F

Improving biomass boiler performance and emissions

uel delivery is a common problem in biomass plants. Wood fuel is non-uniform in shape, density, moisture content, ash content, and energy content. Over-sized or stringy pieces of wood fuel will cause bridging and wet fuel sticks to the sides of fuel chutes, often causing plugs. Fuel delivery problems result in decreased efficiency, boiler upsets, lower steam production and emissions violations. The fact that wood fuel is so problematic is the justification many plants use for resorting to manual fuel control. An operator running a boiler in manual mode will make a fuel adjustment once every five minutes. More often than not, the adjustment is excessive, resulting in predictable system oscillations which negatively affect steam production, fuel efficiency and emissions. For the plants that use automated fuel control, most use a controller based on drum pressure. While a pressurebased controller is likely to have fewer oscillations than would be caused by an operator1, it is inherently slower and less accurate than a controller which uses O2 as a process variable. An alternative to the pressure-based controller is a fuel delivery system based on furnace O2. An O2 controller makes fuel adjustments according to the amount of excess O2 in the flue gas. Instead of trying to directly stabilise drum pressure, it stabilises combustion which results in stable drum pressure. As excess O2 can be used as a leading indicator of drum pressure, an O2 controller can prevent drum pressure errors by adjusting the fuel rate before the error occurs. Nearly all boiler upsets and emissions problems can be traced back to swings in O2. An O2-based fuel controller will stabilise the combustion process, resulting in increased steam flow, reduced average O2 (which leads to greater efficiency), stable temperatures and reduced emissions violations. A well-designed controller will pay for itself within a year. The importance of excess air as a control parameter Air is one of the most critical parameters for attaining good combustion and boiler efficiency. Too little air results in spikes in CO and other unburnt combustibles; too much air reduces boiler efficiency and increases stack losses. Excess O2 is important because it is nearly an exact indication of excess air. [Stultz] Figure 1 shows a set of characteristic curves for a theoretical boiler. The curves represent the effect swings in O2 have on load, efficiency, furnace temperature and the production of CO and NOx. Swings in O2 have a predictable and detrimental effect on efficiency and emissions. NOx increases with O2. Load, efficiency and temperature decrease with O2 and CO increases for variations away from the corner point. For conditions of low O2, CO increases asymptotically. The corner point shown in Figure 1 is the ideal O2 operating point. It is the balance point where boiler efficiency and control of emissions are optimised. Every boiler has a unique corner point. The controller concept The role of the O2 controller is to stabilise the combustion process around the boilers corner point. Given a fixed amount of air flow into the furnace, the controller varies the fuel rate in order to maintain an O2 set point. The amount of air flow is selected by the operator based on the desired steam flow or MW output. If the O2 is consistent, then drum pressure, steam flow and furnace temperature will also be consistent. As the combustion air is constant, the only reason the excess O2 in the flue gas will increase is because the amount of O2 being consumed in the combustion process has decreased. The possible causes of this condition are a decrease in screw speed, fuel piling, fuel bridging or a negative change in fuel quality and/ or moisture. Increasing O2 is a leading indication of decreasing drum pressure and steam production.

biomass boilers Bioenergy

Figure 1: Effects of O2 swings on CO, NOx, temperature and efficiency

Bioenergy Insight

September/October 2013 61

Bioenergy biomass boilers

Furnace temperature varies inversely with O2. A decrease in amount of excess O2 in the flue gas will occur if the amount of O2 being consumed has increased. Decreases are caused by a rise in fuel screw speed or a positive change in fuel quality and/or moisture. Decreasing O2 is an indication of increasing drum pressure and steam production. A decrease in O2 is usually accompanied by an increase in furnace temperature. Unlike an operator, an O2-based fuel controller can continuously monitor the combustion process, making several small adjustments
160 140 Percent of Average Value

O2 and Fuel Variations - Manual Control

DECREASE IN STEAM REVENUE

160 140 Percent of Average Value 120 100 80 60

O2 and Fuel Variations - Auto Control

120

100 80 60 40
INCREASED CO LEVELS POSSIBLE EMISSIONS VIOLATIONS

O2%
FEED%

O2%
FEED%

10:55 10:59 11:03 11:07 11:11 11:15 11:19 11:23 11:27 11:31 11:35 11:39 11:43 11:47 11:51 11:55 11:59 12:03 12:07 12:11 12:15 12:19 12:23 12:27 12:31

