Effective Radius Prediction Method for Gas Extraction Based on Adam+DNN

With the widespread application of artificial intelligence and intelligent algorithms in the industrial field, predicting the effective radius of gas extraction has become an important topic for improving mine safety and resource utilization efficiency. A method for predicting the effective radius of gas extraction based on a combined intelligent model has been proposed. The results showed that the Adam optimizer rapidly converged to 0.02 in the validation set loss value at a learning rate of 0.001 and remained stable. Further testing showed that the deviation between the predicted values and the true values of the proposed model was 1.5m, 0.6m, and 0.3m for samples 2, 4, and 10 in the first set of data, and 0.7m and 1.0m for samples 7 and 8 in the second set of data, respectively. Overall, the fit was good. In practical applications, the relative error between predicted results and measured values ranges from 1% to 6%. Therefore, the adaptability and high accuracy of the model under multi mine conditions provide an intelligent prediction tool for the effective radius evaluation of mine gas extraction process.