INTRODUCTION

This paper primarily addresses the location of the measurement of oxygen in flue gas depending on the type of boiler and if it is a balanced draft type or pressure furnace. Boilers which have negative pressures in the back passes of the boiler or with penthouses operating at a positive pressure referenced to furnace pressure will have entrained air in the flue gases exiting the economizer. This entrained or tramp air becomes significant when low excess air operation is done. An approach using high temperature oxygen probes measures the oxygen content in most cases prior to the air entrainment.

OXYGEN MEASUREMENT

Zirconium Oxide based oxygen analyzers are the latest ma or development in measuring oxygen in flue gas. Zirconium oxide is a ceramic that will allow the transfer of oxygen ions across its matrix at temperatures above 650 degrees Centigrade (1200 degrees Fahrenheit). This transfer of ions and the subsequent electrical charge buildup is proportional to an oxygen differential across the Zirconia cell.

There are three types of Zirconia based analyzers:

A) Heated in-situ (used in temperatures of less than 1200 degrees Fahrenheit)

B) Extractive and close coupled extractive

C) In-situ high temperature

Zirconia must be at 1200 degrees Fahrenheit or higher to respond to the presence of oxygen.

There are two popular ways of maintaining the Zirconia cell at 1200 degrees or greater.

1. Use an auxiliary heater in close proximity to the cell along with a control loop to regulate the power flow for the heater. This method assumes a constant temperature and measures the millivolt output of the cell which is inversely proportional to oxygen. This type is used in both in-situ and close coupled extractive types.

2. The high temperature Zirconia probe uses the process as its heat source. In this case both

the temperature and the millivolt output of the Zirconium Oxide cell are measured and applied to the Nernst equation.

E = .0215 x T x Ln ( O2 Reference/02 Furnace )

Where E = Sensor output in millivolts

T = Sensor temperature in degrees Kelvin

02 Reference = Oxygen concentration on inside surface of the sensor ambient air: 20.9 % oxygen

02 Furnace = Oxygen concentration in furnace atmosphere

The heated in-situ and extractive types are typically mounted in the back pass of a boiler. There are many installations using this type of sensor. There are often problems with the extractive types because of sample system maintenance. The heated in-situ designs do not have as many sampling problems but there are maintenance requirements that in many cases become intense. The maintenance requirements will vary with the fuel burned, with gas requiring the least maintenance. Oil and coal provide increasing problems due to heavier particulate and contaminate inclusions. These are typical maintenance problems and they are dealt with as routine. The more serious problem with these analyzers is not maintenance, accuracy of the device or durability, it is the location in which they are used.

The back pass or economizer outlet installations are subject to the influence of air infiltration anywhere between the burner zone and point of measurement. The effect becomes exaggerated if there is a pressurized penthouse. Leakage may become more severe as the unit ages. Installation of high temperature sensors in various locations in the radiant section of the boiler minimizes the effect of this "tramp' air. This provides a far more representative reading of the actual boiler combustion conditions than back pass measurements. The high temperature sensor also requires far less maintenance than the other types because it does not use any heaters or sample systems. The sensors may be mounted through existing doors or ports but if there is not a satisfactory existing location, waterwall tubes may have to be bent during an outage to provide n opening into the upper furnace area.

Boilers which were built prior to the late 70's have burner and windbox arrangements which were designed to have excess air of 10% or greater. Distribution between burners generally is unequal. Proponents of CO control generally show a text book version of the 02 versus CO curve and recommend that the optimum point to operate is in the 02/CO knee. In actual

application, the knee varies considerably. Some of the factors are:

1. Erratic signal of CO measurement of 200 to 400 PPM on a 0-1000 PPM scale.

2. The fuel/air ratio curve on a per burner basis is very different and non-linear with load.

3. Lack of stability or positioning ability of both fuel and air control actuators.

4. Air in-leakage from pressurized penthouses. This leakage varies with (A) size of the leak,

(B) unit load, or (C) differential pressure between the penthouse and the furnace.

