Research On Performance-based Health Assessment Method For Marine Gas Turbine

Performance degradation can affect the health state of marine gas turbine.It is a hot point of current researches on health management of marine gas turbine,how to evaluate the health state of marine gas turbines in real time,quantitatively and accurately.In this study,the performance degradation of marine three-shafts gas turbines,which is mainly caused by salt spray fouling,is studied.According to the degree of relying on experience knowledge and using historical data,the health assessment technology of gas turbines is studied from three aspects of performance parameter estimation,performance data statistics and performance data relationship,which provides useful research for scientific evaluation of the health state of gas turbines.The idea has certain theoretical significance and application value.The main research contents are as follows:(1)Research on performance degradation of marine gas turbine: The mathematical model of marine three-axis gas turbine is established by volume inertia method.The degradation factor of marine gas turbine components is defined,and the performance degradation trajectory model with salt spray fouling as the main cause is established.The performance degradation simulation of marine gas turbine is realized by coefficient matching method,and the degradation influence of sensor noise,working condition and environmental temperature is analyzed.Compared with the experimental data of washing cycle,the mean square error is less than 0.7,which indicates that the performance degradation model is consistent with the actual degradation process.The simulation data of washing cycle and dynamic working conditions are obtained,which provides a data basis for subsequent health assessment.(2)Research on health assessment method of marine gas turbine based on performance parameter estimation: A health assessment method based on monitoring parameter estimation is proposed.The method is based on multi-model idea.The degradation data in multiple washing cycles are modeled according to the degree of temperature deviation of exhaust gas after low pressure turbine.The estimation of monitoring parameters of marine gas turbine under various working conditions is realized by using generalized regression neural network.Mahalanobis distance is used to recognize the health state of marine gas turbines.A health assessment method based on component parameter estimation is proposed.The method is based on generalized regression neural network to estimate component parameters and principal component analysis to recognize the health state of marine gas turbines.The simulation experiment of marine gas turbine health assessment is carried out.The results show that the health assessment method based on monitoring parameter estimation can achieve the health state assessment under the condition of accurate model,and the health assessment method based on component parameter estimation can achieve the accurate health state assessment.(3)Research on health assessment method of marine gas turbine based on performance data statistics: A hierarchical health assessment method of marine gas turbine is proposed.The method realizes parameter-level health assessment based on probability density function estimation.Fixed weight and variable weight analysis method are used to realize component-level and system-level health assessment step by step.The simulation experiment of marine gas turbine health assessment is carried out.The results show that the health assessment method based on performance data statistics can reasonably assess the health state of marine gas turbine in the case of massive operation data but no accurate model.(4)Research on marine gas turbine health assessment method based on performance data relationship: From the perspective of data internal relationship,a multi-parameter time series correlation network and a single-parameter time series visual graph network are established.Based on CNM community detection algorithm,a multi-parameter time series and a single-parameter time series health assessment method for marine gas turbine is proposed,and the health assessment for marine gas turbine is carried out.The simulation results show that the method of marine gas turbine health assessment based on performance data relationship can realize the rapid assessment of marine gas turbine health state,reduce the dependence on prior knowledge,and the evaluation effect of single parameter time series network is better than that of multi-parameter time series network.