Dissertation > Excellent graduate degree dissertation topics show

Gas Gushing Forecasting Based on Difference Grey Radial Basis Function Neural Network

Author: BaiYu
Tutor: PengXinGuang
School: Taiyuan University of Technology
Course: Applied Computer Technology
Keywords: Difference Grey Theory Radial basis function neural network Gas Emission Forecast MATLAB
CLC: TP183
Type: Master's thesis
Year: 2011
Downloads: 130
Quote: 1
Read: Download Dissertation

Abstract

Gas explosion, coal and gas outburst and gas suffocation is the major disasters that plagued the production safety of coal mining enterprises in China the gas concentration overrun leading to the direct cause of the gas disaster accident, therefore, the level of accuracy of the prediction of coal mine gas to coal mining enterprises Security plays a vital role. As we all know, gas emission by the geological conditions, Occurrence of coal seam gas content of the coal and surrounding rock, the scale of mining, mining technology, mine structure of many complex factors, and show a high degree of non-linearity between these factors, The neural network has developed rapidly in recent years, the high nonlinear mapping and parallel processing capability of gas emission modeling particularly like the BP neural network, radial basis function neural networks, etc. These methods are no doubt very suitable. However, the randomness of the sample data in the neural network has a direct impact on the accuracy of the neural network. Due to the presence of noise, the the Gas Emission historical data generally shows the chaos and the state, which fits well with the gray theory, can be used to solve the problem of neural network randomness of the sample data. While taking into account the radial basis function neural network than BP neural network in the Gas Emission Prediction advantages and radial basis function neural network training is simple, fast convergence, infinite approximation of the characteristics of the function of any form, so gray theory and radial basis function neural network combined. Gray radial basis function neural network model prediction accuracy is mainly affected by two factors, one is a combination of gray theory and radial basis function neural network method, the other is the choice of gray model and radial basis function neural network model itself prediction performance. From two points of view, the difference between the gray radial basis function neural network model, and mainly the following aspects: (1) on a combination of gray theory and radial basis function neural network weakening the sample randomness advantages and radial basis function neural network based on gray theory training is simple, fast convergence, infinite approximation of the characteristics of the function of any form of a more effective difference between the combined method, in some extent, improve the prediction accuracy. (2) taking into account the recent years, although many researchers in the field of gas forecast gray theory and neural networks combined, by various means, and indeed to some extent, improved gas forecast accuracy, but the majority just The traditional model of the two binding both one or both improved less then combined, in particular both improved after binding less, which is the bottleneck of gray neural network hybrid model one. Previous researchers better improved gray theory and radial basis function neural network, so the improved gray GM (1,1) model and improved radial basis function neural network model introduced in. (1 difference binding assay) mentioned two improved models, the difference between the gray radial basis function neural network model, so that the model can improve the prediction accuracy greater extent. Gray model and radial basis function neural network model forecast performance, we selected two improved model with good predictive performance. (3) MATLAB powerful data processing capabilities, taking into account the convenience of easy to use neural network toolbox function and good visualization environment to write code radial basis function neural network model, gray radial basis function neural network model and The difference between the gray radial basis function neural network model simulation experimental results show that the accuracy of the three models in turn improve the relative error of the difference gray radial basis function neural network model control in an ideal range of within this difference gray radial basis function neural network model has been some improvement in prediction accuracy to achieve the desired results.

Related Dissertations

CLC: > Industrial Technology > Automation technology,computer technology > Automated basic theory > Artificial intelligence theory > Artificial Neural Networks and Computing
© 2012 www.DissertationTopic.Net Mobile