Wear detection for progressing cavity pumps with system identification methods

Abstract

We present a model-based approach for a non-invasive online wear detection for progressing cavity pumps. The approach is based on a model of the rotor displacement. All unknown model parameters can be determined from measured data with a recursive-least-squares algorithm, which can efficiently be run on an embedded device. The identified model parameters provide information about the internal wear. Without the model-based approach, wear can only be analysed after disassembling the pump. We evaluate the proposed approach in a laboratory test setup with an undersize rotor, which simulates a worn pump. The results show the proposed approach can reliably monitor wear.