关键词: 钻井, 卡钻, 预测分析

Abstract: A multivariate statistical approach was used to analyze the stuck pipe data and pra diet the chances of pipe sticking.Three groups of data,mechanical sticking.differential sticking and nonsticking were used to study the data and discriminant functions were found to distinguish these three groups.Then Fisher discriminant functions was used to plot the Fisher classification maps.Discussions of the discriminant functions gave the relative importance in discriminant power of each variable.The validity of the discrimipant functions was checked against field data.The test results confirmed the prediction of pipe sticking with the proposed discriminant functions.