Evaluating the Potential for HRSG Cold-End Deposition
Cold-end deposits in the HRSG are a major problem at many plants. They typically comprise corrosion products from the reaction of sulphuric acid with iron on boiler surfaces. The sulphuric acid is created by the reaction of sulphur trioxide (SO3), which is created by burning fuels with traces of sulphur, with moisture in the exhaust gas.
The acid will react to corrode the iron in HRSG components if it condenses on their surfaces and forms deposits of iron oxides and iron-sulfur compounds. Therefore, the objective is to keep the metal temperature of the heat transfer surfaces above the acid dew point. Tubes containing water and/or steam at saturation will have an external metal temperature that will be very close to the inside fluid temperature.
Any deposits will accumulate and clog the finned HRSG heat exchanger tubes. This occurs at the back or "cold" end of the unit, toward the exhaust stack, because that is where the metal temperatures can be low enough to drop below the sulphuric acid dew point. If the plant has a selective catalytic reduction system for NOx control, then deposits of ammonium bisulphate may also be present at the cold end, worsening the fouling problem.
Predicting the absolute rate of deposition is difficult because of numerous uncertainties. These include the amount of sulfur present in the exhaust gas and, more importantly, how much of the sulfur is converted to SO3. Nevertheless, by accurately calculating the metal temperatures on heat exchanger surfaces throughout the HRSG, a relative risk of deposition can be identified and the impact of potential solutions investigated.
A software model of the HRSG was developed and used to predict the cold end tube wall and fin temperatures and investigate the effect on plant performance of raising the cold-end temperature by recirculating inlet feedwater.
Figure 3 shows the cold-end portion of the heat-energy-temperature (QT) diagram from the simulation run with the GT outlet (HRSG inlet) conditions corresponding to the OEM full-load base design case. The green dotted lines show the probable range (160F to 190F) for the sulphuric acid dew point, which is calculated using typical values for fuel gas sulfur and exhaust moisture content.
3. Chilly forecast. The boiler simulation run predicts the cold-end temperature profile. The calculated most-probable range of acid dew point is shown by the green dotted lines. Courtesy: Tetra Engineering Europe
Clearly, the last sections of the high-pressure economizer 1 (HPECO1) and the split intermediate-pressure/low-pressure (IP/LP) economizers are below the temperatures where acid may potentially condense on the tubes. These modules, each comprising seven rows, are located in parallel in the gas path and therefore have the same temperature profiles.
The fluid temperature predicted by the model at the entrance to each row of IP/LP economizer tubes is shown in Table 1. From this data it appears that only row 1 and possibly part of row 2 are at risk for deposition.
Table 1. The acid test. Simulations are performed to determine if the fluid temperatures are such are that acid may potentially condense on the tubes. Inlet water temperatures to rows 1 through 7 of IP/LP economizer are shown. The first value is for row 1 nearest the HRSG exhaust stack. Courtesy: Tetra Engineering
Europe
The temperature profile in a tube row can be obtained from the boiler simulation. It shows that only the lower half of the tube is susceptible to cold-end deposition.
As mentioned above, the metal temperatures in the tube are assumed to be identical to fluid temperatures. This can be verified and the zones that are at risk for deposition further refined. Table 2 shows the tube average temperature profile from the inside wall (which is essentially at the fluid temperature) to the fin tip. The temperature difference is just under 25F. A similar difference is found in the other tube rows. It has been shown that acid condensation in environments with few particulates (such as in HRSGs fired with natural gas) will be small at temperatures that are just within a few degrees of the dew point. This observation would imply that much of the surface area of row 2 would not be at risk of deposition.
Table 2. Looking for trouble. This is the averaged tube temperature profile for row 1 of the IP/LP economizer. Comparing the metal temperatures with the fluid temperatures can reveal potential risk for deposition. Courtesy: Tetra Engineering Europe
A recent visual inspection confirmed that only the last row, and in particular the lower sections of the last row, had significant deposits.
The next step was to investigate the effect of adding a recirculation loop to the feedwater inlet. This would allow the raising of the inlet water temperature to above the sulphuric acid dew point, thereby potentially reducing the amount of cold-end deposition.
The HRSG model was modified to add separate recirculation loops for the HP economizer and IP/LP economizer. In the actual plant this would represent a significant capital cost, requiring the installation of two pumps, two 3-way valves, connecting piping, and additional instrumentation and control.
The control loop was set to raise the feedwater temperature to 194F (90C). This could be considered the minimum that would raise tube metal temperatures above the estimated maximum dew point over all of the heat exchanger surfaces.
The result was a reduction of the HRSG input-to-output efficiency from 87.19% to 86.12%. This translates into roughly a 1% loss in steam turbine output. Changing the recirculation temperature setting to 176F (80C) reduced the efficiency drop to 0.67%, although this is still significant.
Fouling from deposits will also lead to increased backpressure on the GT and a loss of heat transfer on the tubes. The analyzed HRSG has economizer tubes with a relatively high fin pitch (286 fins/m or 7.26 fins/in.). This renders the unit more susceptible to backpressure problems if any deposits form. Backpressure changes can be tracked and trended by logging the measured gas pressure drop across the HRSG with operating conditions normalized to equivalent mass flows.
The heat absorbed in the last few economizer rows is small, and deposition in these areas has little effect on performance. This hypothesis was tested with the boiler model by simulating the fouling of all seven rows of the IP/LP and HP economizers. A fouling factor of 5% in heat transfer efficiency was first applied, implying very heavy fouling. The result was a 0.15% reduction in efficiency. A more realistic fouling factor, again applied over all seven economizer tube rows, caused a 0.05% drop in efficiency. As fouling currently is occurring only across the last one or two rows, these modeling results imply that the actual efficiency losses from reduced heat transfer are negligible.
Email
Comments
Print
Reuse
