Air Pollution Control Equipment Services, Emissions, Gas, Policy & Regulations

Reduce CO2 Emissions and Boost Generation Efficiency

Issue 7 and Volume 115.

Combustion monitoring and optimization help reduce emissions and improve generation efficiency

By Scott Affelt, Zolo Technologies

Improving power generation efficiency is becoming more important for coal-fired power plants throughout the U.S. Resource conservation, energy independence and climate change awareness incentivize owners and operators to improve efficiencies, reduce heat rates, save fuel and reduce emissions. Advanced combustion monitoring and optimization tools combined with plant personnel training focused on generation efficiency best practices are key to achieving these goals.

In 2010, the U.S. Department of Energy’s National Energy Technology Laboratory (NETL) sponsored demonstration projects to reduce CO2/MWh using generation efficiency improvements through combustion balancing at two coal-fired power plants1. The projects, conducted at American Electric Power’s (AEP’s) John Amos Unit 3 and Detroit Edison’s (DTE’s) Belle River Unit 2 and led by Zolo Technologies, used the ZoloBOSS real-time combustion monitor to balance combustion coupled with closed-loop combustion optimization software to sustain improvements.

In February 2010, NETL conducted a workshop, “Improving the Thermal Efficiency of Existing Coal-Fired Power Plants in the United States,”2 where utility owners, equipment vendors, industry experts and research organizations convened to identify technical opportunties for increasing the thermal efficiency of coal-fired power plants as a means to lower green house gases. Several of the key opportunities identified were addressed in the demonstration projects at DTE and AEP, particularly using real-time combustion monitoring and optimization processes with advanced tools and instrumentation. In addition, identification of a dedicated performance engineer and training of plant personnel on efficiency best practices were applied during the demonstration projects. The inter-realtionships between the project components are shown in Figure 1.

Real-time in-furnace combustion monitoring provides enhanced visibility of the combustion process enabling operators to properly balance combustion (temperature, O2 and CO) and therefore safely operate at reduced levels of excess air without the typical adverse impacts of slag and corrosion. By contrast, traditional economizer oxygen probes, positioned downstream of the combustion process, only provide a point source measurement and require independent calibration and tuning of the boiler several times per year using multi-depth extractive gas sampling. Even after tuning is complete, natural variations in combustion performance can lead to lcoalized areas in the boiler that are O2 rich and O2 depleted. This can result in higher localized NOX levels in the O2 rich regions and high CO in the O2 lean regions. The operational tradeoffs of air-to-fuel ratios are illustrated in Figure 2.

Project Components

Three key components were central to the successful execution of the overall demonstration project – the in-furnace combustion monitor to balance combustion, combustion optimization software to sustain improvements and plant personnel training.

In-Furnace Combustion Monitor: At each plant, a Zolo Technologies ZoloBOSS combustion monitor was used to obtain real-time measurement of temperature, O2, CO and H2O concentrations directly in the furnace. The ZoloBOSS is a laser-based sensor that simultaneously measures these constituents within the ultra-harsh combustion environment using tunable diode laser absorption spectroscopy (TDLAS). Two-dimensional distribution profiles of these combustion constituents provides a high degree of visibility into the combustion zone and allows real time balancing to improve combustion.

The ZoloBOSS layouts at each plant were customized to adopt the furnace characteristics, burner configuration and access or obstructions at the furnace wall. At AEP Amos 3, the furnace contains 96 DB Riley Model 90 Venturi burners oriented in 12 burner columns at the front and rear walls. There is no overfire air (OFA) on this boiler. As shown in Figure 3, a total of 14 parallel laser paths were located with paths generally above each of the 12 burner columns and additional paths near the side walls. This enables left to right combustion balancing across the 110-foot-wide furnace and permits monitoring of low O2 near the walls that can contribute to water wall corrosion.


The DTE Belle River Unit 2 furnace has 40 Babcock Borsig BBP #6N Venturi burners oriented in two 7-burner rows and one 6-burner row at the front and rear walls. The boiler has 14 overfire air ports. As shown in Figure 4, seven laser paths were installed above each burner column and four orthogonal paths were used to enable two-dimensional combustion balancing. In addition, three single laser paths were located in the superheater pendants to monitor the temperature level and gradient thus providing an early warning of fouling conditions (based on ash fusion temperatures).


Combustion Optimization Tools: Initially, each boiler was manually balanced using the ZoloBOSS real-time temperature, CO and O2 combustion data. The objective was to identify and eliminate localized “pockets” of high/low O2 and CO so that overall excess O2 could be safely lowered without resulting in increased slag. Once manual balancing was completed, combustion optimization software was utilized to further optimize combustion and sustain the combustion improvement over time.

