971	0.6018

FCCU Regenerator (annual) c	fccuregen	1.0604	1.4765	1.9046	1.1647

FCCU Regenerator (1-hr, 3-hr, 8-hr, 24-hr) c	fccuregen	2.1208	3.6913
16.1407	1.1647

Existing MHC Feed Heaters

(14-H1 & 14-H2)	mhc14h12	0.3748	0.1596

0.0638	1.4904	0.0845

HF Alky Isostripper Reboiler	hfreboil	0.1645	0.3812

0.1525	0.2003	0.2018

SWAATS Unit	swaats	0.0000	0.2322	4.3994

0.3384	0.0000

Ground Flared	gndflare	0.0279	0.0089

0.0057	0.1519	0.0107

Diesel Firewater Pump Engines (annual)	firepump	0.0222	0.0000	0.0130
0.0007

Diesel Firewater Pump Engines (24-hr)	firepump	0.0809	0.0001	0.0474
0.0027

Diesel Firewater Pump Engines (8-hr)	firepump	0.2428	0.0003	0.1422
0.0082

Diesel Firewater Pump Engines (3-hr)	firepump	0.6475	0.0007	0.3792
0.0219

Diesel Firewater Pump Engines (1-hr)	firepump	1.9425	0.0022	1.1375
0.0656

Cooling Tower 1	coolt1	0.0000	0.0000	0.0000	0.0303

Cooling Tower 2	coolt2	0.0000	0.0000	0.0000	0.0303

a Strikeout values are from the December 2006 revised PSD application.

b  Revised sulfur emission rates reflect combustion of refinery fuel gas
at 40 ppmv total sulfur, expressed as H2S.

c  FCCU heater is a limited-use startup heater.  FCCU regenerator
emission rates are larger than those from the FCCU startup heater, so
FCCU regenerator emissions were used in the modeling.

d Flare emission rates are annual averages that include process unit
startups and shutdowns as well as continuous pilot flaring.



Mild Hydrocracker Location

It recently came to our attention that the location of the existing mild
hydrocracker, which has been included as an affected unit in the PSD air
quality impact analysis modeling for the  Clean Fuels Project, was
misrepresented in these dispersion modeling runs. The UTM coordinates
should not have been  311795.2 Easting and 3917118.9 Northing (NAD27) as
presented in Table 5-4 of the December 2006 revised application, but
rather should be approximately 220 m SSE, at 311837.5 Easting and
3916901.5 Northing. 

The dispersion modeling to compare maximum project impacts with Class II
significance levels and monitoring significance levels has been
performed again to take this change into account. The affected tables
and figures from the December 2006 revised PSD application are included
below. As noted above, this revised modeling analysis does not reflect
the new and reduced fuel gas total sulfur content limit of 40 ppmv. 
However, this conservative approach demonstrates that the revised
location of the Mild Hydrocracker does not result in any exceedances of
relevant EPA PSD Significance Levels.  The modeling to determine Air
Quality Related Values (AQRV) impacts on Class I areas has not been
revised, as the effect of a 200 m shift of one emission source would be
imperceptible at the distance of the nearest Class I areas (~80 km).

Figure 5-2: Big West – Boundary Receptors and Source Locations

Figure 5-3: Big West Far Grid of Receptors

Figure 5-5: Maximum Impact Locations with Fine Grid for Maximum 3-hour
SO2 Impact

Table 5-9: Maximum Project Impacts Compared with Class II Significance
Levels and Monitoring Significance Levelsa

Pollutant	Averaging

Period	Maximum Predicted Impact

(µg/m3)	Class II Significance Level

(µg/m3)	Monitoring Significance Level

(µg/m3)

NO2	Annual	0.56b

0.68	1.0	14

SO2	Annual	0.74

0.83	1.0	NA

	3-hour	10.67

10.71	25.0	NA

	24-hour	3.18

3.37	5.0	13

CO	1-hour	181.31

183.42

	2,000	NA

	8-hour	44.08

31.38	500	575

Notes:

     a Strikeout values are from the December 2006 revised PSD
application.

       b EPA default Ambient Ratio Method factor of 0.75 applied.

     NA = Not applicable/not defined



Cost Effectiveness of SCR on VGO-HDS Heaters

The BACT analysis presented in section 4.2.1 of the December 2006
revised PSD application concluded that BACT for NOx for refinery
combustion units less than 50 MMBtu/hr is the installation of low NOx
burners to achieve a NOx emission limit of 20 ppmv @ 3% O2. This
conclusion was reached with the following reasoning:

The most stringent limit found to be achieved in practice or required by
a state implementation plan (SIP) was 25 ppmv @ 3% O2.

