Prediction of Rotating Instabilities in Low Pressure Steam Turbines Operating at Low Load

Abstract

Due to the increasing need for operational flexibility nowadays, the low-pressure (LP) steam turbine might face widely variable operation conditions up to extreme low volume flow situations. Hereby, rotating instabilities (RIS), well known from compressor aerodynamics, can initiate nonsynchronous blade excitations, which can lead to increased vibrations of the last stage moving blade (L-0R). This paper describes a numerical method able to be used in an applicable way for blade designing. To judge the onset and the shape of RIS, a numerical approach based on a transient three-dimensional unsteady Reynolds averaged Navier–Stokes flow computation of the single passage L-1 stage, a full row L-0 stage, and an axis-symmetric diffuser was established. Mesh size, solver settings, etc., were optimized to reduce the computational effort without losing prediction accuracy. The dominant harmonic compound of a spatial Fourier transform of all monitored L-0R blade forces over the blade row's circumference reveals the stall cell count and excitation amplitude for each time-step. The transient change of the corresponding phase lag provided the stall cell speed and, multiplied with the stall cell count, the excitation frequency. Their distribution over decreasing flow rates finally displays the likely onset of RIS. The approach is capable to run the analyses of several load conditions within a couple of days and was successfully validated using measured blade tip deflections and unsteady pressure probes in multistage test turbines. Furthermore, it is shown that the approach can be used to find LP blade designs, where the effects of RIS can be significantly suppressed.