Figure 3: O2 and fuel variations

data were collected in one minute increments. The data presented here are from June 2010. Steam flow, drum pressure, O2, feed rate and control mode

Figure 2a shows that changes in drum pressure follow the changes in screw speed; the pressure response lags the change in fuel rate by about six

160 140 Percent of Average Value

Pressure and Fuel Variations - Manual Control

160 140 Percent of Average Value 120 100 80 60 40

Pressure and Fuel Variations - Auto Control

120

100 80 60 40

PRES%
FEED%

10:55 10:59 11:03 11:07 11:11 11:15 11:19 11:23 11:27 11:31 11:35 11:39 11:43 11:47 11:51 11:55 11:59 12:03 12:07 12:11 12:15 12:19 12:23 12:27 12:31

Figure 2: Drum pressure and fuel variations

in screw speed every few seconds. By relating changes in O2 to changes in screw speed, the controller can establish a baseline fuel grade and can therefore detect subtle changes in fuel quality and make adjustments accordingly. Although the controller cannot discriminate between fuel bridging and fuel piling, it can detect that one of these two conditions is present and alert an operator to impending drops in drum pressure or combustion conditions which might lead to emissions violations. Case study An O2 controller was installed in a cogeneration plant in 2010 and process

(auto or manual) were collected. Variations in screw speed, pressure, steam flow and O2 are represented as a percentage of their respective averages. Results Figure 2a shows 100 minutes of variations in drum pressure and metering screw speeds while operating in manual mode. The metering screws are controlled by variable frequency drives (VFDs); screw speeds are entered by the operator in Hertz. In 100 minutes, the operator made 24 changes ranging in magnitude from -15 to 10 Hz with the average adjustment magnitude being 5.08 Hz. The average VFD speed is 36.05 Hz.

62 September/October 2013

7:22 7:26 7:30 7:34 7:38 7:42 7:46 7:50 7:54 7:58 8:02 8:06 8:10 8:14 8:18 8:22 8:26 8:30 8:34 8:38 8:42 8:46 8:50 8:54 8:58

minutes. The average pressure during this period is 404 psi. The standard deviation of the pressure is 20.66 psi. This variation represents 5.1% of average. The pressure rating for the boiler is 400 psi. The pressure swings represented in Figure 2a range from 354 to 440 psi. The upper end of these oscillations is either near, or in the range of, the high drum pressure alarm set point. If the operator pushed the boiler any harder without reducing the swings, the tops of the oscillations would trip the boiler. Figure 2b shows 100 minutes of variations in drum pressure and metering screw speeds while operating in auto mode. In auto mode the fuel controller is

7:22 7:26 7:30 7:34 7:38 7:42 7:46 7:50 7:54 7:58 8:02 8:06 8:10 8:14 8:18 8:22 8:26 8:30 8:34 8:38 8:42 8:46 8:50 8:54 8:58

40

PRES%
FEED%

responsible for making the changes to the screw speeds. The controller made more changes in the same time period, and the magnitude of changes was much smaller than in manual mode. In 100 minutes the controller made 80 changes ranging in magnitude from -3.5 to 2 Hz, with the average adjustment at 1.18 Hz2. The average VFD speed is 33.24 Hz. The pressure swings shown in Figure 2b range from 391 to 416 psi. The average pressure is 402 psi with the standard deviation of 5.02 psi. This is slightly more than 1% variance from average. Figure 3a shows variations in O2 and metering screw speeds for manual mode. Note that the O2 changes inversely as the metering screw speeds. The O2 lags the changes in fuel rate by about one minute, which means O2 leads pressure by about five minutes. The range of O2 shown is 2.73% to 7.94%; the average O2 is 5.73%. The standard deviation is 1.07%. Figure 3b shows variations in O2 and metering screw speeds while in auto mode. The controller set point for this time period was 4.97%. The range of O2 shown is 3.84 to 6.09% with an average of 4.97%. Figure 4a shows variations in steam flow and metering screw speeds for manual mode. Changes in steam flow lag the changes in fuel by about five to six minutes, which is the same as the pressure. The correlation

Bioenergy Insight

biomass boilers Bioenergy

160 140 Percent of Average Value 120 100 80

Steam and Fuel Variations - Manual Control

160 140 Percent of Average Value 120 100 80 60 40

Steam and Fuel Variations - Auto Control

STEAM%
FEED%

STEAM%
FEED%

60
40

increase by approximately 21C for each per cent decrease in O2. The swings in O2 shown in Figure 3a have a 5% variation which suggests a temperature swing larger than the effective temperature window of the SNCR control. Conclusion