The lowest vendor guarantee that the refinery was able to secure for
state-of-the-art low NOx burners on a refinery heater of this size is 20
ppmv @ 3% O2.

While the addition of SCR would be technically feasible and could
achieve lower NOx emissions, this is not achieved in practice on a small
refinery heater and is not considered cost effective; the cost
effectiveness, calculated at $13,766 and $12,779 per ton of NOx control
for the smaller and larger heaters, respectively, was far above the
SJVAPCD’s cost effectiveness threshold for NOx control.

You and your staff have indicated that EPA may not agree with the cost
effectiveness thresholds as established by SJVAPCD and requested a more
complete accounting of costs associated with the installation of SCR.
Because the design of the project has progressed since the cost
estimates were initially prepared over a year ago, and detailed cost
estimates have been obtained for the other CFP heaters, a more complete
cost estimate can now be provided.  As we have explained to your staff,
a more detailed cost analysis was not provided in the December 2006
revised PSD application because the estimated cost-effectiveness
exceeded the SJVAPCD BACT cost-effectiveness thresholds.  We note that 
the addition of SCR units on the VGO-HDS heaters this late in the
project design would significantly increase these costs – unit
redesign/placement, re-engineering, cancellation charges for parts
already ordered, etc. – none of these schedule- and redesign-related
costs have been included. Only incremental costs between installation of
low NOx burners alone and installation of low NOx burners and SCR have
been included in our revised cost-effectiveness analysis.

Revised Tables C-4 are attached, which provide the new cost estimates
and cost effectiveness calculations. As before, guidance from the EPA
OAQPS Cost Manual, 6th Edition, Chapter 2 regarding cost estimates for
SCR was followed, except where more specific data were available. Cost
effectiveness estimates for the smaller and larger heaters were $45,170
and $39,450 per ton, respectively. 

In summary, there are no existing refinery heaters or boilers <50
MMBtu/hr that are permitted to achieve NOx emission rates lower than
proposed here, the proposed NOx emission limits for the VGO heaters are
more stringent than any applicable SIP or proposed NSPS Subpart Ja
requirements, and the cost effectiveness for SCR control is
significantly more expensive than the BACT cost effectiveness thresholds
required for refinery units or similar sources in California or
elsewhere in the United States (under EPA, South Coast, Bay Area or San
Joaquin Valley air district guidelines).  We therefore conclude that SCR
controls as applied to the VGO-HDS heaters are not representative of the
lowest achievable NOx emission rate.

Proposed NSPS Subpart Ja

On May 17, 2007, EPA proposed amendments to the New Source Performance
Standards for Petroleum Refineries (Subpart Ja and proposed
modifications to Subpart J, see 72 FR 27178, 5/17/2007).  NSPS
requirements are effective based on the date of proposal; therefore,
affected facilities in the Clean Fuels Project will have to comply with
these requirements.  We note that the proposed rule is subject to review
and comment, and may be modified by EPA in light of these comments. 
Nevertheless, as demonstrated in the table below, the affected units
under CFP can comply with Subpart Ja requirements with the following
changes to the proposed project:

Unit	Proposed NSPS Ja Requirement	CFP Controls/Design

FCCU	PM: 0.5 lb/1,000 lb coke burn-off	Same

	PM Monitoring: Method 5 performance test; PM CEMS or control device
operating parameter monitoring	Proposed continuous opacity monitoring;
will incorporate Ja monitoring requirements (PM CEMS or parameter
monitoring)

	NOx: 80 ppmv (dry, 0% O2 ) 7-day rolling average	More stringent: 40
ppmv (dry, 0% O2 ) daily average; 20 ppmv (dry, 0% O2 ) 365-day rolling
average

	NOx Monitoring: CEMS	Same

	SO2: 50 ppmv (dry, 0% O2 ) 7-day rolling average; 25 ppmv (dry, 0% O2 )
365-day rolling average	More stringent: 50 ppmv (dry, 0% O2) daily
average; 20 ppmv (dry, 0% O2 ) 365-day rolling average

Claus Sulfur Recovery Plant	Provides new SO2 and H2S emissions limits
Not applicable to SWAATS unit

