基于产生式规则系统的抽油泵故障诊断

以传统方法对有杆抽油泵系统进行故障诊断存在着很多缺陷,需要寻找可替代的诊断方法。基于人工智能领域中搜索推理技术的原理,分析了产生式规则系统的构架与工作机理。对示功图进行特征提取,将特征参数与抽油系统的抽汲参数、井况数据、故障代码、改进建议等存入总数据库。依据18种典型工况的示功图图谱,借助于领域知识与专家经验将计算所得的特征参数按各典型工况逐一比较匹配,制定推理机和产生式规则存入规则库,最终构造出产生式规则系统。从某油田随机抽取10口油井进行诊断测试,与利用BP神经网络的诊断结果对比表明,本文方法诊断结论正确,改进建议合理,该系统可在线运行,自适应性强,有助于提高采油效率、降低采油成本以及实现远程采油控制的自动化、智能化,建立数字油田。图6表3参13

Fault diagnosis of rod-pumping unit based on production rules system

Defects exist in the fault diagnosis of rod-pumping unit with traditional methods,so alternative diagnosis methods are required.According to the principle of the reasoning-based search technology used in artificial intelligence,this article analyzed the composition and working mechanism of the production rules system.After extracting features from indicator diagrams of pump working,the feature parameters,suction parameters,well data,fault codes and proposed improvements of the pumping system are input to th...