Historically, air preheater repair decisions were based on the boiler approaching operational limitations and inspection results. When administratively imposed performance tests were budgeted, these tests were based on industry practices recommended by the American Society of Mechanical Engineers (ASME). The financial environment for funding these tests has changed considerably in the past 10 to 15 years.

This article addresses some of the drawbacks of applying conventional testing as well as a common-sense approach that can be applied. The alternative air preheater evaluation results can be compared to the ASME testing for both rotary and tubular air preheaters. The primary focus of the article is on tubular-type air preheaters because they are more common, except for larger coal-fired units.

The ability to interpret the air preheater performance and analyze the results is greatly enhanced by understanding the process variability within both the air and gas systems. Examples of the complexities in evaluating boiler and air preheater performance are provided in the following sections. These generic problems can be any combination of air leakage, air bypass, flow limitations and heat transfer. The key to the success of implementing the alternative method is having sufficient time-based operational process data to perform boiler and air preheater performance calculations.

FIGURE 1. Values measured by a four-sensor system on a 200 MW boiler are shown. Sensors B and C are indicating approximately the same average, but the two are not synchronized. Sensor C is indicating a higher value than all the others. This may be an indication of air in-leakage or an imbalance in firing.

The Analysis and Reporting Process

A performance analysis often can be accomplished using customer-provided process data directly from the heater’s human machine interface (HMI) or distributed control system (DCS) historian. The process measurements I/O list is used to select specific unit process variables for performing calculations similar to ASME Power Test Code procedures. Boiler performance with air heater thermal efficiency and leakage are calculated from the process data.

In addition, the emissions reductions may be estimated using basic stoichiometry as part of the process. Gross and net power (or process steam values), representative fuel analysis data and measured combustion gas characteristics are applied with other process measurements. Supporting technical information — heat balance diagrams and cycle flow for air, gas and water, if available — is utilized along with other operating data to establish the model of the unit configuration. This process relies on the ability to properly evaluate the air preheater leakage. In order to accomplish this, a good understanding is required of the O_² readings around the air preheater.

Boiler O2 Measurements

One would expect the combustion gas leaving a boiler to be well mixed after passing from the combustion zone and through a maze of vertical and horizontal tube-bundles. The gases change flow-path directions several times just to enter the air preheater. The combustion gas includes moisture from the fuel and excess combustion air (O₂). From the combustion chemistry, the gas composition is often highly stratified, with zones of variations in concentrations and properties. Boiler sensor probes vary greatly based on the boiler size and operation. While these sensors may be located vertically or horizontally, the number of sensors varies. In addition, interpretation of the data in the control system is not consistent from one boiler to the next.

The point of this discussion is that boiler O₂ measurements do not represent an isokinetic sample arrangement. The result is that boiler control elements are not capable of providing the correct fuel-air ratio for all zones. Essentially, a few square inches sampled represent thousands of square feet of area in the gas-flow path. Figure 1 is typical and indicates values measured by a four-sensor system on a 200 MW boiler. Sensors B and C are indicating approximately the same average, but the two are not synchronized. Sensor C is indicating a higher value than all the others. This may be an indication of air in-leakage or an imbalance in firing. These are dynamic systems providing constant cycling of computation product concentrations. The ability to determine the absolute excess air entering the air heater is challenging.

FIGURE 2. The composed O₂ value at the economizer outlet is shown. It is one of the control drivers for the unit. The test period near the center of the chart (points 151 to 231) identifies a test period where the unit was set to base load for testing purposes.

In other cases, data may be represented as a single trace like the chart in figure 2. That data was collected at the economizer from a medium select O₂ measurement arrangement over a period of just a few hours of a typical unit operating under control near base load. This is a single point driving multiple processes that are not in tune.

Figure 2 represents the composed O₂ value at the economizer outlet and is one of the control drivers for the unit. The test period near the center of the chart identifies a test period where the unit was set to base load for testing purposes. The variability within this stable period is presented as approximately 1 percent; however, it is composed from several points of different values and cycling rates that tend to minimize the variability of any single point.

Based on the two examples shown in the figures, it is easy to see that there is a potential for confusion when analyzing plant data. In cases such as in figure 1, the variation between sensors may be caused by variations in the combustion process, or there may be a leak near the sensor. In those cases, it may be necessary to perform some type of evaluation with temporary instrumentation. Based on what is seen in figure 2, coordinating either a traverse of the duct or installation of a temporary sensor array (as required for ASME PTC testing) is critical to understanding the situation.

Although the unit processes are continually cycling, there is value added from progressing through the process of the slow method traverse testing. Sometimes, anomalies can be detected that may not be indicated anywhere else in the process data. Zones of high CO or abnormal O₂ concentrations sometimes can be identified and evaluated from the traverse data.

