应用人工神经网络自动识别岩性

应用人工神经网络(ANN)模型,输入测井数据自动识别岩性,重要的问题是改进传递函数,提高收敛速度。文中导出的人工神经网络误差反传递算法,较之传统的统计方法具有较高的容错性,广泛的适应性及速度快的特点。用这种方法的识别结果与统计方法、岩性录井分析的结果一致,说明该方法是行之有效的。

Automatic lithology identification using artificial nervous network

An important matter for our using an artificial nervous network (ANN) to automatically identify lithology from inputted logging data is to improve both transfer function and convergence rate. The ANN error inverse transfer algorithm has higher allowance，wider adaptability and faster computation than old statistical ones. This method brings the same result as statistical and lithologic log methods have，which proves it effectual.