Mine Gas Emission Based On Gray Neural Network Prediction Model

Gas refers to the coal production process, from the coal, rock, and gob release the general term for a variety of harmful gases. it is one of the most important factors which endanger the security of the mining production. Also, it is one of the most serious disasters in coal mine. Predicting the gas emission in coal mine is very important work to mine safety workers. It is great significant to avoid gas ultra-limit and gas blast effectively. The law of Gas emission is very complex to the coal mine with complicated geological condition. The traditional predicting measures seem to be insufficient. Searching for new forecast methods in the original technology base becomes the focus of the current study on gas emission predicting.In this paper, takes Shanxi Jincheng Coal Mine as an example. By collecting and collating impact factor data that gas emission over the years. This pager study the main coal layer controlling factors of gas emission by grey relational analysis and let controlling factors as the gray neural network input layer nodes, the controlling factors of the gray relational grade as the input weight. Constitute the input weighted gray neural network model, so as to prediction the coal gas emission.The gray neural network weights and thresholds are randomly initialized, the network's evolution to fall into local optimum is easily, and the predicted results different each time. Using genetic algorithm optimization of gray neural network weights and thresholds can effectively avoid network by evolution into a local optimum.Finally, this article take a simulation experiment for enter the weighted gray neural network model,genetic optimization of the input weighted gray neural network model and the traditional BP neural network model, also conduct a comparative analysis. The results showed that input weighted gray neural network model of genetic, not only can simplify the system modeling, but also Improve the prediction accuracy of gas emission, having some practical value.