应用神经网络技术诊断钻井事故

Diagnosis of drilling faults using neural network technology

针对钻井过程的复杂性、不确定性等特点,提出了基于人工神经网络钻井事故预测与诊断模型.以钻井过程工况参数构成神经网络的输入特征向量,以钻井过程正常运行模式及常见事故模式的监测数据作为训练及检验样本,根据钻井事故诊断特点,确定了神经网络的结构与参数,采用改进算法和学习规则,实现对神经网络系统的训练和模拟,建立能够准确预测事故的神经网络模型.该方案的提出可使现场工作人员及时监测钻井过程,降低事故发生率,节约钻井成本,提高效率.

Because of the complexity and uncertainty of drilling process,it is proposed that the prediction and diagnosis of drilling accidents are carried out by artificial neutral network.The operating parameters of drilling process are taken as the inputs of the network.The monitoring data of the normal drilling condition and the common fault conditions in drilling process are taken as the learning and test samples.The structure and parameters of the network are determined based on the characteristics of the fault diagnosis of the drilling process and the abundant data from drilling spot.The improved algorithm and learning rule are used to train and stimulate the network system.The established neural network model can precisely forecast and diagnose the common drilling accidents.The presented method can make the field workers monitor drilling process in time,which will decrease accident rate and drilling cost,and increase drilling efficiency.