Analysis Of Inclination Mechanism And Research Of Inclination Prediction Technology Based On Data Mining

Inclination is a common problem in the drilling work, which directly influence the drilling well bore quality and drilling speed. From the drilling practice, we can find that it tends to be more difficult to control the inclination prevention of vertical well than the well track of the directional well. Currently, the scholars at home and abroad have developed various inclination prevention technology for the vertical well, while each technology has the limitations which can not be applicable to all the inclination problem. Therefore, most inclination prevention work for the vertical well are implemented by the expert guidance and the experience of drilling technician. Moreover, the factors influencing the inclination are various, thus making the inclination mechanism analysis and inclination trend forecasting difficult. Most inclination prevention work are passive lacking of the pertinence. On the other hand, the drilling data (shift report) recorded in the drilling process contains a large amount of valuable knowledge. However, they are just saved as the raw material for reference in most cases, thus causing the a waste of resource. Data mining is a kind of technology to extract the knowledge from a large number of data. Therefore, the inclination mechanism analysis and inclination forecasting research based on the data mining owe a great theoretical significance.The application of data mining technology on the field of inclination prevention for vertical well has been researched in this paper from two aspects:(1) Analyze the key factors for inclination;With the analysis of drilling data, it can find the key influencing factors for the inclination and obtain the inclination mechanism. It targeted selected the drilling assembly and drilling parameter, or improved the drilling structure.(2) Establish the inclination forecasting model;With the existing drilling data training, the model can get the classification rule for the inclination change trend to predict the inclination trend under the specific condition in the future drilling. The drilling worker can adjust the drilling assembly structure or drilling parameter according to the predicted result, and prevent the inclination surpassing the allowable range.Main work and research achievements obtained in the paper include:(1) Detailed introduced the operation of each step (data preprocessing, model establishing, result outputting) for the data mining in two kinds of softwares of SQL Servers2008Excel Data Mining Assembly and SPSS Clementine. With the application of two kinds of softwares, it established four models of decision tree, association rule, bayesian classification and neural network on the drilling data in Shihu Well, and analyzed the mining result of each model. According to the research objective, the mining results of different softwares and different models were compared, and their application value was evaluated.(2) With the comparison and evaluation,the bayesian model in Excel Data Mining Assembly and the decision tree model in SPSS Clementine were determined to be the most suitable method to seek the key influencing factors for the inclination and analyze the inclination mechanism. According to the mining result, the inclination in Shihu Well was analyzed to be controlled mainly by the formation force, drilling assembly and drilling pressure. Wherein, the formation deviating force played a dominant role in the change of inclination; The pendulum drilling assembly and spiral drilling rig can achieve the obvious straightening effect. However, increasing the drilling pressure may undermine the straightening effect; The pendulum force of common slick assembly is smaller than the formation deviating force, and it is unable to contain the inclination aggravation.(3) The neural network models in Excel Data Mining Assembly and SPSS Clementine were determined to have a higher classification predictive accuracy, which can be applicable to the inclination trend forecasting. On the basis, the main program for "inclination forecasting system based on data mining technology" was designed initially. The system was tested with the data in Shihu Well and achieved good predictive effect.