Study On The Method Of Some Real-time Data Processing For Electrical Condition Monitoring Of Hydroelectric Generating Sets

Along with the enhancing proportion of hydroelectricity day by day in the whole electricity system,people are increasingly concerned about the security and stability of the hydropower plantoperation and utilization. The hydropower units condition-based maintenance which isresearched and developed based on the above needs has made many contributions sincetwenty years ago, but it is also necessary to carry out all-round monitoring of thehydropower generating units, to establish collaborative mechanisms, and to research theway of fusion the multi-source, multi-temporal marked heterogeneous state data. The firstqualification to achieve that goal is to get the state monitoring data which could meet therequirements of full range of integrated co-processing.Electrical state monitoring data is very important of all the condition monitoring dataof hydropower units, it is of great significance to effectively pretreatment them for the fullintegration of co-processing. The paper summarizes the background and necessity of carryingout the study on the method of some real-time data processing for electrical conditionmonitoring of hydropower units, analysis the requirements, the current situation as well asthe lack of the electrical condition monitoring in data processing, determined from threeaspects to study, which are the phase of the alternating electrical signal self-identification andcorrective methods, the high frequency electrical signal pulse capture method, and unitelectromagnetic stability of high-precision frequency domain analysis method, clear theresearch ideas and technology roadmap. With the application of the electrical conditionmonitoring in Gezhouba Hydropower Plant, mainly study the methods of the real-time dataprocessing from these three aspects. Describing objects and their corresponding characteristics, andusing simulation technology and advanced signal processing technology, three studies are carried out,and get their individual results, including the same kinds of signals and signal phaseself-recognition and correction method, high-frequency pulse continuous sampling frequencymeter storage strategy, and high-precision frequency-domain analysis methods, at lastverified in matlab platform.