Field testing of staged combustion on a petroleum process heater

FIELD TESTING OF STAGED COMBUSTION ON A PETROLEUM PROCESS HEATER BY: R. J. TlDONA W. A. CARTER KVB, INC. 7/ *~ IRVINE, CA FOR: AMERICAN FLAME RESEARCH COMMITTEE 1980 FALL MEETING NEWPORT BEACH, CALIFORNIA OCTOBER 27-28, 1980 A mmrdt-Ccttftl Cmpmy IVPCC 4 7 ^ - ^ ° SECTION 1 INTRODUCTION Subscale process heater combustion modification tests were previously conducted on a research scale furnace. These results of these tests were presented at the Third EPA Stationary Source Combustion Symposium. These tests evaluated the effects on N0 X emissions of several types of modifications, including staged combustion air, flue gas recirculation, lowered excess air, altered injection geometry, and low N0 X burner installation. The work showed that staged combustion air appears to be the most cost effective combustion modification for process heaters. Both staged air and flue gas recirculation produced N0 y emission reductions in excess of 60 percent below baseline emission levels. COMBUSTION MODIFICATIONS ON A FULL SCALE PROCESS HEATER The present report summarizes the testing of a full scale natural draft refinery process heater firing natural gas, refinery gas, and No. 6 oil fuels. A staged air injection system was designed by KVB and installed by the refinery for these tests. The system is capable of supplying up to 50 percent of the stoichiometric air requirement. The staged air is injected by means of lances inserted through the heater floor as it was in one of the subscale test configurations. The process heater was tested with and without the staged air modification and over a range of loads and stack excess oxygen concentrations. Stack gas emissions and heater efficiencies were measured for all conditions. Lowered excess air and staged combustion air modifications were applied separately and in combination. NOx emission reductions of up to 52 percent below baseline levels were achieved when firing gas fuel with no short term 1-1 P-263 ill effects on the heater. Heater efficiency was increased at the low NOy conditions due to the lowering of excess air. Staging the combustion air did not appear to significantly alter efficiency; the same level of lowered excess air was attainable both with and without staging. In the most favorable situations the combination of staged air and lowered excess air costs $64/Mg N0 x reduction. Staged combustion air applied separately is expected to be considerably more expensive-approximately $2600$3000 per Mg NOx reduction. OBJECTIVE AND SCOPE The objective of the program is to develop advanced combustion modification concepts requiring minor hardware modifications that could be used by operators and/or manufacturers of selected industrial process equipment to control emissions. The development is aimed at equipment on which the modifications will be most widely applicable and of the most significance in mitigating the impact of stationary source emissions on the environment. The program involves investigation not only of emissions but also multimedia impacts and control cost effectiveness • The program includes both subscale and full-scale testing. Subscale testing is a necessary part of development of new hardware to ensure acceptable performance, which is a vital aspect of emissions control. Full-scale testing is also necessary on more than one process design configuration (e.g., forced draft and natural draft) before equipment manufacturers and the process industry can employ a given emission control technology. At the conclusion of the study, a final engineering report will be prepared summarizing the accomplishments of the subscale and full-scale demonstration tests. A series of guideline manuals will be prepared to acquaint equipment manufacturers with the most promising emission control methods that have been demonstrated and to offer technical guidance that can be directly applied in their process equipment design. 