An Emerson SmartProcess combustion optimizer system was delivered to AEP Amos 3. The SmartProcess system balances combustion using the ZoloBOSS data along with advanced analytics and artificial intelligence algorithms. Initially, the SmartProcess system has been limited to maintaining combustion balance across the furnace while operators retained direct control over excess O2 levels. As the plant gains more experience with the optimizer, closed-loop operation will be expanded to include other manipulated variables.

At DTE Belle River 2, an existing NeuCo CombustionOpt system was integrated to maintain combustion balance and further improve efficiency and reduce emissions through optimization of fuel and air mixing. The CombustionOpt system produces real-time closed-loop combustion optimization using a combination of neural network and model predictive control technologies. Control of fuel and air biases is achieved using data from the ZoloBOSS combustion monitor and other plant data.

Plant Training and Project Management: Black and Veatch (B&V), Zolo Technologies and the plant champions at each plant provided efficiency training to plant personnel to increase awareness of CO2 reduction and boiler efficiencies. Additional training encompassed operation and maintenance of the combustion monitor with a live balancing exercise using the ZoloBOSS tomography. Operation and support of the optimization software was also included in the plant training program.

Performance Results at AEP Amos 3

Performance tests at AEP Amos 3 consisted of a baseline test followed by a post-manual balancing test. During both tests, unit stabilization was established during the first hour followed by four hours run at valves-wide-open conditions typical for the Amos 3 normal operation. During the four-hour test periods, fuel and ash samples were acquired for evaluation. For the baseline condition, real-time temperature and concentration measurements of O2, H2O and CO across the furnace left-to-right walls are shown in Figure 5. The unit burns bituminous coal and typically operates at excess air levels of 21% to avoid slagging conditions. The baseline combustion profile shows low O2 along the outside walls of the boiler, a condition typical of wall-fired boilers. Constituent concentration is represented by color shown on the images with blue being 2% and red being 9% for the case of O2.

Manual combustion balancing performed by the plant and Black and Veatch engineers focused primarily on the redistribution of excess air from the middle burner shrouds to the left and right sides of the boiler and reducing the total air flow. Also, manual tuning included the distribution of more fuel to the lower burners. These manual balancing steps are shown schematically in Figure 6.

Air redistribution was accomplished by closing center shrouds to force more air to the sides as this furnace has a common windbox and side shrouds on were already at maximum open. The manual combustion balancing also minimize O2 depletion zones and temperature hot spots in the furnace effectively decreasing the propensity for boiler slagging.

Increasing fuel delivery to the lower burners was achieved by increasing their pulverizer feed rates as a means to stage combustion as this boiler does not have overfire air. This provides more residence time for combustion before reaching the superheat and reheat pendants, and for maintaining the flue gas temperature below the ash fusion temperature to minimize the likelihood of slagging.

Once combustion was effectively balanced, the AEP plant champion, operations staff, and corporate combustion expert determined that the overall O2 set point could be safely reduced from 3.58% to 3.08% corresponding to an excess air reduction from about 21% to 17%. At this point, the post-manual balancing test was conducted and the profiles for temperature and concentration of O2, H2O and CO are also shown in Figure 5.

Manual balancing and excess air reduction at Amos 3 provided the plant with savings in fuel and power through improved boiler efficiency and reduced CO2/MWh. For this demonstration project, CO2/MWh was the key metric as it underscores societal interest in reducing carbon emissions. Further, the CO2/MWh metric serves to provide a straightforward quantification of efficiency taking into account only the tons of CO2 out the stack and the net power produced by the unit. Measurement of CO2 out the stack was determined using a CEMS ultrasonic flow meter and an NDIR CO2 analyzer.

Using the CO2 CEMS method, the baseline performance testing vs. post-manual balancing testing showed a 1.80% decrease in CO2/MWh which represents 184,000 tons/year in CO2 savings. Also, a heat rate improvement of 181 Btu/kWh and a fuel savings of 73,000 tons/year were determined.

Two alternative methods for determining efficiencies were evaluated: Input/Output (IO) and Heat Loss (HL) methods both confirmed significant improvements in heat rate. Lowering overall excess O2 results in a direct reduction in dry gas losses (since less energy is required to heat up the extra air and the reduced volume of flue gas passing over the same heat exchange surface results in improved heat transfer) which improves the net plant heat rate. There is also a reduction in fan loads (ID, FD and Primary Air) since less air is being moved. In this instance, the combination of the dry gas loss improvements and lower auxiliary power requirement are significant. Overall results for unit improvements at Amos 3 are shown in Table 1.

