Research On Engine Fault Diagnosis Of Excavator In Open Pit Mine Based On Machine Learning

With the promotion of national policies,open-pit mining gradually develops to mechanization,intelligence,and green direction.Among them,the safe operation of mining machinery equipment is the top priority to ensure the normal mining of mines.As the main mining equipment of open pit mine,the excavator is prone to wear and even fatal failure because of its long-term harsh and complex operating environment and improper daily maintenance,which affects the working performance of the excavator.The traditional engine fault diagnosis technologies based on bench test and vibration signal have technical barriers in open pit mine,and the available engine fault recognition methods have the defects of low recognition accuracy and low efficiency.Therefore,based on the actual production environment of the mine,this paper determines the corresponding relationship between the fault types of the excavator engine with frequent faults and the corresponding performance parameters and constructs a robust and accurate fault diagnosis model for the excavator engine of the open-pit mine,which provides a new idea for the fault diagnosis of the mine mechanical equipment.The main work of this paper includes the following aspects:(1)Research progress and basic theory of engine fault diagnosis.We first systematically expounds the research situation of engine fault diagnosis techniques and now available fault identification methods.Then we describes the typical fault mechanism of excavator engine.Simultaneously,we summarizes the basic theories of fault identification method,data preprocessing and feature dimension reduction of excavator engine,which establishes a theoretical basis for the proposed fault diagnosis model and real fault data acquisition and processing.(2)Mining excavator engine real fault data acquisition and processing.Firstly,we analyzes the fault characteristics of the mining excavator engine,and summarizes the fault types and corresponding performance parameters of mining excavator engine with frequent faults combined with the actual mine production environment and the running state of the excavator.Then we collects the engine fault data with the help of Weichai Zhiduoxing acquisition device,and the corresponding data processing is carried out on the fault data.(3)Establishment of mine excavator engine fault diagnosis model.On account of the now available research,the defects of the previous engine fault identification methods based on machine learning are analyzed,and the Ro FGBM classification model is constructed by optimizing the Rotation Forest based on its potential defects.Simultaneously,the improved fruit fly algorithm(IFOA)is introduced to optimize the parameters of the Ro FGBM model to get the final IFOA-Ro FGBM fault diagnosis model,and the classification performance of the model is confirmed on the public datasets.(4)Application simulation of excavator engine fault diagnosis in open pit mine.The proposed diagnosis model is applied to the excavator engine fault data of the M open pit mine in Henan Province for simulation experiments.The experiment results indicate that the raised fault diagnosis model has better diagnostic performance and the fault-tolerant ability for the real mining excavator engine fault data,meets the fault diagnosis requirements of mining excavator engine,and can offer a reference for the fault diagnosis of mining machinery and equipment.