Prediction of Lateral Dynamic Characteristics of Shifting Belt Conveyor in Open-Pit Mines Using Long Short-Term Memory Network

Within open - pit mining activities, guaranteeing the precise anticipation of the lateral dynamic traits of mobile belt conveyors is of utmost significance for safeguarding operational security and enhancing efficiency. Initially, this study employs the LSTM network to elucidate the hierarchical structure of the network and the quantity of neurons arranged at each layer, and contrasts its prediction efficacy with that of the traditional autoregressive integrated moving average (ARIMA) model. The coefficient of determination (R²) value of the LSTM model is nearer to 1, indicating that the LSTM model exhibits robust data fitting capabilities and can more comprehensively account for the variability within the data. Ultimately, this paper systematically recapitulates the core content of the research, profoundly analyzes the experimental results, and anticipates potential ameliorative measures that might be executed in the future, with the aim of further refining the research.