The combustion gas stream can be imagined as a river with fruit trees lining both banks. Flowing within the river are varieties of fruit that have fallen out of trees. Imagine the the sampler is a hungry traveler wading across the river. He catches a few pieces as they pass him and use those he is able catch to determine which tree they came from and which fruit is of high quality.

Some plant operators attempt to utilize traverse data to locate areas to target for inspections or repairs. When doing so, they may make the mistake of assuming that the leaking area must be located near the high O₂ indications. Many inspections and repairs have proven that an air preheater tube can develop a leak and then plug because the single tube was cooled below the dewpoint downstream of the leak. When that occurs, high pressure air sometimes flows back toward the boiler and follows the path of least resistance. This creates a high O₂ concentration at the end of that path not related to the location of the leak.

In this case shown in figure 3, the air preheater had dewpoint corrosion near the hot gas inlet. To add to the difficulty of the analysis, in this design, the cold air is admitted to the air preheater on the hot gas side. There is a possibility of high pressure backflow artificially raising the inlet O2 level, thus reducing the calculated air preheater leakage. Once again, this is a situation where understanding the operation of the unit and air preheater, having access to inspection reports and having multiple O₂ measurement locations helps to better understand and define the performance.

The two charts in figure 4 indicate the real drawbacks of attempting to utilize traverse test data alone. By now, it should be clear that the O₂ concentrations are constantly cycling. By the time the equipment is moved from one point to another and from one sample port to another, the indicated differences make it impractical to accurately identify localized defects.

Operational Analysis

A multiyear dataset often is reviewed and analyzed in order to determine the operation and any performance degradation of the air preheater and other downstream equipment. If there are measurements at the stack, it is possible that leakage may be occurring in multiple locations such as ash collection devices, ductwork and inner-diameter fan seals.

This additional leakage often is defined by evaluating key parameters such as differential pressures and temperatures across the air preheater and all downstream equipment. Simple O₂ concentration levels of all the air preheater downstream equipment may locate other sources of air in-leakage. While this method shows an increase in total system leakage, often spot checks of the O₂ levels are required to separate the effects, and this gets back into the O₂ measurement issues already discussed.

Other parameters such as the air preheater exit gas temperature and power output are used to track overall performance. Figure 5 shows a typical plot of increasing O₂ difference between the inlet of the air preheater and the stack.

FIGURE 3. An air preheater tube can develop a leak and then plug because the single tube was cooled below the dewpoint downstream of the leak.

A Performance Analysis Method

Once the operational analysis is complete and a reasonable understanding of the plant operation is gained from reviewing the data with plant operators, the determination of the boiler and air preheater performance can begin. To accomplish this, an Excel spreadsheet has been developed that utilizes a suite of preconfigured thermal properties. The calculations are based on the ASME PTC 4.1 and PTC 4.3 methodologies. This calculator is configured to identify and isolate the impact related exclusively to air heater leakage and to correlate the direct impact to specific equipment. For example, the volume of additional fuel consumed is calculated, and specific analysis is provided on upstream and downstream equipment as well as processes. This includes the volume of additional pollution-control additives injected to compensate for the deterioration of the air heater. If sufficient process measurements are not available within the historian, site measurements will be required. In most cases, the missing data will be related to air heater outlet O₂ concentrations.

FIGURE 4. The two charts indicate the drawbacks of attempting to utilize traverse test data alone. The O2 concentrations are constantly cycling. By the time the equipment is moved from one point to another and from one sample port to another, the indicated differences make it impractical to accurately identify localized defects.

FIGURE 5. This typical plot shows increasing O₂ difference between the inlet of the air preheater and the stack.

The information for this type of analysis often comes from the data collected for the operational analysis. Ideally, data from when the air preheater was not leaking is available and can be used as a reference case. If not, estimations of the performance gains are made based on operational experience and documented assumptions. Using the difference between the operational case and the reference case, a determination of the change in operational costs then can be calculated. These costs are amortized over a set time period — say, five, seven or 10 years for the payback. Ultimately, the level of savings must be greater than the repair cost to entice management to authorize the air preheater repairs.

FIGURE 6. This figure examines the information and shows an amortized costs for repairs.

At best, these measurements and methods may not provide the most precise tests for the air heater; however, they are extremely valuable and can be used to assess additional leakage points between the boiler outlet, including the air heater, to other measurement locations. In order to understand the actual process, it is desirable to accomplish the near impossible with the current environment, which is to be able to view all of these variables in a single snapshot quickly. While not optimum, there is substantial value in collecting the historical data from the plant historian and continuous emissions monitoring systems (CEMS) data to support the performance evaluation and financial impact. These alternative methods are economical and powerful enough to be used for pre-/post-repair studies and performance monitoring to determine the optimum time to initiate a major repair. With this information and discussions with plant operators, experts can assist processors by understanding the unit constraints and applying these to the financial impact determination. With this information as well as physical inspection of the air preheater, and expert can propose a viable solution.