The Design And Development Of Online Monitoring And Warning System For Kick Foreboding On "Three High" Wells

Blowout is the most serious accident during drilling process. Kick is the foreboding of blowout, hence it is prerequisite and guaranteed to protect blowout and detect kick for drilling safely. While the traditional methods of kick detection are simple and available, their instantaneity and reliability is so poor that they are unable to meet the need of safely drilling. In order to improve the instantaneity and reliability of kick detection, the thesis concluded and analyzed the existing kick detection methods, kick identification methods and the system of kick detection. And it described the causes of kick, the signs of kick and other accidents, and made comparisons about the instantaneity and reliability of kick detection parameters. On that basis, an online monitoring and warning system for kick foreboding on “Three High” wells was designed, and the system software was completed through VC++6.0.Regarding the poor instantaneity and effectiveness of kick detection by conventional means which only one parameter or one mode was used, a new method was proposed that made use of three kinds of parameters which include micro-flux, PWD and comprehensive logging parameters, and applied the multi- parameters, multi-means and verified mutually mode to detect kick. And it is based on monitoring the change of down-hole annulus pressure, combines with the changes of mud flow and comprehensive logging parameters. It could discover kick in advance and improve the instantaneity and reliability of blowout warning before the formation fluid backing out of the ground. According to the anomalous changes of monitoring parameters, it could detect not only kick, but also seven other accidents that are related to kick, such as lost circulation, drilling break, sticking, drilling tool washout, string parted, bit jets losing and bit jets blockage.An identification method of kick and other drilling accidents based on Expert System in combination with the hierarchical Bayesian discriminant model was proposed in the thesis. Because of lacking field data, the training of the hierarchical Bayesian discriminant model could not be completed. So Expert System is applied to indentify kick and other drilling accidents in the initial application, while field data is stored for training the hierarchical Bayesian discriminant model. If the field data is enough, Expert System combining with the hierarchical Bayesian discriminant model is applied to identify kick and other drilling accidents. Regarding the defects of the traditional Bayesian discriminant model, such as training complex, requiring the attribute variables completely independent and lacking of flexibility, an identification method of kick and other drilling accidents which is based on the hierarchical Bayesian discriminant model was presented. The application of the hierarchical Bayesian discriminant model is more flexible, and the training is easier. It could also reduce the misjudgment caused by the attribute variables which are not independent completely, and improve the identification accuracy.The laboratory test and field experiments results showed that the system could detect kick and other drilling accidents instantaneously and effectively, and the system software could operate normally and stably.