Process Heater and Other Fuel Gas Combustion Device	NOx: 80 ppmv (dry,
0% O2 ) 24-hour rolling average	More stringent (<20 ppmv @ 3% O2 15
minute average for CEMS)

	NOx Monitoring: CEMS	Proposed periodic sampling on VGO-HDS units to
verify compliance; will install CEMS

	SO2: 20 ppmv (dry, 0% O2 ) 3-hour rolling average – or fuel gas limit
of 160 ppmv H2S 3-hour rolling average; 8 ppmv (dry, 0% O2 ) 365-day
rolling average – or fuel gas limit of 60 ppmv H2S 365-day rolling
average	More stringent: 40 ppmv total sulfur limit, expressed as H2S  on
a 4-hour average



As summarized above, the current BACT emission limits for the CFP units
are equal to or more stringent than the proposed Subpart Ja
requirements; however, a new short-term NOx standard will be
incorporated to address the proposed Ja standard for new FCCUs. 
Further, additional monitoring will be required to meet the proposed
Subpart Ja monitoring requirements; specifically, installation of NOx
CEMS for the VGO-HDS heaters and installation of a PM CEMS or parametric
monitoring of PM emissions from the FCCU.

Other Issues

Separately, in recent email correspondence, Ms. Kathleen Stewart raised
several issues regarding the ground flare, for which our responses are
provided below:

The flare is designed to handle only process upset gases, taken to
include gases released during startup, shutdown and malfunctions; it
therefore is not designed to comply with Subpart J or proposed Subpart
Ja. Furthermore, it will not be permitted to handle releases subject to
Subpart J or Ja under its federally enforceable operating permit;

The minimum heat content of gases that will vent to the flare during
process upset conditions, startups and shutdowns will be 300 Btu/scf. 
We do not anticipate any instance where the heat content of process
upset gases will lower than 300 Btu/scf;

No pressure relief devices will vent directly to the flare;

The presence of a pilot flame on the ground flare will be monitored with
thermocouples, which will record temperature, and hence, the presence of
a pilot flame;

The flow of gases released to the ground flare will be monitored with a
GE Sensing ultrasonic flow meter (product brochure and technical data
are attached).  Please note that this unit does not require daily
calibrations; Big West will calibrate the unit consistent with
manufacturer’s recommendations; flow accuracy and repeatability data
are provided in the brochure;

The heat content and sulfur content of gases released to the flare
during process upset conditions will be monitored by a sampling system
that will consist of evacuated cylinders, which will sample gases during
a release.  We are working with vendors to define the specifications for
this sampling system, which we understand is used by other facilities in
California and required under the Motiva Consent Decree to measure heat
content and sulfur concentrations in gases that are released to a flare.
The sample gas obtained from the automated sampling system will then be
analyzed for heat content (ASTM Method D2382-88; D3588-91 or D4891-89)
and sulfur content (EPA Method 15/16 GC-FPD or equivalent); and

Finally, we wish to confirm that the sulfur content of the torch oil
used in the startup of the FCCU will be verified through testing.

A CD of the modeling input and output files will be sent to you under
separate cover; copies will be provided directly to Ms. Carol Bohnenkamp
as well. Please contact Mr. Everard Ashworth of Ashworth Leininger Group
(805.370.1469) to discuss any concerns or questions raised by this
letter.  Thank you again for your continued assistance on this important
project.  Thank you again for your continued assistance on this
important project. 

Very truly yours,

Eugene Cotten

Vice President-Refining

Enclosure

	

cc:	Carol Bohnenkamp, USEPA

	Kathleen Steward, USEPA

Vince Memmott, P.E., Flying J Inc. 

Bill Chadick, HSE Director, Big West

Everard Ashworth, ALG

Richard Karrs, SJVAPCD

Leonard Scandura, SJVAPCD

Perry Fontana, QEP, URS

Mike McCorison, US Forest Service

Table C-4: BACT Annual Cost Analysis – Refinery Combustion Units <50
MMBtu/hr

Table C-4: BACT Annual Cost Analysis – Refinery Combustion Units <50
MMBtu/hr (cont.)

Big West of California, LLC

A Subsidiary of Flying J Inc.

6451 Rosedale Highway, Bakersfield, CA 93308                            
  Page   PAGE  1 	

Tel. 661-326-4205  Fax 661-326-4382

          

	

Big West of California, LLC

A Subsidiary of Flying J Inc.