Performance Results at DTE Belle River 2

Belle River 2 burns 100 percent Powder River Basin coal and experiences slag formation in the re-heater and super-heater pendants due to combustion imbalances. This slagging also caused frequent outages. As a result, the boiler was normally operated at higher excess O2 settings to prevent slag build-up even though this lowered the furnace exit gas temperatures, increased the NOx levels and adversely impacted the heat rate. A similar test protocol was used at Belle River 2 to establish baseline combustion profiles using the ZoloBOSS at normal operating conditions as shown in Figure 7. The ZoloBOSS measurements revealed a local pocket of high O2 slightly left-of-center.

Following baseline performance testing, plant operators, Zolo Technologies and B&V engineers used the ZoloBOSS system to balance combustion profiles using the real-time ZoloBOSS data. Based on the ZoloBOSS measurements, O2 distribution was adjusted to remove areas of high O2 and improve the temperature and CO distribution profiles as shown in Figure 7. Once combustion was properly balanced, total air flow was reduced to improve heat rate and reduce auxiliary power. Shroud positions and overfire air dampers were varied based on temperature and CO uniformity as shown in Figure 8.

Following manual balancing, the CombustionOpt optimizer was initialized to maintain balance combustion and further optimize combustion. At Belle River 2, ZoloBOSS data was used to adjust shroud and fuel bias changes to maintain combustion imbalance. With the ZoloBOSS data, the optimizer could successfully operate in a much smaller search space and therefore produce better results. For example, if a boiler imbalance was observed in a particular location, the optimization search space was constrained allowing rapid balancing by adjusting the burner column directly underneath the imbalance. Reducing the optimizer search space from all forty burners and fourteen OFA ports to three burners and two OFA ports significantly reduced the time required for balancing.

The specific steps used to optimize combustion using the CombustionOpt system involved combustion balancing (CO, temperature, and O2) by minimizing deviations in the ZoloBOSS measurements for side to side and front-to-rear. In addition, optimization included adjustment in total air flow by bias of excess air based on the average ZoloBOSS CO and O2, stack CO, and the DCS excess O2 set point. Following optimizer tuning, temperature and CO distribution was further improved and O2 uniformity was maintained as shown in Figure 9.

Boiler efficiency was improved with savings in both fuel and auxiliary power as a result of manual and optimizer tuning. Reductions of CO2/MWh of 2.43 percent and 2.34 percent were calculated following manual and optimizer tuning, respectively, using the CO2 CEMS method of determining CO2 reduction. These improvements correspond to an annual CO2 savings of 127,000 tons following manual tuning and 122,000 tons after optimizer tuning. These results clearly indicate that plant training, an in-furnace combustion monitor, and a software optimizer can provide a 2% reduction in CO2/MWh.

Controllable losses and heat loss methods for determining efficiencies provide more pronounced improvements when manual tuning was followed by optimizer tuning. Overall results for unit improvements at Belle River 2 are shown in Table 2.


The project documents how an in-furnace combustion monitor, a combustion optimizer, and plant training can effectively reduce CO2/MWh density and improve efficiencies. Viewing real-time temperatures and combustion constituent profiles in the ultra-harsh combustion zone provides a better understanding of the combustion process versus standard methods of analysis. Manual balancing and combustion optimization software serve to further enhance this information and provide tools needed to capture savings in reduced emissions, fuel savings and operating costs.

It is estimated that implementation of manual and optimizer tuning technologies to reduce CO2 emissions across the fleet of U.S. coal-fired power plants would reduce CO2 by 37,000,000 tons per year. This level of CO2 reduction represents more CO2-free energy generated than that from the current level of installed solar and wind power in the U.S. These savings in CO2 would be realized at a cost of 1/100th of that associated with large scale solar PV installations and 1/25th of that associated with large scale wind farms.


1. Huelson, E., Logan, N., Sappey, A., Tanck, G., Steiger, C., Jakinovich, N., Scott, J., Alleshouse, T., Spinney, P., Grott, J., Winn, H., (2010) “Carbon Management for Existing Power Plants via Measurement and Control Optimization”, DOE/NETL-2010/RES1000046, Jan. 28, 2011.

2. DiPietro, P., Katrina, K., (2010) “Improving Efficiency of Coal-Fired Power Plants for Near Term Greenhouse Gas Emissions Reductions”, DOE/NETL-2010/1411, April 16, 2010.
Power Engineerng Issue Archives

View Power Generation Articles on