中文摘要：

传统钻井动态风险评估严重依赖于专家主观判断、结果大多是定性或半定量,无法满足深井复杂地层钻井安全需求。针对该问题,研究建立了基于PSO优化BP神经网络的钻井动态风险评估方法。通过对录井资料的监测分析,实时判断井下风险发生的类型并定量计算风险发生概率,可以在风险发生的早期给出预警信息,及时指导风险调控措施的开展。海上BD气田的实例分析表明,基于构建的动态风险评估模型得到的风险预测结果与实际风险发生情况相符合,说明建立的模型是合理可行的。该模型对于钻井作业过程中动态风险评估具有一定的参考价值。

英文摘要：

The traditional methods on dynamic risk assessment of drilling rely on the subjective judgment of experts seriously, and the results are mostly qualitative or semi-quantitative,which cannot meet the needs of drilling safety in the complex for-mation of deep wells. Aiming at this problem,a dynamic risk assessment model of drilling based on PSO algorithm optimized BP neural network was established. Through the monitoring and analysis of the logging data,the type of underground risk can be judged in real-time,and the probability of the risk can be calculated quantitatively,so the early-warning information can be given in the early stage of the risk,which can timely guide the implementation of risk control measures. The case analysis of offshore BD gas field showed that the risk prediction results obtained by the dynamic risk assessment model were consistent with the actual risk situation,which showed that the model was reasonable and feasible. The model has a certain reference value for dynamic risk assessment during the drilling operation process.