基于 BP神经网络的井壁坍塌预测

Prediction of well wall collapse based on the BP neural network

中文关键词: 坍塌预测  神经网络  井壁稳定  绥中区块

英文关键词: collapse prediction  neural network  well wall stability  Suizhong block

中文摘要:

      一定合理的井壁坍塌对于缓和井壁稳定与储层保护之间的矛盾具有重要作用。本文通过分析影响井壁坍塌的主控因素，采用 BP神经网络算法实现对井壁坍塌程度的预测。相较于依靠传统经验公式预测井壁坍塌程度，该种方法的预测结果更为科学可靠。结果表明:通过 BP神经网络模型对井壁坍塌程度的预测结果与现场实测值绝对误差率为 0.37%~14.22%，平均误差率为 7.54%。预测精度满足工程精度，证明了 BP神经网络预测井壁坍塌程度的可行性。

英文摘要:

      Some reasonable wall collapse to ease the contradiction between wall stability and reservoir protec-tion plays an important role. Through the analysis of the main factors affecting the wall collapse. Using the BP neu-ral network algorithm to predict the degree of wall collapse, than the traditional experience formula to predict the de-gree of wall collapse, the prediction result of this method is more scientific and reliable. The results show that theabsolute error rate of the prediction result of BP neural network model is 0.37%-14.22%, and the average error rateis 7.54%. The prediction accuracy satisfies the engineering accuracy, proving the feasibility of the BP neural net-work to predict the degree of well wall collapse.