Study On Coupled Vibration Characteristics And Response Prediction Of Underground Powerhouse

The capacity and size of modern hydropower generator units become largerwith the rapid development of the hydroelectric enterprise. The vibration ofpowerhouse structure has increasingly become a major problem. For the sake of theparticularity of the underground powerhouse’s structure in hydropower plant, moreemphases should be placed on the generator set and structure. But because of thecomplexity of the system dynamics, nonlinear coupling, random characteristics andetc., at present, the study on dynamic characteristic of the powerhouse structure,vibration-source identification and vibration response prediction in the problem isstill inadequate. So, it is necessary to make further study on problems above. Basedon prototype observation data of Ertan Hydropower Station, and using technology ofprocessing dynamic signal, finite element analogy and intelligence algorithm, thepaper studies the problem of hydropower plant vibration. The principal contents andachievement of this dissertation are as follows:(1)According to field test of underground powerhouse in Ertan HydropowerStation, data of vibration is analyzed using technology of processing dynamic signal.Study the vibrating law of generator set in conditions of variable-speed,variable-excitation and variable-load, and then identify vibrating attributes andoperational status of the generator units. Depending on the vibrating law ofgenerator units and powerhouse, make a preliminary division for the operationalranges of the tested generators which could be a dispatching reference for the powersystem and power plant.(2)Identify the vibration sources, depending on field testing results ofunderground powerhouse. Study the influences of various dynamic loads for thepowerhouse structure and find out the key one. Then evaluate the effects of thevarious dynamic loads resulting in the vibration of the underground powerhouse.(3)The modals of the model of powerhouse structure are calculated by thethree-dimensional finite element numerical analogy. Then check the source ofresonant excitation of the generator units and propose reasonable measures foranti-vibration. (4)Because the pressure fluctuation of tail water always happens with vibrationof generator units, the relations between pressure fluctuation and vertical vibrationof units are mainly analyzed. Through the analysis of the relationship between unit-vibration and powerhouse-vibration, the nonlinear coupling laws of vibration arefurther discussed.(5)In accordance with the coupling effect between vibration of units andpowerhouse, two models for forecasting response of the vibration of the powerhouseare built based on BP neural network and LS-SVM respectively. Then verify theprecision of the two forecasting models through the testing data of vibration in ErtanHydropower Station. Based on this, the paper firstly introduced the running waterhead as an input divisor into the intelligent prediction model and the forecastingrange is extended. At last, the design mind for building prediction system ofvibrating response of powerhouse structure is proposed based on LS-SVM.