مدل پیش بینی غلظت گاز بر اساس تئوری آشوب

By means of chaos system predictability in the short term, model of coal gas concentration forecast was constructed. Based on Takens theorem, the phase space was reconstructed from the time series of gas concentration, and the optimal time delay and embedding dimention was proposed by using C-C arithmetic. In high dimention phase space, the model of gas concentration forecast using add-weighted one-rank local-region method was constructed, the real gas concentration data was analyzed, and the future data of the coal mine were forecasted. The results show that maximum Lyapunov exponent is 0.049, the time series is chaotic, and in the phase space, time delay is 7, embedding dimention is 2, the model parameter a is 0.0228, b is 1.0859, the relative error is -0.2∼0.2, and RMSE(root mean square error) is 0.0423. The predictive results tally with the real ones, which can be used to forecast the coal gas concentration in the short future.

Gas explosion is one of the major disasters threatening coal mines. Prediction of gas concentration change trend and adoption of corresponding measures to prevent gas concentration is the effective means to prevent gas explosions. At present, gas concentration prediction methods include statistics prediction method[1,2], nonlinear prediction method[3,4] and integration method[5]. The nonlinear prediction method dominates. Chaos theory is widely applied in the field of electric load, stock prices and slope displacement, etc.[6]. NIE Bai-sheng et al[7] analyzed the characteristics of electromagnetic emission and acoustic emission in coal and rock fracturing. CHENG Jian[8], CUI Xiaoyan et al[9] successfully applied the chaos theory in gas prediction. Reasonable selection of time delay τ and embedding dimension m is the key to correct prediction. This paper employed C-C arithmetic in simultaneously determining time delay τ and embedding dimension m, and used weighted one-rank local-region method to establish gas concentration prediction model used to predict the changing trend of gas concentration in a short period. The research is of important significance to preventing gas explosions in coal mines.

Based on chaos theory, weighted one-rank local-region model for gas concentration prediction was established. Short time prediction was made for the actual value of the gas concentration in the working face. The results indicate that, maximal Lyapunov index of gas concentration time series is λ=0.049>0, maximal prediction step is MaxStep=1/λ=20, the relative error of the prediction results ranges [-0.2,0.2], and mean square root error is 0.0423. The prediction results are in good agreement with the actual data. The model meets the requirement of engineering accuracy, and can be used to predict the data in abnormal conditions (e.g., coal and gas outburst, rockburst).