论文部分内容阅读
提出一种基于多源信息融合技术的钻井事故诊断预警方法。首先,建立钻井过程故障征兆子空间作为相应的子神经网络的输入,然后进行初步诊断,诊断所得的相关系数可反映对辨识框架中各故障的证据支持程度。再利用D-S证据理论融合各子神经网络初步诊断所形成的证据,最后得到各故障模式的信任度区间,很好地实现了钻井事故诊断识别。将神经网络与D-S证据理论相结合,既能充分利用神经网络的自适应和自学习能力,又能够发挥D-S证据理论处理不确定信息的强大推理能力,提高了事故诊断预警的准确度和可靠性。
A method of early warning of drilling accident diagnosis based on multi-source information fusion technology is proposed. Firstly, the fault symptom subspace of drilling process is established as the input of the corresponding sub-neural network, and then the initial diagnosis is carried out. The correlation coefficient of the diagnosis can reflect the degree of evidence support for each fault in the identification framework. The evidences formed by the preliminary diagnosis of each sub-neural network are fused by the D-S evidence theory. Finally, the confidence interval of each failure mode is obtained, and the diagnosis and recognition of drilling accidents are well realized. The combination of neural network and DS evidence theory not only makes full use of the adaptive and self-learning ability of neural network, but also can play the powerful reasoning ability of DS evidence theory to deal with uncertain information and improves the accuracy and reliability of accident diagnosis and early warning .