Research On Container Health Prediction And Multi-Objective Maintenance Decision Technology Based On Big Data

With the implementation of the "The Belt and Road" strategy and the rapid development of multimodal transport,the market share of railway freight transport will become larger and larger.As an important carrier of railway transportation,containers must be guaranteed to be in good condition.However,the traditional "post-repair" and"planned maintenance" methods cannot guarantee timely and effective maintenance of the container and cause unnecessary losses.The development direction of future equipment maintenance guarantees will be based on "disregarding maintenance" and"preventive maintenance" as the future development direction.Based on historical maintenance data and freight data of containers,this paper conducts research on container health prediction and maintenance decision optimization.The main contents are as follows:(1)For the container repair,it is impossible to quickly and effectively locate the damaged parts.This paper introduces the health index concept to characterize the health status of key components and establish a health index calculation model.Firstly,the indicator data related to the health index of the key components of the container is selected from the container historical maintenance data and the cargo data.Then the health index of the key components is obtained through the health index calculation model.XGBOOST prediction algorithm is used to predict the health index of key components with strong generalization ability and fast calculation speed.Then,for the problem that the parameters of XGBOOST prediction algorithm are difficult to adjust,the simulated annealing algorithm is used to find the XGBOOST prediction algorithm.Parameters to complete the health index forecast for key components.(2)For the problem that the "post-repair" method currently used by the container is inefficient and the container health status is not high enough,this paper introduces the health index concept to characterize the overall health of the container and establish a health index calculation model.Using the health index of the key components obtained in the previous section and its predicted value and the weight of the health weight of each key component,the overall health index of the container and its predicted value are obtained,and the container to be repaired is judged based on the predicted health index of the container as a whole.At the best time,take the initiative to repair.(3)In view of the high maintenance cost of the container and the insufficient maintenance effect,this paper establishes a multi-objective maintenance decision-making optimization model for the container to reduce the maintenance cost and improve the maintenance effect as the main objective function.The NSGA-II algorithm is used to optimize the maintenance time.The health status of the point solves the optimal combination of maintenance methods for dangerous critical components,thereby optimizing container maintenance decisions,reducing maintenance costs and improving maintenance.(4)For the key component health index calculation health prediction process and container health index calculation,health prediction process and multi-objective decision model solving process long time,this paper uses Spark big data technology to optimize these processes,Reduce its running time.Through these research contents,the time and location of the container to be faulted are predicted in advance,the remaining life is predicted,the container usage rate and the use safety are improved,and the container maintenance effect and maintenance efficiency are enhanced.