Fault Prediction Model And System Design For Cloud Edge Collaboration Of Industrial Internet Of Things

In the industrial field,equipment failure or even downtime may cause immeasurable economic losses.If the occurrence of equipment failures can be predicted in advance,technicians can formulate maintenance plans in advance to reduce the impact of failures.Collecting production data and predicting faults in the central cloud of the Industrial Internet of Things(IIo T)is an effective means to improve the quality and efficiency of factory production.However,the transmission of large amounts of data will cause unpredictable delays due to the limitation of network bandwidth,and will also cause huge computing power consumption and storage pressure on the central cloud.Therefore,how to design a high-performance fault prediction model and complete the fault prediction task with reasonable delay and energy consumption is an urgent problem to be solved.In order to solve the above problems,In this paper,this paper divides the fault prediction into two parts.The first part is the prediction of equipment state data through the original data,and the second part is the fault diagnosis using the predicted state data.Aiming at the application requirements of fault prediction in actual production,this paper designs a fault prediction system framework for IIo T cloud edge collaboration based on the respective characteristics of cloud computing and edge computing.In order to improve the accuracy of data prediction,this paper builds a data prediction model based on Deep Learning(DL).On the basis of CNN-LSTM,the MKCNN-LSTM model and the MKCNN-LSTM model based on the attention mechanism are constructed.Due to the introduction of multi-scale convolution kernels and attention mechanisms,the above models have significantly improved feature learning capabilities and can achieve high-precision data prediction.In order to accurately diagnose the possible faults of the equipment,this paper proposes a fault diagnosis model based on the fusion of multi-source data features based on deep learning,which improves the accuracy of fault diagnosis.In this paper,the features of singlesource data are extracted by the hybrid convolution model,and on this basis,two fault diagnosis models based on multi-source data feature fusion are constructed,namely the MFCNN model and the MFSCNN model.The two models fuse the features of the input data in different ways to make the final diagnosis more accurate.Based on the fault prediction system framework,this paper designs a fault prediction system for IIo T cloud edge collaboration,and implements it in combination with the proposed fault prediction model.The system can utilize the limited computing resources of edge computing nodes and cooperate with the central cloud to perform high-accuracy fault prediction,which speeds up the inference speed of the model.The realization of the above fault prediction system endows the fault prediction model with practical application value,and lays a foundation for equipment fault prediction in the future production environment.