基于改进的反向传播神经元网络复杂煤岩层对比方法研究

为了提高复杂地区煤岩层对比的准确率,解决由于反向传播神经元网络(BP神经元网络)连接权值和阈值的初始值选择不合适而导致的无解问题,本文把求全局最优解近似值的遗传算法(GA)和求局部最优解精确值的传统BP神经元网络所使用的梯度法有机地结合起来,取长补短,用于复杂地区煤岩层对比。首先,用GA求得BP神经元网络权值和阈值的全局最优解的近似值;然后,把该近似值作为初始值,训练该神经元网络;最后,用训练好的BP神经元网络进行复杂地区煤岩层对比工作。本文利用测井曲线采用小波变换分析沉积旋回,并用于煤岩层对比。该新方法在钱营孜煤矿的复杂煤岩层对比中,取得了较好的效果,比常规的单一方法和简单的组合方法效果要好。

Research on correlation of complex coal and rock seams based on improved back propagation neural network

In order to improve the detection rate of correlation of complex coal and rock seams and solve the problem that the back propagate neural network （ BP neural network） is invalid when initial weight and threshold values of BP neural network are chosen impertinently, Genetic Algorithms （GA） ＇ s characteristic of getting whole optimization value was combined with BP＇ s character- istic of getting local precision value with gradient method. After getting an approximation of whole optimization value of weight and threshold values of BP neural network by GA, the approximation was used as the first parameter of BP neural network to train the BP neural network again. The trained BP neural network was used to correlation of complex coal seams. The accuracy of correlation of complex coal seams can be improved by the new method. In the paper, the sedimentary cycles with muhiscale characteristics based on wavelet transform of logging data was applied to correlation of complex coal seams. The practical experiment results shown that this method was useful and applicable for correlation of complex coal seams of Qianyingzi＇ s coal mine.