Gas Emission Prediction of Working Face Based on PCA-FOA-GRNN

Working face gas emission quantity affected by many factors, it is difficult to accurately predict through a linear method. Generalized regression neural network is a feed-forward neural network, with good robustness and high fault tolerance, and only one adjustable parameter, so we adopted GRNN to build predictive model. Then the improved FOA was applied to optimize the traditional GRNN model. Meanwhile, the PCA was adopted to simplify the sample data to reduce interference by secondary factors on prediction result. Chosen Xiaoming mine data to validate the model, the prediction is good, and the average absolute error was 3.45%, less than 10.06% of traditional GRNN model.