1-2 P-263 SECTION 2 TEST HEATER AND EMISSIONS SAMPLING TEST UNIT DESCRIPTION The test unit was a natural draft, vertical cylindrical crude oil process heater which is used to supply a partially vaporized charge to a crude oil distillation column. A maximum load of 108 m /h (16,250 bbl/d) may be sent through the heater in two passes. A sketch of this crude oil heater is presented in Figure 1. The maximum firing rate of the heater is 16.1 Mw thermal input (55 x 10 Btu/hr). It is fired by six John Zink DBA-22* natural draft burners. The burners are combination gas/oil burners rated at a maximum of 2.68 Mw (9.14 x 106 Btu/hr) each with a turndown ratio of 3:1. Although combination gas/oil burners are used, some gas must always be fired because the unit is base loaded at constant oil firing rate and an automatic temperature controller adjusts the gas fuel flow to maintain crude oil outlet temperature. Three parameters may be controlled in the heater: furnace draft), firing rate, and load. excess oxygen, (by The stack damper is the main control for the pressure drop across the furnace. The pressure drop may also be controlled by opening or closing the secondary air registers which are in the base of the heater. Each burner has a set of primary and secondary registers •Mention of trade names does not constitute endorsement by the Environmental Protection Agency. 2-1 P-263 which can be adjusted independently of the other burners. The stack damper & and register adjustments establish the excess oxygen. The load is controlled by pumps and valves on the inlet and outlet of the heater. STAGED AIR SYSTEM DESCRIPTION This application of staged combustion utilizes air lances in the firebox to supply air to the flame zone a given distance above the base of the flame. Figure 2 presents a schematic of this system. The system consists of twenty-four vertical 316 stainless steel pipes 3.18 cm (1-1/4 in.) diameter arranged four per burner at 90° apart. A 45° elbow is placed on each pipe to provide better mixing across the flame. A fan supplies air to the lances through a manifold and flexible tubing. The lances may be varied in height up to 1.2 m (four feet) from the base of the burners. Extensions for the lances allowed staging heights up to 2.4 m (eight feet) for oil firing tests. EMISSIONS SAMPLING INSTRUMENTATION | Monitoring of the required gaseous and particulate emissions was performed with an EPA furnished mobile laboratory. The laboratory's monitoring capabilities are presented in Table 1. A schematic of the continuous monitoring system is presented in Figure 3. A detailed description of a similar mobile emissions laboratory has been presented in a previous report (Ref. J_) . Continuous gaseous emissions analyzers provide the capability for measurement of 0 2 , C02, CO, S02, NO, NOx, and HC (as methane) in the flue gas. Particulate total mass and sizing as well as S03 measurements are noncontinuous. S0 3 measurement is by a controlled condensation technique using a Goksoyr-Ross type coil• TI Hunter, S.C., et al. "Application of Gbmbustion Modifications to Industrial Combustion Equipment," EPA Report 600/7-79-015a, NTIS Order No. PB 294214, January, 1979. 2-2 P-263 ! i The 0 2 , C0 2 , and CO concentrations are measured on a dry basis. S0 2 , and N0 2 measurements are made through a heat-traced sample line and are de mined on a wet basis. NO may be sampled either wet or dry. 2-3 P-263 SECTION 3 DISCUSSION OF RESULTS LOAD AND EXCESS OXYGEN VARIATIONS A series of tests was conducted to evaluate the performance of the process heater with regard to N0 X emission and efficiency over varying load and excess oxygen conditions. These tests were first made while firing a gaseous fuel mixture consisting of natural gas and refinery fuel gas and then while firing residual oil simultaneously with the gas mixture. The NO emissions as a function of excess oxygen are shown in Figures 4 and 5. The highest NO emission generally occurred at the intermediate load condition, 70% of rated capacity. NOx emission increased with increasing excess oxygen up to approximately 4-6 percent 0 2 for all fuel combinations. In the 4-6 percent 0 2 range, NO emission was relatively insensitive to excess 0 5 level and, at higher excess 0 2 , NO emissions decreased. The nitrogen content of the gaseous fuels was negligible whereas the nitrogen content of the oil fuel was 0.8 percent by weight, hence, the difference in absolute NO emission between Figures 4 and 5. oil/gas are reported in Figure 5. Several ratios of These ratios indicate the approximate fuel split by percentage of total heat input. Changes