HRSG problem-solving 101: Don’t jump to conclusions

Sometimes you have to dig deeper into the detective work to identify not-so-obvious issues with your heat-recovery steam generators, HRST Inc’s Sam Shaw told the editors. Occasionally, captured DCS data are limited and more information is required to properly diagnose a nagging reliability issue. It’s easy to blame failures on an obvious root cause, but with additional data review, issues initially unsuspected sometimes reveal themselves.

Examples of additional data collection techniques may include the installation of temporary thermocouples (T/Cs) in strategic locations to identify any thermal abnormalities.

Recently, Shaw and his HRST colleagues investigated the root cause of several tube failures on a double-wide HRSG installed in 2002. Failures were occurring at the lower tube-to-header connections of the reheater, comprised of two-row, two-pass panels.

Early in the investigation, the engineers suspected the observed tube deflection could be a result of the temperature differential between the down-pass (cooler) and up-pass (hotter) tubes. However, it was later verified by inspection that the reheater had spring support hangers (not shown on drawings) to accommodate differential expansion. Additionally, calculations revealed little difference in the temperature of adjacent tube rows. Thus the initial thought of large pass-to-pass temperature differences proved an unlikely source of stress.

Temperature variations across the width of the individual harps then were suspected by the HRST engineers as the root cause of the failures. The non-uniformity of the observed stresses further supported this theory. The tubes near the center of the harp (in line with the inlet) showed bulging and scale exfoliation. Clearly, data beyond those available from the DCS were necessary.


The required information would come from temporary thermocouples installed by HRST on the bare sections of tubes in the lower crawl spaces (Figs 1-3). Of primary concern were the temperature differences across the width of an individual harp. Keep in mind that the average temperature of a tube along its length can be calculated knowing the temperature at the bottom of the tube.

The strategy for positioning the T/Cs was as follows:

      • Focus on one harp and instrument it thoroughly. Look for differences occurring across the width of the panel.

      • Install T/Cs directly under the inlet because those tubes revealed they were stressed.

      • Install T/Cs at the ends of the harp to identify any issues associated with side-wall or center bypass of turbine exhaust gas.

      • Place T/Cs on the drains to spot water that might be flowing in the wrong direction.

      • Concentrate instrumentation on down-flow tubes; instrument a small number of up-flow tubes for reference.

      • Record temperatures for a cold start, hot start, and steady-state conditions.

Feedback. Under steady-state and warm-start conditions, the tube-to-tube temperature differences were minimal and unlikely to cause failures. However, under cold-start conditions, temperature differences in excess of 500 deg F were measured between tubes under the inlet and at the side wall. This delta T is more than sufficient to cause stress-induced failures at the tube-to-header connections.

HRST D5 4Fig 4 shows the tube-metal temperature measured by installed T/Cs at a single moment during a cold start. Reheater pressure at the time of this event was approximately zero. The measured temperature of the tubes below the inlet drops to just above 200F—close to the saturation temperature at atmospheric pressure. This proves water was entering the tubes and cooling them.

A rookie investigator’s first instinct, the experienced Shaw said, might be to blame the water infiltration on a leaking or improperly operating desuperheater spray system. But further review of the data suggested otherwise. Fig 5 presents DCS data for the cold start. The information is not particularly troubling by itself. However, when compared to Fig 6, which shows the measured tube-metal temperature of the same period of time, problems become apparent.

Fig 6 reveals a significant disturbance corresponding to first indicated steam flow. The tubes in line with the inlet are cooled suddenly, while those at the ends of the harp are largely unaffected. Refer back to Fig 4 for a snapshot of the temperature distribution during the disturbance.

Engineers believed that improper desuperheater operation likely was not the source of water infiltration. Rather, water was being carried into the reheater with the initial steam flow during a cold start. Oftentimes the source of such water can be undrained condensate which builds up in steam lines while the unit is offline. Further review of the system revealed there were no drains in the cold reheat piping. Thus any condensate collected in that piping must pass through reheater harps before reaching a drain.

HRST D5 5,6

By further investigating the issue, asking more questions, and gathering additional data, engineers were able to identify the root cause of the failures. Lesson relearned: Sometimes the first suspect may not be the one you are looking for. It may take more digging and verification to get the information you want.

Posted in D5-D5A Users |

Comments are closed.