Research On Health Assessment Of Marine Gas Turbine Generator

As a typical complex system,ship integrated power system is often affected by many factors during its operation.If the system is in sub-health state,it will not only cause economic losses,but also endanger the satdty of workers.The maintenance technology based on reasonable health condition assessment can accurately know the actual condition of the system,optimize the maintenance strategy,save energy and improve the stability and reliability of the entire integrated power system.Therefore,it is imminent to carry out the research of health condition assessment.Considering the power generation equipment represented by the gas turbine generator is the source of the energy of the ship’s integrated power system,the health condition will directly affect the performance of the integrated power system,so the health condition assessment system of gas turbine generator is established,and a theory and prediction model suitable for module characteristics is proposed.The main work of this paper is as follows:(1)Taking the gas turbine generator as the research object,the health condition assessment system of the gas turbine generator is established.The meaning of the health degree is clarified based on the definition of the system health condition;based on the functional subsystem,the health condition assessment architecture of module is divided into target layer,project layer,sub-project layer and assessment parameters layer;based on the improved Failure mode,Effects and Criticality Analysis(FMECA)method,the typical failure of the module and its corresponding response parameters are analyzed,then select the risk priority number(RPN)more than 12 of the response parameters corresponding to the fault as a health condition assessment index,finally,the health condition assessment parameters system of the gas turbine generator is got.(2)For the parameter level assessment,the probability density function of the varying paremeters of the assessment parameters is obtained by the non parametric estimation method.In this way,the distribution of parameter changes under various conditions is obtained.The health degree function of the assessment parameters of the gas turbine generator is obtained by combining the general form of health degree function and the calculation method of the function eigenvalue.Finally,the health assessment of the parameter level is completed.(3)For the system level assessment,the assessment strategy using fuzzy comprehensive combined with analytic hierarchy process is established.First,the membership function form is defined,and a post-processing method is proposed.Then,the weight of the same level elements is allocated according to the assessment framework constructed before.And the variable weight theory is introduced on the basis of constant power to make the weight distribution more practical.The gas turbine generator as an example,the parameter level,subsystem level,equipmental level and system level health condition assessment is completed.The assessment results show that the health condition assessment model constructed in this paper can reasonably assess the health condition of the gas turbine generator.(4)Considering the less prediction research started from the overall performance degradation trend,this paper predict the health condition of the gas turbine generator based on the health degree,and build the prediction model based on the RBF kdrnel function SVR.Taking the health degree of a gas turbine generator at a certain period of time as an example,the prediction of the health degree for a period of time is completed.And the SVR prediction model based on different kernel functions is used to predict the same data,and the prediction accuracy of the results is analyzed.(5)The test platform of the health condition assessment of gas turbine generator is built based on the LabVIEW.Through the data transmission of the simulation system,the health condition assessment system is verified,and the function of real-time assessment,monitoring and short-term trend prediction of health condition is realized.