in the fuel split occurred because a) gas fuel composition and heating value could change in a short time period and b) total heat input required to maintain the load could change due to changes in the composition of the crude charge to the heater. At the intermediate load, the baseline excess oxygen was 4 percent which was regarded by the plant as a normal operating level. The excess oxygen variation tests indicated that continuous operation of the heater at 2 percent 3-1 P-26 excess 0 2 was possible without additional operator supervision. level the heater draft tended to become unstable. Below this Fluctuations in the draft caused occasional smoking of the unit and resulted in positive stack or convection section pressures which are dangerous in a natural draft heater because they can cause flashback. Flashback occurs when a flame encounters a back pressure which forces it downward and out the bottom of the heater through the air registers. STAGED COMBUSTION AIR TESTING The next phase of testing at the process heater site involved the evaluation of the staged combustion air system. This evaluation included the variation of three important parameters: 1) burner equivalence ratio, *B, 2) excess oxygen level, and 3) staged air insertion height. For the staged combustion air tests, due to changes in plant operation, it was necessary to fire a different refinery gas from that used in load and excess oxygen variations. This gas, called "adsorber gas," contained a greater percentage of higher hydrocarbons than did the "fuel gas" or natural gas and, therefore, a higher heating value. | ' * Baseline NO emission with adsorber gas was about 8 percent higher (8 ppm) compared with fuel gas. The NO emission is graphed as a function of burner equivalence (A/F) ratio, <> = / F ) B u r n e r • The staging height for Stoichiometric this test series was 1.2 m (4 ft.) and the load was 64m3/h (9600 bbl/d), 60 percent of rated capacity. At each overall excess 0 2 level * B was decreased in steps to its minimum value which was determined by the limitations of the staged combustion air fan. Figure 6 shows that at 4 percent 0 2 the minimum 4>B (maximum staging) obtained was 0.74 which decreased NO emissions 35 percent below the baseline of 105 ppm, dry at 3 percent 0~. At 2 percent 0 2 and minimum * B of 0.65 the NO concentration dropped to 51 ppm, dry, corrected to 3 percent 0 2 . This represented a reduction of 51 percent below the 4 percent 0 2 baseline condition and a 43 percent below the 2 percent 0 2 (non-staged) level. Table 2 summarizes the N0 X * 3-2 P-263 reductions and efficiency changes for the various combustion modifications to the process heater while firing adsorber gas. Figure 7 shows similar trends in NO emissions for varying * B and excess O2 at a higher load. The same degree of staging could not be achieved at the intermediate load because of fan capacity limitations, however, the curves in Figure 7 indicate that roughly the same N0 X reductions would have been obtained had the minimum *B values of Figure 6 been reached. The results of another test series in which overall excess oxygen and burner equivalence ratio were kept constant while varying staged air insertion height are shown in Figure 8. (The staged air height is defined as the height above the heater floor at which the staged combustion air is injected. This is approximately equal to the height above the burner gas tips and oil gun.) There was little, if any, effect of staging height on NO emissions. Since the burner tile top was about 0.23 m (0.75 ft) above the floor of the furnace and since the staged air pipes were located on a diameter outside that of the burner tile injection heights of less than about one foot above the heater floor resulted in impingement of the staged air on the burner tile. Thus, the minimum staging height was approximately 0.3 m (one foot). The results of staging the combustion air while firing the No. 6 oil/adsorber gas mixture at intermediate load are shown in Figure 9 and 10. While lowering the excess air reduced N0 X by a percentage similar to that obtained firing gas fuel only, the use of staged combustion when firing oil with gas did not produce as large a percentage decrease in NOx emissions as it did when firing adsorber gas only. The