论文部分内容阅读
针对钻井工程监测参数与井下复杂事故的相关性及其本身的不确定性,基于BP神经网络建立钻井工程风险的监测评估方法。首先,通过样本训练确定钻井工程风险的隐式非线性功能函数;其次,通过仿真推断相应工况下井下复杂事故的风险类型。最后,考虑到监测参数与井下复杂事故的映射关系、监测参数本身的不确定性及风险监测模型的可靠性,基于可靠性理论的Monte-Carlo方法,计算相应井下复杂事故的风险概率,并以风险柱状图描述井下复杂事故的风险。实例分析表明,用该理论方法计算得到钻井工程风险监测评估结果与工程实际基本吻合。
Aiming at the correlation between monitoring parameters of drilling engineering and complex accidents in the underground and their own uncertainties, a monitoring and evaluation method of drilling engineering risk is established based on BP neural network. Firstly, the implicit nonlinear function of drilling engineering risk is determined through sample training. Secondly, the risk types of complex accidents under the corresponding working conditions are deduced through simulation. Finally, taking into account the mapping relationship between the monitoring parameters and the complex accidents in the underground, the uncertainty of the monitoring parameters and the reliability of the risk monitoring model, the Monte-Carlo method based on the reliability theory is used to calculate the risk probability of the corresponding downhole complex accidents. The risk histogram describes the risk of a complex accident downhole. The case study shows that the result of drilling engineering risk monitoring and evaluation calculated by the theoretical method is in good agreement with the engineering practice.