Research On Real-time Monitoring Method Of Drilling Kick

In the process of drilling,the bottom hole pressure difference often fluctuates due to the influence of the field engineer's operating level,geological environment,drilling conditions and other factors.Once the negative pressure difference occurs at the bottom of the well,formation fluid will invade the wellbore and kick will occur.After the kick occurs,the formation fluid invading the wellbore will expand due to the decrease of wellbore pressure in the annulus upward phase,especially near the wellbore,with a faster expansion rate.If well control measures are not taken as soon as possible,wellbore gushing and blowout will easily occur,which will disrupt the production order of drilling operations and threaten the safety of life and property of field personnel.Therefore,how to detect the kick in time and prevent the occurrence of well kick and blowout has become one of the key research topics at home and abroad.At present,in the aspect of early kick warning,the most widely used method in China is to monitor kick by using surface monitoring parameters.Among them,the timeliness and accuracy of threshold method for kick monitoring depend too much on the professional quality and experience of field engineers,and so it has strong subjectivity.Artificial neural network and Bayesian model also need a lot of mud logging data training work,which is inconvenient for field application.In the aspect of kick severity monitoring,when multiphase flow model is used to calculate kick allowance,the parameters required by the model are difficult to obtain,and the computational process is complex,so it is difficult to meet the requirements of rapid prediction and analysis of on-site real-time monitoring.Therefore,based on the analysis of the shortcomings of existing real-time monitoring methods,this paper uses data mining technology to analyze real-time data for real-time warning and real-time monitoring of kick severity.This paper first summarizes the causes and manifestations of kick in drilling process,then uses analytic hierarchy process to screen the early warning parameters of kick,and analyses the changing rules of early warning parameters before and after kick,and then analyses the changing rules of the characterization parameters after kick.The research on real-time monitoring method of drilling kick is divided into two parts:real-time monitoring of kick using kick warning parameters and real-time monitoring of kick severity using kick characterization parameters.Based on normal distribution model,the criterion of distinguishing control charts has great advantages in engineering anomaly detection.Five criteria of distinguishing control charts are used to identify the state of spillover warning parameters,and the corresponding table between the state value of spillover warning parameters and the probability of spillover occurrence is established.Finally,the relationship between warning parameters of real-time spillover warning model and sample data is analyzed to optimize the stability of spillover warning parameters.The established early warning model realizes the real-time early warning of kick.After the kick occurs,the kick volume,casing pressure and bottom hole pressure are selected as the evaluation parameters of kick severity according to well control regulations and empirical formulas,and the corresponding maximum allowable value model is established.Then,the real-time prediction model of kick severity development trend based on polynomial curve fitting method is established and optimized by setting model accuracy value.Finally,the real-time prediction model of kick severity development trend based on polynomial curve fitting method is established.The model predicts the kick volume before shut-in and the severity of casing pressure and bottom hole pressure after shut-in,and analyses the safety status in the future.According to the established real-time kick warning model and real-time monitoring model of kick severity,the real-time monitoring software of drilling kick is developed by using Visual Studio 2013 software programming platform and SQL Server 2008 database management platform.The real-time kick warning model and real-time monitoring model of kick severity are tested and analyzed by using the software.The field data test results show that the real-time early warning model of drilling kick based on probability model can detect the state of kick early warning parameters and find the kick in time;the real-time prediction model of kick characterization parameter development trend established by polynomial curve fitting method can predict the severity of kick in real time through kick volume,casing pressure and bottom hole pressure,and analyze the security status in the future.