absolute amount of NOx reduction, however, was about the same for combined fuel firing as it was for gas alone. For example, for the case of staged combustion air combined with lowered excess air firing oil and gas, Figure 9 shows that the NO level dropped from a baseline of 219 ppm, dry corrected to 3 percent 0 2 , to 166 ppm, dry at 3 percent 0 2 for a drop of 53 ppm. For the same conditions firing adsorber gas only at the intermediate load the drop was 46 ppm. This behavior indicates that fuel nitrogen conversion to NO x was reponsible for a large fraction of the observed emissions when firing the oil/gas mixture. It appears then that approximately half of the baseline N0 X 3-3 P-263 emission firing oil with gas is due to fuel nitrogen and half is due to thermal N0 X since NO emissions firing this mixture were about twice those observed firing a nitrogen-free refinery gas. The estimated fuel nitrogen conversion efficiency, based on the ratio of oil to gas in the fuel and the oil fuel nitrogen content (-0.8%), is approximately 19 percent. Figure 10 shows that NO x emission decreased slightly as staging height increased firing oil with gas. Very little decrease was observed at heights greater than 1.2 m (4 feet). During all testing with the staged combustion system operating, careful observation of the flame and furnace draft was made. There did not appear to be any problems with coking of the process tubes and at no time was there any emission of carbon monoxide or unburned hydrocarbons even at 2 percent stack excess oxygen. There were certain instances in which the draft in the convec- tion became slightly positive at the low 0 2 condition with maximum staging, however, flashback was never observed. The long term effect of this change in furnace draft in terms of maintenance or operational costs is not yet known. COMPARISON OF PRESENT RESULTS WITH SUBSCALE RESULTS Table 3 summarizes the N0 X reductions obtained at the present full scale process heater with the emission reductions observed at the subscale test site. For gaseous fuels the trends are the same for both subscale and full scale tests except that the percentages are somewhat lower in the present data. For tests with oil fuels lowered excess air alone was more effective in reducing NO emission than was staged combustion alone - unlike the trend observed in the subscale test where staging the combustion gave a significantly greater NO reduction than low excess air. Also, the percentages are much lower for the full scale heater than they were for the subscale heater firing No. 6 oil alone. The reasons for the lower percentage NOx reductions firing gas fuel only in the full scale heater are not altogether clear. Furnace bulk gas and wall temperatures were much lower in the full scale heater than they were in the subscale unit (~700°F vs. 1700°F). longer in the full scale heater. Residence times, however, were probably The baseline NO emission with the same 3-4 P-263 burner type in the subscale heater firing natural gas was 131 ppm, dry, corrected to 3 percent 0 2 compared with the present 105 ppm, which is reasonable in light of the temperature differences mentioned. There may be a decrease in the effectiveness of combustion modifications in reducing thermal NO x as temperatures are decreased. The mixing of the combustion air and fuel nay have been different enough to cause the changes in NOx reductions from subscale to full scale. The precise location, angle of injection, and spray angle associated with the staged air system are probably also important parameters in the N0 X reduction process. Unfortunately, direct comparisons of the oil-firing results are not valid since the full scale unit was firing a considerable amount of gas along with the residual oil. In the subscale tests no gas was fired with the oil. Although the fuel nitrogen in the subscale tests was much lower (~0.3%) than it was in the full scale tests, the apparent fuel nitrogen conversion in the former tests was about twice that of the latter. The fractions of thermal and fuel NOx are thought to be about the same in either case-about 50-50. Further work is planned to determine whether or not the performance of these combustion modifications can be improved for oil-firing applications. 