Research On The Prediction Of Sticking Based On Time Series

Pipe sticking accident has been one of the main obstacles that influencing the drilling efficiency, in order to reduce drilling costs and reduce the occurrence of sticking accidents, it is necessary to work hard in studying on the method of predicting sticking accident. Time series modeling analysis method which has a strong data processing ability to forecast the sticking was selected based on the analysis of a large number of drilling data, and summing up causes, characteristics, prevention and solution for various types of sticking.For the prediction of sticking accidents in the process of drilling, analysed mechanism of sticking deeply, developed a new method of sticking prediction based on time series. Described specific methods and steps that using the Eviews software to establish sequence model of the drilling parameters in detail, through the time-series model to predict and process future drilling data, and for each of the ARMA model to make the power spectrum estimation, calculate the parameters of the comprehensive power spectrum superimposed deviation and a numerical simulation, as a basis of forecasting sticking.In terms of sticking type judgment, analysis the different performance of drilling parameters when sticking, different characteristic parameters corresponds to different types of sticking. Introducing the multi-factor time series modeling method, to calculate the prediction interval of main parameter, when finding change of single-factor model power spectrum of one of the parameters is unusual and the predicted value of multi-factor exceeds the prediction interval, then further to determine the occurrence of the corresponding categories of sticking.Finally, using the relevant real drilling data of the Qinghai regional exploration well and the Yulin regional geological exploration well to analyse and verify. The result shows that using this method can realize a judgment on the drilling process in future sticking accident, has good function of advance prediction and better alignment, in the actual drilling has higher scalability and application value.