Research For Operation State And Maintenance Cost Prediction Method Of High-end Equipment For Operation And Maintenance Process

With the proposal and continuous promotion of "industry 4.0" and "made in China2025" strategies,the research of high-end manufacturing systems and advanced intelligent equipment is receiving more and more attention.The high-end equipment manufacturing industry has become an important force driving China’s industrial transformation and upgrading.High end equipment has also been applied to various fields such as industry,electric power,medical treatment and so on.High-end equipment has the characteristics of complex structure,numerous components,and mutual coupling between components.In case of failure,it may cause huge losses.Therefore,its safe and reliable operation has become a research hotspot.Preventive maintenance and post-event maintenance of high-end equipment based on the predicted results of operating state and maintenance cost can effectively reduce the risk of failure and improve the stable and reliable operation of high-end equipment.In addition,it can also reasonably formulate maintenance strategies after a fault occurs and saving maintenance resources.Remaining useful life prediction,fault diagnosis and maintenance cost prediction are three key links for safe operation and scientific maintenance of high-end equipment.Affected by the mutual coupling between the components of high-end equipment,its monitoring signal shows strong nonlinearity and nonstationarity.There is also a complex mapping relationship between the monitoring signal and the operation state.In addition,its high reliability and long life make the samples incomplete and the fault samples insufficient.Therefore,based on the quality characteristics of the operation process,the study of the state prediction method has important engineering application value to ensure the safe operation and scientific maintenance of high-end equipment.Taking high-end equipment as the research object,based on the quality monitoring data generated in the operation process,this paper studies the operation state and maintenance cost prediction of high-end equipment.First,aiming at the nonlinearity and non-stationary of data series in the degradation process of high-end equipment,we study the remaining useful life prediction method of high-end equipment.Then,considering the problem of low recognition accuracy of fault samples caused by the imbalance between normal samples and fault samples of high-end equipment,we study the fault diagnosis method of high-end equipment.Finally,considering the low prediction accuracy of high-end equipment maintenance cost in the case of small samples,we study the prediction method of high-end equipment maintenance cost based on quality characteristics.The specific research contents and innovations of this paper are summarized as follows:(1)The remaining useful life prediction of high-end equipment based on time series analysis is studied.The degradation process data series of high-end equipment show strong nonlinearity and non stationarity.Therefore,the remaining useful life prediction of high-end equipment is a typical nonlinear time series prediction problem.In order to identify the potential operation state of high-end equipment,the phase space reconstruction algorithm is used to reconstruct the data space of degradation process.A variable neighborhood search algorithm is designed to dynamically select the important parameters of the prediction model.Using the reconstructed data space,a remaining useful life prediction model based on support vector regression is established.Finally,based on the background of the aero-engine operation and maintenance process,the remaining useful life of the aero-engine is predicted,and the performance of the prediction model is verified.(2)The problem of high-end equipment fault diagnosis for unbalanced samples is studied.A large number of data are produced in the operation monitoring process of high-end equipment,which brings challenges to the prediction accuracy and calculation efficiency.The compressed sensing theory is used to compress the signal,which reduces the amount of redundant data and improves the speed of data processing.In practice,the probability of failure of high-end equipment is small.There is a serious imbalance between normal samples and fault samples in the samples.An adaptive mixed sampling method based on K-means clustering is proposed to improve the identification ability of the model to fault samples.A fault diagnosis model based on Bi-directional LSTM is established on each sub sample,and the weighted voting method is used to integrate the model.Finally,based on the background of the aeroengine operation and maintenance process,the fault of the aero-engine is predicted,and the performance of the prediction model is verified.(3)The point and interval prediction of maintenance cost of high-end equipment considering quality characteristics is studied.There are many quality characteristics in the operation process of high-end equipment,and there is a linear relationship between quality characteristics.An improved recursive feature elimination algorithm is proposed for feature selection of quality characteristics,which can remove the redundant information in quality characteristics while retaining the original attribute characteristics of data.The point and interval prediction model of high-end equipment maintenance cost based on residual bootstrapped support vector regression is established.The particle swarm optimization algorithm is used to optimize the model parameters.The interval prediction performance of the model is verified by PICP,PINAW and CWC.Finally,based on the background of the aero-engine operation and maintenance process,the maintenance cost of the aero-engine is predicted,and the performance of the prediction model is verified.This paper deeply studies the state prediction problem of high-end equipment in the operation and maintenance process.Three new high-end equipment state prediction models are constructed.Some innovative research results have been achieved.The research in this paper provides a decision-making reference for the safe operation and scientific maintenance of high-end equipment,and has important theoretical and practical significance.