Cavitation Feature Extraction Method of Hydraulic Turbine Based on NS-TEMD and Cross Fuzzy Entropy

As a key equipment for hydropower energy conversion, water turbines may encounter cavitation problems during operation, which will lead to equipment performance degradation, efficiency reduction, and even failure. Monitoring and accurately identifying the cavitation state of water turbines is of great significance for the safe operation and performance optimization of equipment. Therefore, the cavitation feature extraction method based on Nonuniformly sampled Trivariate EMD (NSTEMD) and cross fuzzy entropy is used in this paper to analyze the cavitation phenomenon of large Francis turbines combined with acoustic emission detection method. The specific areas of inlet edge cavitation, rim clearance cavitation and vortex cavitation in the draft pipe are monitored, and various cavitation conditions and different degrees of cavitation areas of the turbine are clearly identified. The research results show that the method proposed in this paper can better capture the cavitation characteristics in the monitoring noise compared with other algorithms. The calculation results reveal that under the same working head conditions, within a certain load range, the cavitation degree of the turbine will intensify as the output power increases.