On-line Cavitation Monitoring And Analysis For Large Turbines--Methodology & Application Studies

Cavitation erodes equipments and decreases efficiency of turbines. The investigation on cavitation has been made for long time. However, cavitation in hydro turbines is still a focused and difficult issue due to its complexity and multidisciplinary involvement. Based on summarizing the existing researches on the cavitation, this dissertation is focused on on-line monitoring and analysis of turbines'cavitation in Gezhouba Hydro Power Station. The methodology and technique are comprehensively investigated on integration of theory with practice.The mechanism, types, and location of cavitation as well as its external features and influence factors are analyzed in detail after the existing researches have been reviewed. Accordingly, based on the association with the operating states of turbine-generator sets, the cavitation sound wave detection, including audible sound and ultrasound, are proposed to meet real time, reliability and non-destruction of on-line monitoring, which is meaningful for the accurate evaluation of cavitation, especially for the identification of cavitation inception. Considering the structure of turbine-generator sets, cavitation source, flow stability, the covering range of sensors as well as the reliability and the convenience of mounting sensors, the sensors are positioned. Besides, a programmable signal modulation module is designed based on high reliability after detailed investigations on the characteristics of sensors, the signal transmission and field application. The bandwidth and the amplification of the module are programmable. The power supply of the sensors is included and is protected from short circuit.The traditional monitoring systems only accomplish data acquisition (DAQ) and signal characteristic extraction. The evaluations of turbines'condition are mainly implemented in dependence on off-line analysis. Based on modeling the relationship between turbine system structure and its behavior, the test process are extracted or imitated in the normal operation of turbine-generator sets according to the field test rules and experts'experience. Consequently, the turbine performance is automatically estimated via identifying the operating states. Furthermore, the statistic of operating states of turbine-generator sets is carried out so as to optimize turbine operation. In addition, the integrate cavitation intensity is calculated periodically with respect to the large Kaplan turbines in Gezhouba Hydro Power Plant, which can figure out the degree of cavitation erosion approximately. And pivotal concerns are discussed on the evaluation of the metal loss caused by cavitation through analyzing the relative researches.Aiming at the cavitation monitoring for turbines in Gezhouba Hydro Power Station, an on-line cavitation monitoring and analysis system based on sound wave detection has been developed, which is composed of an on-line cavitation monitoring and analysis unit for hydro turbines (TrbMAU) and a mobile workstation. the mobile workstation is not only an off-line analysis platform, but also is an operating platform for TrbMAU, for example, data sampling can be controlled and relevant parameters can be set according to the needs of clients and field tests. Particularly, considering the influence of the operating states on cavitation, adaptive data acquisition (DAQ) and storage is proposed to capture all sound features in different operating states with less data redundancy. The DAQ period varies with operating states and the storage mechanism allows for operating states and indexes of cavitation. Moreover, the sensors'health is diagnosed automatically to ensure the data creditability. TrbMAU implements as a part of Optimal Maintenance Information System for Hydro turbines (HOMIS) in Gezhouba Hydro Power Plant. Therefore, the operating states and time are synchronized in HOMIS so as to comprehensively estimate the health of turbine-generator sets.In terms of the failure modes of Kaplan turbines and the influence factors of cavitation, model tests and field tests have been implemented in order to build the knowledge database of cavitation characteristics. Considering the frequency bandwidth analysis technique based on wavelet decomposition, the characteristics of wavelet basis that is applicable for the processing of sound wave emitted by cavitation are summarized according to the characteristics of these signals. Moreover, local decomposition of wavelet packet is proposed to improve the efficiency of calculation. The frequency bandwidth analysis technique has been successfully applied in analyzing monitoring results in model tests. The relationships are achieved between the amplitude modulation in time domain as well as energy distribute in frequency domain and cavitation coefficients. And the characteristic sound waves are extracted and the influence of turbine servicing time on cavitation is analyzed via field tests.TrbMAU has been operating on turbines in Gezhouba Hydro Power Plant for more than two years. Thus the characteristic knowledge on cavitation has been achieved. The cavitation intensity on different water head and power has been traced out and the degree of cavitation erosion is estimated according to the cavitation intensity at a fixed operating condition. Especially, some phenomena obtained in field tests have been solved with the results of model tests.Since October 2005, TrbMAU has been successfully put into on-line operation on 6 turbines in Gezhouba Hydro Power Plant in succession. The monitoring and analysis results can describe the cavitation level well. All the achievements are meaningful for carrying out condition-based maintenance and improving cavitation research on hydro turbines.