基于多源信息和智能计算的钻井异常自适应预警方法研究

Aiming at the limitations of the present abnormal monitoring pre-warning method based on compound logging technology, we study the method of self-adaptive and intelligent pre-warning for drilling abnormal conditions. Firstly, gets regional earthquake data and adjacent well data (including logging, well log, real-time drilling, well history and other information), as well as all kinds of relevant information from comprehensive logging, measurement, and well log during drilling; mix and integrate multi-sources information together through multi-sources information blend technique; eliminate the contradiction and redundancy between information. Based on the fusion information, establish real-time dynamic update geological mechanics model and real-time prediction and analysis (including abnormal lithology, abnormal pressure, etc.) models for drilling abnormal conditions before drilling, in order to the deficiency of lower forecast accuracy by traditional prediction techniques due to thinking about geological factors insufficiently; establish drilling abnormality prediction model with the function of self-adaptive evolutionary learning; and establish dynamic self-adaptive pre-warning mechanism, improving accuracy of pre-warning, trying to realize intelligent automatic pre-warning. On the basis of above, develop a self-adaption pre-warning intelligent system for drilling abnormal conditions with distributed technique under the distributed network environment, and provide a collaborative work platform for onsite or rear experts and technical personnel, making them carry out long-distance monitoring, decision analysis and abnormality control collaboratively through network, and realizing the purpose of safe, low-cost and high-benefit drilling.

针对常用的综合录井异常监测预警方法的局限性，研究钻井异常自适应智能预警方法。先获取区域地震与邻井资料（包括测井、录井、实钻及井史等信息）以及钻井过程中综合录井、随钻测量与测井等相关信息，利用多源信息融合技术进行融合与集成，以消除信息间的矛盾和冗余；基于融合的信息，建立可随钻更新的地质力学模型以及基于该模型的钻前异常预测分析（包括异常岩性、异常压力等）模型，以克服以往的预测方法中因对地质因素考虑不全而导致预报准确率不高的缺陷；利用智能计算技术构建具有自适应进化学习功能的钻井异常预测模型，并建立动态自适应预警机制，以提高预警的准确性，力图实现自动化智能预警。在此基础上，利用分布式技术研制一个分布式网络环境下的钻井异常自适应预警系统，并为井场及后方基地的多方专家、技术人员提供一个协同工作平台，使他们能通过网络进行远程钻井监控、决策分析与异常控制等，以实现安全、低成本、高效益钻井目的。

通过分析钻井工程异常的成因、影响因素及其特征，对钻井异常进行了分类，确定了钻井过程中常见的异常类型，并建立了常见异常的成因模型。针对常用的综合录井异常监测预警方法的局限性，研究钻井异常自适应智能预警方法，首先获取区域地震与邻井资料（包括测井、录井、实钻及井史等信息）以及钻井过程中综合录井、随钻测量与测井等相关信息，利用多源信息融合技术进行了融合与集成，消除了信息间的矛盾和冗余；基于融合集成的信息，建立了可随钻更新的三维地质力学模型，以及基于该模型的钻前异常预测分析（包括异常岩性、异常压力等）模型，从而克服了以往的异常预测方法中因对地质因素考虑不全而导致预报准确率不高的缺陷；利用智能计算技术构建了具有自适应进化功能的钻井异常动态识别与预测模型及方法，并建立了动态自适应预警机制，可提高异常预警的准确性，为实现自动化智能预警提供了技术基础。在上述基础上，利用分布式技术研发了一个分布式网络环境下的钻井异常自适应预警系统，并为井场及后方基地的多方专家、技术人员提供了一个协同工作平台，使他们能通过网络进行远程钻井监控、决策分析与异常控制等，从而达到安全、低成本、高效益钻井目的。