Calculation Of Mine Tunnel Friction Coefficient Based On Multilayer Feedforward Neural Networks

As the ventilation theory basic parameter, mine roadway friction coefficient value is of great significance for mine ventilation management. Multilayer feedforward artificial neural network is the most common and most sophisticated artificial neural network, the error back propagation (BP) algorithm is easy to become the first choice due to the network before. Different roadway friction coefficient formulas are summarized. Using the regression equation solves the shotcrete roadway friction coefficient and improve its formula, providing a more accurate way to establish the mine roadway friction coefficient. Collect ventilation chart. Then compare the chart calculated results with the actual results, combining actual. Finally, prove the applicability of the formulas.In the fully turbulent region, friction coefficient has nothing to do with Reynolds, and just about the relative roughness of the pipe. BP neural network can fit the known data well, forecasting function is strong. Supporting different from the factors that influence the type of mine roadway friction coefficient of departure, select the appropriate parameters as input parameters BP neural network, collect large amounts of data, use BP neural network model to support different types of mine roadway friction coefficient pattern recognition. And use the improved Bayesian regularization method to train the networkin order to predict friction coefficient.Compare the network simulation and measured values with BP neural network, basically,error meets the engineering needs. BP neural network converge speed is fast, prediction accuracy is high. It plays an important role in the determination of mine tunnel friction coefficient, to facilitate mine shaft ventilation safety management implementation.