摘要
准确地预测瓦斯涌出量对于指导矿井设计和安全生产有重要意义,而瓦斯涌出量是一个与自然因素及开采技术等多因素有关的非线性建模问题。鉴于传统神经网络方法解决非线性问题收敛速度慢,易陷入局部最优解的缺陷,笔者提出一种既充分利用小波变换的时频局部化性质,又能结合神经网络的自学习能力的小波神经网络预测瓦斯涌出量的方法,并建立了预测模型。在此基础上,采用Delphi语言,设计了小波/BP神经网络仿真器。通过实例分析表明该方法较传统神经网络收敛迅速,预测精度高。
Precisely predicting the amount of gas emitted from the mine, which is a matter of nonlinear model related to many factors such as nature characters and mining technology, is of great importance in the design of mine and production safety. As BP neural network has the shortcomings of slow convergence and being prone to fall into local optimums , a new method of wavelet neutral network which can make full use of part characteristic of wavelet time-frequent and combine the ability of self-study of neutral network, is presented to form a model for predicting the amount of gas emitted from the mine. Based on this model, using Delphi language, the simulator of wavelet and BP neural network is designed. The results obtained from the simulator slow that the new method can achieve faster convergence and more accurate prediction compared with that of BP network .
出处
《中国安全科学学报》
CAS
CSCD
2006年第2期22-25,共4页
China Safety Science Journal
关键词
小波神经网络
瓦斯涌出量
非线性
仿真器
预测
wavelet neural network
gas emission quantity
nonlinear
simulator
predicting