3-5 P-263 SECTION 4 COST ANALYSIS In this paper we will concentrate on the cost analysis of lowered excess air and staged combustion air applied separately and in combination to a process heater with a thermal input capacity of 16.1 MW (55 x 106 Btu/h). will consider only the gaseous fuel application. We Costs for oil-fired units will be determined pending the results of further experimentation. INITIAL CAPITAL COSTS There is no initial capital cost associated with the lowered excess air (LEA) modification at the levels of 0 2 used in the present tests. Such costs would be incurred if operation at stack oxygen contents of less than 2 percent were desired since an automatic oxygen control and analyzer would then be required. The initial capital costs of a staged air system such as that of Figure 2 are given in Table 4. The installed costs include direct labor and overhead charges at contract labor rates. Required equipment and materials include a high pressure blower, a damper, piping, fittings, valves, stainless steel lances for insertion into the firebox, and temperature and flow measurement probes. Shipping costs are also included. The annual operating costs for combustion modifications to a full scale heater are given in Table 5. The fuel savings were based on the efficiency increases for the various modifications shown in Table 2 and a market price of natural gas of $2.50/106 Btu. (The latter assumption may not be valid in a 4-1 P-263 case where the refinery uses an off-gas which is not readily marketable and would otherwise be flared. In that case, there would be no cost savings associated with the increased efficiency) . In determining the total annualized costs of combustion modifications several assumptions regarding state and federal taxes, insurance, depreciation method, and financing constraints have been made. These are listed in Table 6 and the total annualized costs for the process heater modifications are given in Table 7• Note that there are two columns for the staged combustion air (SCA) modification only. The one labelled "4% 0 2 " applies to the SCA modification at a normal operating condition which results in 4 percent 0 2 at the stack. The fuel savings and annualized costs are calculated relative to those of a heater having a 4 percent 0 2 condition without staging. For the column labelled "2% 0 2 " the fuel savings and annualized costs are calculated relative to those of a heater having a normal operating condition of 2 percent 0 2 at the stack with no staging. For those refinery heaters which normally operate at -2% 0 2 the column "SCA Qnly-2% 0 2 " applies. For those which normally operate at -4% 0 2 the column "SCA 0nly-4% 0 2 " applied. The "SCA + LEA" column shows fuel savings and annualized costs relative to a 4 percent 0 2 normal operating condition without staging. One may ask why all process heaters do not operate at 2 percent 0 2 rather than 4 percent 0 2 since it appears to be cost effective to do so. Refinery gas varies significantly in composition and thus requires a varying amount of combustion air. To avoid the possibility of air deficiency and to minimize the need for continuous operator attention, many process heaters are operated at a higher excess oxygen than would be necessary with a more consistent fuel. With the current need to conserve energy, this practice is being revised by installing oxygen controllers and increasing operator attention. Table 8 gives the annual N0 X emission reductions and cost effectiveness of each combustion modification. The annual N0 X emission reduction is calculated using the values given in Table 2 for percent N0 X reduction and assuming continual operation at 70 percent capacity for 80 percent of the year. The 4-2 P-263 3 cost effectiveness is simply the total annualized costs divided by the annual N0 X reduction capability. Table 8 shows that for a gas-fired heater which normally operates at 4 percent excess oxygen the combination of staged combustion air and lowered excess air is an economically and environmentally attractive modification. Lowered excess air alone is very economical but is only 1/3 to 1/4 as effective in reducing N O . For a heater which normally operates at 2 percent 0 2 the costs of staged combustion air would not be greatly different from those arising