By Justin Bennett
Many utilities struggle to balance pipe-pipe fuel flow distributions as part of a low-NOx burner replacement or as part of combustion optimization. It has often been presented that balanced pipe-pipe fuel flow distributions are the only path to uniform burner–burner stoichiometries. There is a common belief that balancing the pipe-pipe coal flow distributions leaving each pulverizer will directly lead to uniform combustion. Combustion uniformity is the key to the lowest emissions as well as the best boiler efficiency. However, this belief assumes there is adequate secondary airflow to maintain uniform burner-burner stoichiometries. In essence, it is the uniformity of the burner-burner stoichiometries and not balanced pipe-pipe fuel flow distributions that dictate the combustion uniformity.
Achieving uniform burner-burner stoichiometries involves matching with available fuel and air flows. Often there are limitations within the windbox due to either the original design or modifications that do not permit equal quantities of secondary airflows to each individual burner. Therefore it is possible for some burners to be more sub-stoichiometric than others even with balanced pipe-pipe fuel flow distributions and air registers at 100% open. This situation can often lead to increased carbon monoxide, opacity and flyash loss on ignition emissions as well as localized furnace slagging and other operational concerns.
These concerns were challenges for one of SAS Global Corporation’s clients. The unit is a B&W wall-fired design. It has four (4) EL-64 mills, and each mill has three (3) burner lines. It is capable of approximately 95 MW net while firing an eastern fuel. The unit is air limited due to limitations with its induced draft (ID) fan during warm weather. The unit has twelve ACT-modified burners and an overfire air (OFA) system.
SAS Global Corporation (SAS) was contracted to improve the unit’s combustion, lower its CO emissions and reduce flyash loss on ignition (LOI) while not increasing NOx emissions. SAS applied its Total Solutions Approach (TSA) to combustion optimization. The SAS’ TSA involves three phases. Phase I is an as-found assessment of the present combustion characteristics including pulverizer performance.; Phase II involves installing equipment and upgrades (as required) backed by a 100% customer satisfaction guarantee; and Phase III is the optimization of the combustion with the equipment.
The results from Phase I indicated significant pipe-pipe fuel flow distributions as seen in Table 1. In addition, high temperature video recordings of the burner flames indicated significant non-uniform combustion with coal streaking and uneven ignition points as seen in Figure 1.
The burner in this figure represented two distinct characteristics. First, the side closet to the camera has a pronounced coal streak. This is a fuel rich and air lean condition, which is producing much more CO and LOI than would be seen with a well-mixed flame. Second, the opposite side is fuel lean and air rich, which is producing more NOx than would be realized with a well-mixed flame. In essence, this is worst case for any burner.
Based upon these Phase I results and observations, it was decided to install the patented SAS V-Style Adjustable Orifices and patented In-Line Diffusers (ILDs). The V-Style Orifices allowed the tailoring of the pipe-pipe fuel flow distributions with the mills in-service. Further, the results of these adjustments could be seen in immediately using the MIC-One, a real-time, microwave based, portable coal flow measurement system.
The ILDs promote increased fuel and secondary airflow mixing as they break up any heavy coal concentrations (roping or streaking) as the fuel enters the burner. Other benefits of the SAS’ ILDs include more uniform wear characteristics of internal burner equipment, such as the burner barrels. The increased mixing allows for more uniform combustion with lower excess oxygen levels for lower NOx emissions.
Additional components for Phase II included the optimization of the operating primary air flow curves and adjusting the classifiers for proper fineness of approximately 99.0% to 99.5% passing 50 mesh and approximately 72.0% to 75.0% passing 200 mesh.
Table 2 presents the results of the balanced pipe-pipe fuel flow deviations using only the SAS’ V-Style Orifices. The results indicate that the overall deviations were greatly reduced to within approximately ±6.3% from approximately ±33.1% for Mill 4C. More importantly was the significant reduction in the column-column fuel flow deviations with all deviations within ±1.5% from the as-found of approximately ±24.4%.
There was also a remarked improvement in the appearance of the flames as seen in Figure 2. This example is typical of the other burners, and it presents as a well-mixed plume without any coal streaks and uniform ignition within the burner throat.
Once the SAS equipment was installed, Phase III (combustion optimization) began. The combustion optimization focused on balancing the individual burners’ stoichiometries using only the individual burner’s single air register. Figure 3 presents the as-found emission contour plots for this unit.
In this figure the economizer exit plane is replicated three times with the concentrations for excess oxygen (top) carbon monoxide (middle) and nitrogen oxides (bottom) displayed as contours. The goal is to reduce the contours for each plot to as few as possible since fewer contours translate into better combustion uniformity. The results indicate that the carbon monoxide (CO) emissions are concentrated along the South side with corresponding lower excess oxygen (O2) and nitrogen oxides (NOx) in the same areas, Therefore, the South area is fuel rich and air lean while the North part of the furnace is more fuel lean and air rich.
A series of combustion diagnostic and optimization tests followed. These test were performed using the proprietary SAS Multi-Stream Combustion Diagnostic Analyzer (MSCA). The MSCA has the ability to sample up to eighteen (18) independent flue gas streams simultaneously and in real time. The data and results are logged on a laptop computer, and the software corrects for dilution. It is possible to investigate many cause-and-effect relationships much more quickly with the SAS’ MSCA than traditional point-to-point sampling.
As an example, consider a test grid of twenty-four sample points. A traditional point-to-point approach would require at least 120 minutes (2 hours) to obtain five minutes of data per point. The MSCA would perform the same sampling in only 10 minutes. This short duration helps ensure that all data are representative of the actual test condition since it is not uncommon for a furnace’s combustion and emission to change within two hours due to slagging and other heat transfer changes.
The results of the optimization efforts with balanced pipe-pipe fuel flow distributions are presented in Figure 4. The results indicate that the contours have decreased, but a “bull’s eye” is still present on the CO and O2 contours. This “bull’s eye” represents one single burner producing CO emissions in excess of 1400 ppmc. The corresponding burner register was set at 100% open. Hence, no more secondary airflow could be added or forced to this burner by closing other burner air registers.
The only remaining avenue to optimize the combustion was to change the pipe-pipe fuel flow distributions for the lower two firing elevations, which are for Mills 4A and 4D. The approach was to slowly adjust one V-Style Orifice at a time, and then optimize the burner’s stoichiometry with revised air register changes. This process was iterative but effective as seen in Figure 5.
Figure 5 presents the results with many fewer contours, but a CO pocket is still present at the same relative location. The V-Style Orifices could not be closed any further than the values in Table 3 due to minimum velocity constraints within the burner lines.
In summary, this case history indicates that it is more important to optimize combustion based upon local burner stoichiometries. The data and results indicate that it may be necessary to degrade pipe-pipe fuel flow deviations in order to optimize combustion with apparent windbox limitations. These efforts have allowed this unit to produce approximately 4-5 more megawatts with reduced opacity, significantly lower LOI (approximately 19% as-found versus approximately 6-8% as-left) and much lower CO (480ppmc versus 108ppmc) at comparable NOx levels. The SAS’ TSA has also been applied to the other units at this station with similar or better results.