ABSTRACT

Gas analysis close to the boiler exit often gives the most valuable process control information but poses serious problems for the measuring instrumentation. The gas typically will be at high temperature, will be strongly stratified and may contain high burdens of particulate material.

The normal choice to be made is between extractive and in situ gas monitors but both have disadvantages in this type of application. Extractive systems Use conventional gas analyzers in conjunction with sophisticated sampling systems which clean and dry the gas. These can lead to unacceptably long response times and unpredictable changes in sample composition. More importantly, sampling systems comprising filters, pumps and bubblers require frequent and expensive maintenance. In situ analyzers, although more robust, do not match up to conventional analyzers in terms of accuracy and reliability when operating in such harsh conditions.

This paper describes a hybrid of the two types of systems utilized to measure the concentration of carbon monoxide (CO) at the economizer outlet of a gas-fired utility boiler. Samples are extracted from the gas duct using multiple probes which are strategically located to ensure a sample composition which is representative of the bulk stream. This sample is cooled, passed through a simple water ex-tractor and analyzed by an in situ gas monitor without further treatment. Instead of an inductor or vacuum pump, the sample is extracted by using the negative pressure of the forced-draft fan.

Results are presented from a trial application on a Louisiana power station and their use and interpretation are discussed.

INTRODUCTION

On-line measurement of the constituents of flue gas has for a long time been troublesome for the utility industry. This paper addresses some of those problems and a new solution for measuring carbon monoxide (CO). We have combined the best features from both the ex-traction type and the in situ type for this solution. The goal is to have a measurement which can be used for closed loop control and be acceptable for environmental reporting. Because of the slow speed of response, stack mounted methods are not addressed.

EXTRACTION TYPE

The extraction method removes a sample from a vessel and delivers it to a measurement instrument - usually in a different location. This method has served the industry for a number of years even though there was considerable maintenance expense on the system. A variety of measurement -methods were used including chemical cells, thermal conductivity, thermo-magnetic and infrared. Most of these methods worked well when the sample was delivered clean and at a dew point lower than the operating temperature of the measuring device.

The sample conditioning is the main problem with extraction systems. To clean up the sample, filters are used on the sample probes and throughout the sample system. Filters add time delays to the sample stream. Other sample conditioning methods are necessary, such as liquid bubbler chambers, flow and pressure regulators, all of which add volume to the sample system. Added volume adds delay to the sample stream. Delays of 30 to 120 seconds are common. Time constants of this magnitude added to the inherent delays of the flue gas path, make closed loop control very difficult and of questionable value.

Moving the sample point as near as possible to the furnace exit will reduce the overall delay but it creates a problem with gas sample temperature. The higher sample temperature requires greater cooling capacity and larger volume bubblers for sample conditioning. It is essential to maintain the sample dew point below the analyzer operating temperature. Allowing condensate to form in the analyzer will quickly damage the analyzer cells. These systems are notorious for their high maintenance cost.

The best feature of extraction systems is the flexibility of averaging multiple sample points for a truly representative sample.

IN SITU TYPE

In situ type analyzers measure the flue gas constituent in place. For CO measurement some type of infrared analyzer is generally used. In situ analyzers have good response time and work well in closed loop control of CO. The most widely used location of infrared in situ CO analyzers is across a duct carrying flue gas. For control, the preferred location is at the economizer outlet provided the gas temperature will be below 700'F. Above 700'F, the accuracy of these instruments falls off rapidly.

There is another problem with across the duct in situ measurement due to horizontal and vertical stratification. Stratification in a duct has many widely varying patterns. To obtain a representative sample would require several sighting paths each requiring a complete infrared analyzer system. Obviously this would be very costly.

The sensitivity and accuracy of the infrared in situ analyzer has proven to be sufficient for tile required environmental documentation when it can see a representative sample.

