基于人工智能的自动送钻监控技术

Bit automatic feed monitoring by artificial intelligence.

人工智能及其应用已具有潜在的经济价值,将其应用于自动送钻过程的监控中,不失为一种新的途径。提出了把专家系统(ES)和神经网络(ANN)结合起来用于石油钻井自动送钻监控的新方法。这种方法采用预处理结合法,使钻进过程中的特征信号首先被送到BP神经网络进行处理,BP网络的输出值作为专家系统的输入,最后由专家系统通过产生式规则作出监控的结论。以卡钻事故为例的研究结果表明,对专家系统而言,知识的获取可以通过神经网络进行学习得到;对于神经网络纯粹的数值活动,通过专家系统再处理变得更加明确。这种方法可以用于自动送钻过程的在线实时监控,也可用于离线的事故分析与诊断。

This article presents a new approach to bit automatic feed monitoring based on artificial intelligence, that is, combining expert system (ES) and artificial neural network (ANN) to monitor bit automatic feed during drilling operation. The characteristic signals in drilling operation are processed by back propagation (BP) network first, the output values of BP network are taken as the input of ES, then ES draws a conclusion of monitoring by means of the production rule. The research result shows that, knowledge acquisition of ES, which can acquire through the learning of ANN, is not again a difficult problem; numerical values of ANN become more definite through rehandling by ES. This method may be used in not only on line real time monitoring of bit automatic feed, but also off line analysis and diagnosis of troubles.