from the use of SCA alone at 4 percent 0 2 . This cost is, however, much greater than the cost of SCA + LEA. The question of which modification is most cost effective in a given instance will depend on each individual plant situation and each individual heater. O Each plant must determine the following: 1) the credit, if any, it should allow for fuel savings, 2) the normal excess oxygen in the heater, and 3) whether or not a lower excess oxygen can be obtained without additional cost. J 4-3 P-263 3 SECTION 5 SUMMARY Combustion modifications proved effective on a natural draft process heater rated at 108 m3/h (16,250 barrels/day) of crude capacity when firing refinery gas. In addition, heater performance was measured for other fuel mixtures including natural gas and No. 6 oil. While firing refinery gas a maximum N0 X emission reduction of 51 percent was observed, below a baseline emission concentration of 105 ppm, dry corrected to 3 percent 02» An increase in efficiency of 2.37 percent was observed for this condition and, in general, efficiency was improved by the application of combustion modifications. Staged combustion air gave a significant N0 X reduction when firing refinery gas regardless of excess oxygen level. Staging height did not have a major effect on N0 y emissions over the range of 0.3 to 2.4 meters. The NO emissions occurring for the various modification techniques followed the trends observed at the subscale level, however, the percentage reductions were slightly less. In absolute magnitude, the NOx emission reduc- tions occurring for gas-firing were approximately equal to those occurring when No. 6 oil was included in the fuel. The percentage reductions were less, however, for oil apparently because of the fuel nitrogen conversion. J 5-1 P-263 A cost analysis covering initial capital and annual operating costs, total annualized costs, and cost effectiveness of each of the combustion modifications was made. The results indicate that staged combustion air may be an economically and environmentally desirable modification, however, the cost effectiveness for any single application will ultimately depend upon several factors unique to that situation. 5-2 P-263 LCCATION 7 - NATURAL DRAFT REFINERY PROCESS HEATER i \ C CD 1 m JE ti STACK 4 , -6 M O.D. (1.4m) • • o ^ CONVECTION SECTI0: v E 1- n NOT TO SCALE I » «-H r - RADIANT SECTION 16,-94" O.D. (5.1m) 15,-9i" I.D. (4.8m) AIR PLENUM AND SOUND SUPPRESSION BOX O E I I CD Figure 1. Sketch of the process heater at Location 7 P-263 c:TATJn A l R ^ T F M l <®S> 3 PVC BALL VALVE Itf XflfUtLJ -»' D 1(7 OUTLPT &PVC *ou: 8F4SADDIE 'PVCITO CONNECT PP\CTO 4*3 P\C FLUSHING < -PI TOT MI-TUBE [ T tt-»Ei JPVC = = A •I VClOntXlBLf TUBIM6 T» M roc nucfn 'BALI VAIvr = -scntou -ff PVC f>- it I Vi -L'-J :•» Y1 -JPVC -JPvC DETAIL A SIDE V I E W TYPICAL I t I K) Figure 2. n Flow Schematic of Staged Combustion Air System for a Natural Draft Process Heater. **- DETAIL B TOP VIEW TYPICAL Hot Sample Lina © Not Pump Hot Pump Praaaura VACUUM 0 o Haatad Llna Hot Pump HOTBOX Dry tampla Lines (Typical Sat-Up 81a LlnaaJ. Condanaar . I* Piltar -tj [ Hot/coia * Switch Plowmatara Id .Hl-Span SarolSpai *|(Dry Sampjta) tarot&ISpan tfti |x«ro| Lo-Span ftafrlqaratlon Condanaar «h | Manifold ' > NC 80. 1 M)sampla Praaaura Praaaur 1 .»- Vant terotJfaspan Q MO. taro tjfcspan Vacuum Pump KVB P-263 Figure 3. Flue gas sampling and analyzing system. -© N O emissions at three different loads as a function of stack excess oxygen for a natural draft process heater firing a natural gas/refinery gas mixture. •a E Q. Q. 6 z o D A Fuel: Natural Gas + Fuel Gas 18.2 kg/s of crude (11,000 bbl/day) 23.1 kg/s (14,000 bbl/day) 13.2 kg/s (8,000 bbl/day) Stack Excess Oxygen, %, dry Figure 4. KVB P-263 N O emissions at three different loads as a function of stack excess oxygen for a natural draft process heater. 300 8 STACK EXCESS OXYGEN, %, DRY Figure 5. KVB P-263 N o emissions as a function of burner equivalence ratio at two excess oxygen levels with constant staged air insertion height. ( 120 -I 1 r- T T T Fuel: Adsorber Gas Staging Height: 4' (1.2m) Load: 9600 bbl/d (15.9 kg/s) 04% O, 100 • 7/12-1 2 % 03 90 7/12-10 80 70 7/12-5 7/12-6 7 12-7 60 7/12-8 7/12-9 50 40 t 0.2 0.4 0.6 0.8 1.0 1.2 <HB = (A/F) ACT/(A/F) STOIC. Figure 6. KVB P-263 NOemissions as a function of burner equivalence ratio at two excess oxygen levels with constant staged air insertion height. 