EXTRACTION / IN SITU METHOD "HEUSZOMETER"

The Heuszometer consists of a multi-probe sample extraction system, a simplified -moisture removal device and a measurement chamber using an infrared in situ type CO analyzer. The probe assembly is designed to extract a sample from a cross section of the duct at the economizer outlet. Each probe is an assembly of 3 to 5 sample tubes. A combination of probe assemblies and tube lengths is strategically placed to give a good representative sample. Each of the sample tubes in one duct is connected to the moisture removal device. Care is taken to maintain equal flow in each tube by maintaining equal resistance in each flow path. The tubing is run in such a way as to obtain some sample cooling before reaching the moisture removal device. This device serves several functions, including mixing the multiple sample streams and cooling the sample. The internal design consists of cone-shaped discs placed in the path of the incoming samples so that proper mixing of the samples occurs. At the same time, additional sample cooling is accomplished by chilling the dispersion cones using an air operated energy conversion device. This chamber cools the sample below its dew point and the moisture is dropped out of the sample and discharged to a drain.

The composite sample is sent to a measuring chamber. An infrared CO analyzer is mounted on the measuring chamber. The design of the measuring chamber is such that ample sensitivity is obtained to comply with environmental requirements and to provide sufficiently rapid response for use in closed loop control. The exhaust line from the measuring chamber connects to a point on the unit with a lower pressure than the point of the sample. At Gulf States Utilities, we installed the exhaust line to vent into the intake of the forced draft fans. By this method, we have been able to have response times as short as seven seconds and with average times of 10 to 12 seconds. These response times are well within the realm of good closed loop control.

RESULTS OF TRIAL APPLICATION

The Heuszometer combines the best of two worlds in the use of its sample probes, which have been proven in the extraction systems and the in situ measurement which works well when it can see a representative sample. The Heuszometer Mark II system was installed in 1982 at Nelson Station in southwestern Louisiana. The system has worked well since that time and has provided much information for operating the unit as well as for designing changes incorporated into the Mark M. The Mark M version has just been installed on one unit at Sabine Station. Plans are to install this system on five additional units in Texas.

The Heuszometer Mark U provided operator guidance for the set point on 02 automatic control from 1982 through 1991 with a minimum of maintenance. When the analog analyzer began needing maintenance, it was replaced with a newer digital version. During the nine year trial period, operations occasionally questioned the CO reading because the O., reading at the economizer outlet would be higher than the textbook version of CO versus 0.2. The operators became believers only after several verifications of the readings were made by our performance department's test van. High temperature oxygen probes were installed in the furnace which eliminated the penthouse leakage in the 02 measurement and this convinced all doubters that the CO measurement was correct.

Since the utility industry must comply with new environmental standards, a reliable CO measurement will be essential. To reduce NOX, the logical approach is to do those things which have the least cost, like low excess air firing, burners out of service, individual burner flame shaping techniques, and other attempts to change the temperature profile in the furnace. Under several of these conditions, stoichiometric or even sub-stoichiometric may be required. When this is the situation, a fast, reliable CO measurement is required.

The Heuszometer Mark HI was recently tested at Sabine Station. Its accuracy was verified by standard calibration gases as well as in actual operation on the unit at 400 megawatts (MW). The system was operated in parallel with the performance test van and it became a question as to which system was the standard. The Mark III had about a 10-second response time which was 40 seconds faster than the test van.

CONCLUSION

As utility companies tighten up the firing of their- units and get reliable CO measurements, the inadequate wind box distribution and unreliable burner air registers will become apparent and troublesome. One technique which the Heuszometer uses is to take single sample tube readings so that a duct profile of CO is possible. The proper use of this method will be a big assist in adjusting burners.

The Mark D[I Heuszometer is now being built by the analyzer supplier. This new system should fill the need for CO measurement and control on most fossil fuel fired utility boilers. Our testing has been done on natural gas fired boilers. We see little, if any, changes needed for use on oil fired units. A patent has been applied for both the hardware and the process.

Plans are being made under a separate project to do testing on coal fired units for possibly using filters on the probes or increasing the tube size of each probe. The Heuszometer is a proven design for oil and gas. When further testing is done on coal fired units a paper will be submitted on that outcome. In the meantime, the 'Heuszometer' is available for gas or oil fired units.