Research on drilling kick and loss monitoring method based on Bayesian classification

The kick and loss accident is the critical risks of drilling engineering,and that will seriously affect the speed of drilling,quality of well construction and efficiency of exploration. to this end,this work is based on the characteristics of the pressure transmission speed is faster and the correlation between kick & loss and change trend of standpipe pressure (spp) and casing pressure (cp),the change trend is used for monitoring the kick and loss accident. in this research,the value of neighbor well accident prior probability is combined with the spp and cp of neighbor well whose block,horizon and drilling way is similar to target well,then extracting the pressure variation characteristics value vector,using the multidimensional variable normal density function for conditional probability density function (df) and obtaining discriminant function based on bayesian statistics method,finally using the df to monitor the drilling status (ds) of target well. the actual field application showed that the monitoring system is timely,accurate,and able to meet the needs of field applications and provide effective technical support for field kick and loss accident monitoring,it can be used as a new monitoring method of ds. © 2014 pakistan journal of statistics.