I am sorry to say that figures are not able to be shown at this present time, please refer to the magazine MPS March 02.
Achieving low NOx in coal-fired plants will require tighter control of the air and fuel balance at each burner in the system, a trend comparable to the switch from carburettors to fuel injection systems that we have seen in the car industry.
Optimising the balance requires on-line measurement of the coal flow in each conduit. This technology is thus one of the vital steps towards ultra-low-NOx firing.
One approach to on-line coal flow measurement is Electric Charge Transfer technology (Figure 1). This measures the electric charges present in the coal flow and uses the signals to determine the following characteristics of the flow:
• Relative coal distribution between the conduits. In addition, the system can be configured to measure the flow velocity and the absolute flow in each conduit. Coal flow balance is the key measurement for continuous ultra low NOx combustion.
• Unsteady phenomena in coal conduits that can cause problems during plant operation. This is possible because of the high data collection rate of the ECT system. The signals can be used to detect coal conduit layout due to insufficient primary airflow from the mill and coal conduit surging which results in furnace pressure and emissions fluctuations.
• Particle size changes in the coal flow. The antennas and hardware used for this are the same as those for the coal flow distribution application. Thus, the ECT system can be used to monitor mill performance on-line assuring proper coal fineness is maintained to minimise unburned carbon.
The ECT system consists of receiving antennas in each coal conduit that are connected to a signal conditioning unit. The signal conditioning unit is in turn connected to a PC for data processing and analysis.
Proprietary software is used to determine the balance between the conduits of one mill, to display the results to the operator and to feed the data via a network to the plant’s DCS system or continuous combustion optimisation software running on a separate computer.
The antennas are installed through the horizontal or vertical wall of the existing conduit and inserted into the coal stream. Three antennas in each conduit are needed for coal flow balance measurement and six for coal flow and velocity measurement. Their location in the pipe wall is determined so that the effects of coal ropes on the measurement results are minimised. Antennas are made of tungsten carbide to ensure long operating life.
The ECT approach has the following advantages:
• all information is continuous and on-line;
• the ECT measurement is not affected by coal type, moisture, ash content or coal roping;
• the electronics can be located up to 1200 ft (about 400 m) from the conduits and no cabinets are needed on the burner decks;
• the abrasion resistant antennas in the coal conduit are passive and need no power supply; and
• installation is relatively easy, taking a matter of hours.
The ECT system was originally developed by TR-Tech of Finland. Foster-Wheeler is partnering TR-Tech to market the system worldwide and is the global distributor.
The technology has now been installed on ten utility boilers. These include wall and tangential fired furnaces in Europe (four boilers) and in the USA (six boilers).
Six more installations are scheduled to start up during the first half of 2002 in the US.
In addition, Foster Wheeler and TR-Tech participated in seven earlier demonstration projects, several of which were funded by the UK Department of Trade and Industry and EPRI.
The Logan case
The ECT system was installed at PG&E’s 219 MWe Logan station in the spring of 2000. Logan, erected in 1993, has a natural circulation Foster Wheeler boiler with 12 low NOx burners arranged in four rows in the front wall. For additional NOx reduction the unit has front and rear wall OFA ports and SCR. The plant fires a high volatile eastern bituminous coal from Virginia, with an average heating value of 12 700 Btu/lb, ash content 10 per cent and volatile matter content 33 per cent.
The boiler is served by two D-10D double-ended FW ball tube mills (Figure 2). The fuel from each mill end is supplied to a row of three burners via an adjustable three-way distributor mounted at the outlet of the heart-shaped classifiers. The distribution of the fuel to the three pipes is influenced by fingers mounted in side pockets of the distributor vessel.
The Logan boiler was prone to high LOI values and firing system imbalances resulting in increased NOx and CO emissions. The negative effects of the combustion imbalances were compounded by a relatively small furnace height and the need for high overfire air flow to provide a large part of the NOx reduction in the furnace. Also, front-fired furnaces generally suffer from limited mixing of the flue gas in the furnace and subsequently reduced carbon burnout. Unburned carbon in the fly ash ranged originally from 20 to 25 per cent.
Logan conducted an extensive optimisation of the firing system in spring 2000. Some shortcomings of the air distribution system in the windbox were corrected and more overfire air flow was biased to the rear wall to complement the flow of the furnace flue gas.
The ECT system was installed to address the combustion imbalances and as a tool to balance the air and fuel flow to the burners. The Logan ECT system was originally configured only to display the coal flow balance between the twelve burners. An upgrade to include velocity measurement was subsequently performed to gain additional benefits from optimising primary air flow.
Figure 3 shows the relative flow balance of six coal pipes of a mill and the comparison to Rotorprobe sampling. The conduits A1 to A3 are serving a lower burner row from left to right and the remaining conduits an upper burner row. The coal flow distribution varies between 13 and 20 per cent. The ideal value is 16.7 per cent of the total flow. It is apparent from Figure 3 that the boiler has a systematic fuel imbalance favouring the left of the boiler. The coal distribution from this mill is 38, 34 and 28 per cent of the total flow for the left, middle and right burner column, respectively.
The differences between the ECT measurement and the Rotorprobe values are small, but with ECT the information is available at all times. In addition, the coal flow variation in individual pipes due to changes in mill load or other influences is tracked with a high data rate and thus unstable signals can clearly be detected, allowing mill performance problems and poor mill condition to be readily identified.
