Case Study: Relief Valve Replacement Analysis

Problem: A plant was replacing two pressure relief valves at their facility with larger capacity units. The plant needed an analysis to determine if the existing piping could tolerate the larger reaction forces associated with the high capacity valves, or if not, what modifications would be required.

Solution: KHE developed numerical pipe stress models of the existing piping systems, associated with the two relief valves targeted for replacement. Due to a lack of available documentation on the existing piping design, KHE personnel measured the existing piping and supports to enable development of the models. Pipe stress analysis was performed in accordance with API B31.3 for various load cases, including cases that involved relief valve reaction forces. Sensitivity studies were performed to ensure the boundaries of the model were sufficiently extended to not impact the results of the analysis.

Result: The pipe stress analysis determined that, one of the piping systems passed code requirements for the increased reaction forces, without requiring any modification. The other piping system was failing to meet code requirements, although the failure did not occur on a load case involving the increased reaction forces. KHE found that the analysis was failing on a sustained load case, commonly referred to as the “hot sustained” case, or “alternate sustained” case, in which sustained loads are assessed with the pipe supports acting as they would under operating conditions. The analysis was indicating that the piping was lifting off of a certain vertical support under operating conditions, due to thermal expansion. The loss of this support when the piping lifted off of it, was resulting in a failure, when assessing “hot sustained” loads. While KHE cannot be certain as to why this issue was not addressed in the original design, it could likely be due to the fact that the ASME B31.3 has only mandated assessment of this “hot-sustained load” since 2014, well after the piping in question was originally designed. KHE specified a variable support (spring can) for use in this location that would resolve the issue, such that the piping would meet all code requirements.