Research And Implementation Of Life Prediction Of Large Equipment Assets In University Laboratory

As an important means of production for carrying out scientific research and teaching tasks in Colleges and universities,large-scale instruments and equipment are the basic conditions for the orderly progress of laboratory projects and are essential for the evaluation of scientific research environment in Colleges and universities.With the rise of industrial Internet technology,through the use of new research methods,the real-world machines,equipment and networks are interconnected through advanced sensors and software to promote the automation of information flow processing and the improvement of the ability of equipment status prediction.Based on the asset management system,this thesis uses the existing large-scale equipment online monitoring system to collect equipment operation data,and combines data processing algorithms to obtain a comprehensive health index,which is used to train the optimal prediction model to achieve equipment life prediction.Through the service life prediction of large-scale equipment in colleges and universities,it is more effective to manage the opening time setting of laboratory scientific research equipment,the formulation of maintenance plans,and the calculation of depreciation of equipment in the later period,so as to indirectly improve the efficiency of equipment use and optimize colleges and universities Management methods of laboratory equipment.As one of the classic machine learning models,Hidden Markov Model(HMM)has a good performance in modeling time series problems.In this thesis,the construction of the device life prediction model uses a hidden Markov model based on maximizing expectations.Based on the original data,the Relief F algorithm and PCA analysis method are used to perform feature screening and weight fusion to obtain a health index that characterizes the operating state of the device.It is used as the model data set to train the HMM optimal model,combined with the Viterbi algorithm to calculate the exponential probability optimal path,to obtain the health state fitting curve,to achieve the prediction of the remaining service life value of the equipment,and to verify the model prediction by combining with the simulation experiment comparison Accuracy.The construction of the prediction model uses a combination of theoretical analysis and system implementation.The final prediction model is determined through experimental results and incorporated into the prediction system.Based on the determination of the factors that affect the life of the device and the construction of an effective prediction model,this thesis designs and implements an ETL model based on big data technologies such as Hadoop,Hive,and Sqoop.Based on this model,it designs and implements a combination of Python prediction services and ASP.NET Web pages and other technologies for the prediction system of the life expectancy of college equipment.The design and implementation of the system relies on the university asset management system and fully considers the technical implementation and operation of its asset management system,which realizes low coupling to the original system,guarantees the time-varying prediction and supports the processing of multi-characteristic data patterns,so as to The construction of an asset forecasting system on the university asset management system provides examples for reference.