5. Horizontal and vertical stratification in the duct where the conventional 02 measurement is made.

6. Most curves given in previous papers showing the knee (crossover point) of 02/CO are based on 02 readings taken at the economizer outlet. These measurements include any entrainment air (tramp air) and, therefore, do not accurately represent the 02 of the combustion gases in the furnace.

CO measurement is normally only Applicable at loads above 70% of full load, therefore, a reliable 02 measurement is essential for set point control at loads below 70%. A high temperature 02 probe in the furnace is a new application of a proven probe in the steel and glass industry. It offers a lot of possibilities and solutions to some of the problems listed above.

At GSU the first installation was two probes in the front wall across from the archway on a supercritical unit. As the unit picked up load above 75 %, the 02 was being controlled at 2.8% at the economizer outlet. The two furnace probes indicated 0% and 1.2%. This generally indicated a difference between furnace 02 and economizer outlet 02 as much as 2 %.

Further tests revealed that the difference in 02 between the two measurement points was directly affected by the penthouse to furnace differential pressure (DP). The DP is normally held between 0.5 inches H20 and 1.0 inch H20. This DP control is frequently inadequate on many boilers, therefore, the error of 02 becomes a function of leakage area and DP.

A later installation with four probes gave a better understanding of the difference of 02 from side to side. It was found that in each case when a large difference between sides was seen, it was a result of a bad operating burner or air register.

Operating with these high temperature probes in place revealed the following interesting bits. of

information:

1. Different burners in service could change the 02 reading between sides as much as 2% 02 at loads where burners could be removed. This proved that the air distribution to the individual burners was inadequate and air register adjustments were required.

2. The probes also have a temperature measurement, and at approximately the same load during the daily load cycle, a temperature difference of 40 degrees Fahrenheit would change sides. This was with all burners in service. After an outage on the unit, during which time air register travel was adjusted and considerable other boiler work was done, the situation cleared up.

3. The plant operations people were seeing things for which they had no explanation and in some cases were happier not knowing.

Of the three B&W gas fired boilers operating with high temperature probes, two are of the El Paso design and both have four probes each, the one Carolina design is a universal pressure unit and has four probes installed.

Installation into the boiler is difficult and must be done properly or breakage will occur. Excluding breakage, the probe life is in excess of one year each. Maintenance consists of assuring that -- small amount clean, dry reference air is flowing to the cell.

We presently use high temperature probes, on automatic control, on two B&W boilers. The 02 control is performing quite well and controls the 02 to the ramped set point from 3 % at low load to 0. 4 % at high load. The band width of 02 on the chart is approximately 0. 15 % 02. the much faster response and the location of the probes makes a much tighter control loop than we have experienced before.

In the beginning, considerable effort was placed into designing a system to facilitate on-line

probe replacement. The design utilized an aspirating air mechanism to contain hot cases required by our pressurized furnaces. Although this system functioned properly, safety hazards became a major concern. Handling a probe from at least 1200'F furnace gas temperature was the major concern. In addition, the installation of a new probe requires careful insertion at a slow rate to prevent breakage. It was decided to abandon the idea since the probes have an 18-24 month life and are routinely replaced during an outage.

CONCLUSION

The high temperature 02 probes have proven a valuable tool as a faster response than the back pass measurement. Measuring 02 in the furnace area allows for burner and air register adjustment which up to now has been a guess at best. In conjunction with a good CO measurement, it is a tool for evaluating burner performance. When the high temperature 02 probes are used in conjunction with economizer outlet measurement of 02, the difference is a measurement of in-leakage or tramp air. There is still more work to be done to obtain the best performance from these probes. With the environmental requirements and the probability of reduced O2 firing, this should become a vital measurement.

ACKNOWLEDGMENT

1. The Nernst equation was taken from the user's manual for oxygen sensors from AACC/Zircoa, Solon Ohio.