基于多元回归分析的钻速预测方法研究

在钻井过程中,优化钻井技术可以降低钻井成本和减少施工事故,而钻速预测是优化钻井的基础。目前的钻速预测方程繁多,所考虑的钻速影响因素各有侧重,大多数方程做了诸多简化,预测精度不能满足现场需要;少数方程即便考虑比较全面,但由于结构复杂,可操作性差,实用性受到一定限制。以Bourgoyne&Youngs方程作为理论基础,从元坝地区钻井数据库抽取300组录井数据,利用多元回归分析方法进行钻速方程系数的求解,针对不同地层得到的回归系数也不相同。采用同区块井的钻井数据进行验证,发现其预测钻速与实际钻速基本相符,平均相对误差小于13%,说明所建立方程能满足钻井施工需求。该方法提高了钻速方程系数的求解精度,且计算过程相对简单,所用数据易于得到,对优化钻井技术的推广具有重要意义。

Study of Method for Predict Rate of Penetration Based of Multiple Regression Analysis

During the drilling process, drilling optimization technology can reduce drilling cost and construction accidents, and drilling rate prediction is the basis of optimum drilling technology. The current equations of drilling rate prediction consider different emphasis about drilling rate influencing factors, most of the equations are simplified so prediction accuracy can not meet the demand of the drilling operations, a few equations considering them comprehensively lead to poor operability and limited practicality because of complex structure. Based on the Bourgoyne & Youngs equation, the paper extracts 300 sets of mud log data from YuanBa drilling database and calculates the coefficients of drilling rate equation by multiple regression analysis. Differentiation of formation results in different regression coefficient. The drilling data of offset well is adopted to prove accuracy, the result is found that real drilling rate and predicted drilling rate are consistent and mean relative error is less than 13%. The method improves solving precision of coefficients of drilling rate equation, computing processes is relatively simple and data needed is easily obtained. Hence the method has important implications for popularizing drilling optimization techniques.