摘要
在实际应用中发现,胡-陈-张(HCZ)模型的线性回归参数估算法存在一定的局限性,使求解出的参数值出现了一定的偏差,导致该模型预测精度降低。在研究HCZ模型的基础上,提出应用麦夸特法和全局最优算法对该模型进行非线性参数估算;并通过实例计算与线性回归法计算结果进行对比。结果表明,非线性法计算出的参数值更加准确,产量和累计产量与实际生产数据拟合效果更好,误差仅为7.59%和5.05%,可作为求解HCZ模型参数值的新方法。
In the practical application, it is found that the linear regression method to calculate HCZ model parameters has some limitation, which leads to the solving parameter values has a certain bias and the model forecast accuracy turns lower. Based on a rigorous analysis of HCZ model, a new method which including Levenberg-Marquart method and global optimal algorithm was presented to compute model parameters. Through a practical example, the new method results show a much higher accuracy than linear regression method and the predicted results of production and cumulative production accord well with the actual data, the relative error is only 7.59% and 5.05 %, thus this method can be a new way to calculate HCZ model parameters.
作者
黄全华
付云辉
陆云
童凯
HUANG Quan-hua FU Yun-hui LU Yun TONG Kai(School of Oil ＆ Natural Gas Engineering, Southwest Petroleum University, Chengdu 610500, P. R. China)
出处
《科学技术与工程》
北大核心
2016年第36期157-160,共4页
Science Technology and Engineering
基金
国家科技重大专项(2011ZX05030-005-06)资助
关键词
HCZ模型
线性回归
非线性拟合
产量预测
HCZ model linear regression nonlinear fitting production prediction
