Coal mine accident prediction based on seasonal-WNN combination model

In order to accurately predict the evolution trend of the safety status in the mine production field, considering the seasonal factors in the actual production, this paper constructs a seasonal-WNN combination model to predict production accidents.Among them, the X-12-ARIMA model is used to adjust the safety accident sequence seasonally, and it is divided into a stationary time sequence and a seasonal factor sequence.The seasonal factor sequence show that the same law every year.The WNN model is used to predict the stationary time series, and then the multiplication model is used to restore the real forecast sequence.Based on the production safety of coal mines in our country, the monthly death toll of coal mines from 2015 to 2019 is modeled as the observed value, and the monthly death toll from coal mine accidents in 2020 is predicted, and the true value of the monthly death toll from coal mine accidents in 2020 is verified.Several typical comparative analysis of the predicted model and the constructed model.The results show that our country's coal mine accidents have significant seasonal characteristics.The average relative error of the seasonal-WNN combination model is 1.1%.The forecast accuracy is significantly better than the single forecast model, and it is more consistent with the actual trend of coal mine accidents in my country, and have a good forecasting effect.The prediction model can provide methods and guidance for the prediction of safety accidents, and can also provide a basis for coal mine safety production supervision and decision-making.