10:55 11:00 11:05 11:10 11:15 11:20 11:25 11:30 11:35 11:40 11:45 11:50 11:55 12:00 12:05 12:10 12:15 12:20 12:25 12:30

7:22 7:26 7:30 7:34 7:38 7:42 7:46 7:50 7:54 7:58 8:02 8:06 8:10 8:14 8:18 8:22 8:26 8:30 8:34 8:38 8:42 8:46 8:50 8:54 8:58

Figure 4: Steam flow and fuel variations

between oscillations in steam flow and fuel delivery is apparent. The average steam flow shown in Figure 4a is 58,595lbs/hour. The standard deviation is 3,674.4lbs/ hr. This represents a 6.3% variance from average. Figure 4b shows variations in steam and metering screw speeds for auto mode. The average steam flow shown in Figure 4b is 67,129lbs/hr with a standard deviation of 698.9lbs/hr. This represents approximately 1% variance from average. Effect of swings on performance The difference in drum pressure trends in Figures 2a and b is significant. For manual fuel control the pressure trend is almost sinusoidal. In auto control there are still some pressure swings, but the frequency is much lower and the magnitude smaller. The most obvious effect of auto control is stability. The three spikes in O2 represented in Figure 3a can be directly correlated to decreased steam production. The average O2 in auto mode was 4.97%; in manual mode it was 5.73%. This difference of 0.76% corresponds to an increased load on the boiler fans, increased stack losses and decreased furnace temperature. The increased stability is also reflected in steam flow. The average steam production for the manual control data

was 58,595lbs/hr; the average steam production for the auto control data was 67,129lbs/ hr. This is an increase of 14.6%, or approximately 1MW increase in electrical production. At $0.06 (0.05) per kWh, this represents a potential increase in revenue of $480,000/year.

By contrast, the emissions based on the manual mode data in Figure 3a were likely to be very significant. The three pronounced dips in O2 represented in Figure 3a can be directly correlated to spikes in CO. This in particular is important because maximum achievable

An O2 controller makes fuel adjustments according to the amount of excess O2 in the ue gas. It can prevent drum pressure errors by adjusting the fuel rate before the error occurs

In order to optimise boiler efficiency and emissions control, the combustion process must be stable. The benefits of stable combustion are increased steam flow, increased efficiency, and the avoided costs associated with emissions violations. An O2-based fuel controller is designed to stabilise combustion. While a pressure-controlled system may be better than a manual system in some respects, the pressurecontrolled system is designed to correct an error in drum pressure only after it has already happened. In contrast, an O2 controller can predict and prevent the error from happening in the first place. Furthermore, a pressure-controlled system has no means of detecting problems in combustion such as bridging and/or piling. l

The data for these two examples were taken eight hours apart and no information is known about the possible differences in fuel quality. While improved stability will result in an increase in steam production, it cannot be stated with certainty that the improved boiler stability is responsible for the entire increase in steam production. Effect of swings on emissions For the auto mode data represented in Figure 3b the CO and NOx emissions were likely to be negligible or non-existent. The O2 was consistent, staying within 1% of the set point.

control technology (MACT) uses CO as a surrogate VOC as an indication that other hazardous air pollutants (HAPs) are present. The O2 spikes in Figure 3a can be related to increases in NOx production. The three dips in O2 can also be correlated to increases in furnace temperature and decreases in NOx control efficiency. Selective non-catalytic reduction (SNCR) is temperature dependent, having an effective window of approximately 777C to 902C for a O2 concentration of 5%3. SNCR control loses its effectiveness at temperatures above 955C. Furnace temperatures

Footnotes and references:

1 For poor fuel conditions a pressure-based controller may be less effective than manual control 2 It is likely that the controller is actually making more than 1,000 speed changes in 100 minutes, however the slow data collection is unable to accurately represent all the changes. 3 Yonghun Park and Jerald Caton, The Use of Urea for Selective Non-Catalytic Removal (SNCR) of Nitric Oxides: Laboratory Biomass Boiler Performance and Emissions

For more information:

This article was written by Andrew Gentile, P.E. and Sheldon Schultz, P.E. of Yanke Energy, www.yanke-energy.com

Bioenergy Insight

September/October 2013 63