Some of the benefits from the active balancing of the combustion system are shown in Figure 4. The long-term average of the unburned carbon was reduced from 19.3 to 13.6 per cent. This significant improvement results in a very short payback time for the investment in ECT. An additional benefit is the improvement of the fly ash quality. Although Logan does not sell the ash, ash from many power plants has to be below a maximum unburned carbon value of 4-6 per cent. As shown in Figure 4, balancing also results in reduced standard deviation of the carbon in ash and therefore helps to achieve the sales goal.
The improvement in furnace NOx emissions resulted in a reduction of the SCR ammonia spray rates of about 10 per cent. The better combustion balance achieves not only reduced NOx emissions but also a more even emission and oxygen profile at the inlet of the SCR. Thus, ammonia slip and ammonia bisulphate deposition in the air heater are reduced. At Logan, the reduced deposits in the air heater allowed a doubling of the intervals between air heater washings, significantly increasing plant availability.
In addition to the long-term benefits from the reduction of unburned carbon and ammonia spray rates, the ECT system provided valuable data for assessing firing system operation. It was observed that the DC raw signals trended with coal quality. The voltages measured by the system decreased when a coal from another mine source was fired. This was not initially anticipated and required changes to the ECT calculation that significantly improved the robustness of the ECT algorithms.
The changes in DC signal levels can be used to detect coal quality changes on-line. This is helpful when high moisture coals result in CO increase and higher NOx and generally more difficult boiler operation.
On one occasion, a hot burner alarm was tracked with the help of the ECT system to a problem at one of the adjustable distributors. Due to a loose set-screw the adjustment had reduced the air flow in the pipe, which resulted in increased burner temperatures. The ECT signal clearly showed the changes in the distribution and the problem was resolved on-line without taking the mill out of service and reducing load.
Velocity measurement
As is being done at Logan, the ECT system can also be used to measure the velocity of the particles in the coal conduits.
This function is important to determine if a conduit is in danger of coal layout and subsequent damage from conduit fires. In addition, an optimised primary air flow balance is needed for improved low NOx burner performance since burner exit velocities should be close to the design velocities. The velocity information can also be used for calculating mass flow in the individual coal conduits.
To determine the velocity in the conduit the time of flight of the particles between two sets of antennas mounted at a defined distance apart is measured and used to calculate the velocity (Figure 5). No calibration is needed for the velocity measurement.
Figure 6 (which is not Logan) shows the velocities in eight pipes of a mill versus time. The conduit velocity averages vary between 75 and 92 ft/s and show significant fluctuations. The velocities in pipe 2 are quite low when coal is transported. However, as the mill is taken out of service, all velocities are similar, which shows that the primary air balance without coal is quite good.
This example and the one shown in Figure 3 show how significant the differences in coal distribution and transport velocity in existing pulverised fuel systems can be. It is also interesting to note that the fuel flow imbalance is almost never coincident with the primary airflow imbalance in a pulverised fuel system.
Particle size monitoring
Particle size detection is another application for the ECT system. Changes in fuel fineness have an impact on the unburned carbon in the fly ash and can offset improvements that careful burner air and fuel balancing might yield. The pulveriser fineness performance is impacted by many variables such as hardgrove index, moisture, mill wear, grinding force or classifier setting. Almost all of these constantly change, which makes it difficult to maintain constant fuel fineness and to monitor the mill maintenance condition.
The ECT system can be configured to measure the fuel fineness on-line. To obtain the relevant signals two sensors are mounted in a coal pipe right after a pipe bend (Figure 7). The natural particle segregation in the pipe bend is used to measure the particle fineness. An example of this measurement is shown in Figure 8 as fineness in per cent passing Mesh 100 (149µm) versus time.
To show the influence of the rotating classifier speed on fineness, several different speeds were tested during the day.
It can be clearly seen that the ECT fineness data changes with the mill classifier settings. The two Rotorprobe samples compare very well to the on-line data. The trend shows not only the improvement of fineness with increased classifier speed, but also that the fluctuation in fineness changes. The output is much more consistent at high classifier speeds. A decrease of rotating classifier speed shows the expected increase of the coarse particles.
The ECT particle size monitoring allows operators to evaluate pulveriser performance on line. The signals arising from changes in product fineness can be used for classifier control and preventive maintenance scheduling. Any short-term changes are signs of pulveriser system upsets and allow early troubleshooting, before the unburned carbon levels in the fly ash increase.
Expanding applications
As these examples show, the application of ECT technology is growing beyond the original focus of pulverised coal flow measurement. For example, the velocity and coal particle fineness measurement expands the use to a mill diagnostic tool, allowing on-line mill performance assessment. Benefits from these measurements include mill primary air flow and unburned carbon reduction.
Coal flow measurement technology is also advancing quickly towards systems that allow an active control of air and fuel for combustion balancing. A development step in this trend is the automatic adjustment of secondary air flow to the measured coal flow.
This milestone was recently demonstrated with a Foster Wheeler TLN Low NOx tangential combustion system using a Foster Wheeler/TR-Tech control strategy coupled with Foster Wheeler’s CADM air flow measurement and ECT coal flow measurement. The result is reduced CO and opacity emissions for ultra-low -NOx firing systems.