120 1 1 1 110 Fuel: Adsorber Gas Staging Height: 4 ft. (1.2m) Load: 11,500 bbl/d (18.9 kg/s) 7/9-1 O 4% Q, r-j 2 % O2 100 7/9-2 90 7/9-5 80 7/9-4 70 60 50 40 i 0.2 0.4 0.6 0.8 1.0 1.2 ft Figure 7. KVB P-263 N O emissions as a function of staging height at constant excess oxygen and 6 . D 120 100 r- 7/13-1 a 80 - 7/13-4 7/13-3 7/13-2 -O-O 7/13-5 60 40 20 Fuel: Adsorber Gas Load: 11,500 bbl/day (18.9 kg/s) 4 % 02, © = 1.00 -B 1 I 1 1 2 3 Staging Height, ft. Figure 8. KVB P-263 i . NOemissions as a function of burner equivalence ratio for a gas-oil fuel mixture. 220 210 T - T T T T Load: 11,500 bb!/d (19.0 kg/s) Excess 0 2 : 2 % Staging Height: 4 ft (1.2m) Fuel: 4 7 % Adsorber Gas + 5 3 % No. 6 Oil '0 2 = 4 % ' 7/18-8 Baseline 200 •o E" 190 - CL Q. 6 z 180 7/18-4 7/18-7 (No staging) 170 165 I Mr 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 0B Figure 9. KVB P-263 N O emission as a function of staging height for gas-oil fuel mixture. 220 Load: 11,500 bbl/d (19.0 kg/s) Excess O?: 4 % 0 B : 0.9 Fuel: «~75% Adsorber Gas + ~ 2 5 % No. 6 Oil 210 6 CO t3 CO 200 - •o 7/15-9 E* CL CL 190 - O Z 180 - 1 (0.30) 2 (0.61) 3 (0.91) 4 (1.22) 5 (1.52) 6 (1.83) 7 (2.13) 8 (2.44) Staging Height, ft (m) Figure 10. o KVB P-263 ^ ) TABLE 1. Species EMISSION MEASUREMENT INSTRUMENTATION Manufacturer Measurement Method Hydrocarbon Beckman Instruments Flame Ionization 402 Carbon Monoxide Beckman Instruments IR Spectrometer 665 Oxygen Teledyne Polarographic Carbon Dioxide Beckman Instruments IR Spectrometer 864 Nitrogen Oxides Thermo Electron Co. Chemiluminescent 10A Particulates Joy Manufacturing Co. EPA Method 5 Train EPA Sulfur Dioxide DuPont Instruments UV Spectrometer 400 Sulfur Oxides KVB Equipment Co. Controlled Conden. Smoke Spot Bacharach ASTM D2156-65 Particulate Sizing Anderson 2000, Inc. Cascade Impactor 32 5A 2V -7006 Mark III TABLE 2. COMPARISON OF NOx REDUCTION AND EFFICIENCY CHANGE IN FULL-SCALE PROCESS HEATER FIRING REFINERY GAS Test Condition N0 X Reduction Lowered Excess Air (4% 0 2 • 2% 0 2 ) 15% •2.15% Staged Combustion Air (SCA) (4% 0 2 ) 35% •0.09% SCA at 2% 02 43% •0.22%1 SCA • LEA 51%' •2.37%' 1 relative to 2% 0 2 baseline level (no staging) relative to 4% 0 2 baseline level (no staging) 2 TABLE 3. COMPARISON OF N0 X REDUCTION OBTAINED IN SUBSCALE TEST WITH FULL-SCALE TEST Subscale Test Condition M " Sctle No. 6 Oil "? Natural No. 6 Oil Adsorber 4 (0.8% N) Gas (0.3% N) Gas Adsorber Gas 10% 15% 16% Staged Combustion Air (SCA) 46% 35% 35% 7% SCA • LEA *7% 51% 51% 24% Lowered Excess Air (LEA) S/)Sf- JLwe 27% / O^&r" sosFp A/OX 9^ S/^tr A~i*- tf0 M-Wxi/iefirtcr- CO^UC TABLE 4. INITIAL CAPITAL COST OF STAGED COMBUSTION MODIFICATION FOR A NATURAL DRAFT PROCESS HEATER (1980 $) RATED AT 16.1 MW THERMAL INPUT Installed Cost of Equipment Engineering Design $24,385 6,400 Total $30,785 TABLE 5. ANNUAL OPERATING COST OF COMBUSTION MODIFICATIONS FOR A NATURAL DRAFT PROCESS HEATER (1980$) LEA Only SCA Only (4% 0 2 ) SCA Only (2% 0 2 ) SCA • LEA Fan Electricity 0 700 700 700 Maintenance on Fan and Staged Air Pipes 0 4000 4000 4000 Fuel Costs (Savings) (14,749) (617) (1,509) (16,259) $(14,749) $4,083 $3,191 $11,559 Total TABLE 6. ASSUMPTIONS MADE IN DETERMINING TOTAL ANNUALIZED COSTS 1. State and Federal Property Taxes » 11% of Initial Capital Cost 2. Insurance Charges - 0.5% of Initial Capital Cost 3. Straight-line Depreciation Over 12 Tears 4. State and Federal Income Taxes * 50% 5. Investment Tax Credit (First Year Only) • 10% 6. 100% Equity Financing (Debt/Equity • 0) 7. 15% Rate of Return 8. Capital Recovery Factor » .2773 TABLE 7. TOTAL ANNUALIZED COSTS OF COMBUSTION MODIFICATIONS FOR A NATURAL DRAFT PROCESS HEATER LEA Only Annual Capital Charge, $/y SCA Only (4% 0 2 ) SCA Only (2% 0 2 ) SCA • LEA 8,537 8,537 8,537 Annual Expenses (Savings), $/y (14,749) 7,623 6,731 (8018) Total Annualized Cost, $/y (14,749) 16,160 15.268 519 TABLE 8. ANNUAL NOx EMISSION REDUCTION CAPABILITY AND COST EFFECTIVENESS OF COMBUSTION MODIFICATIONS LEA Only SCA Only SCA Only SCA • LEA (4% 0 2 ) Annual NO Emission Reduction (Mg NOj/y? 2.34 Cost Effectiveness Savings of 2960 2679 64 ($/Mg NO Reduction) $6303/Mg 5.46 (2% 0 2 ) 5.70 8.11