脉冲试井模型多参数自动拟合方法

脉冲试井是干扰试井的一种，参数解释常规方法可归纳为切线法、割线法、图版拟合法及参数匹配方法，以上方法解释过程均是运用了生产数据中一些特殊点进行图版解释，因此存在数据利用率低、解释速度慢、拟合精度不高等缺点。建立脉冲试井井储表皮模型，结合非线性回归技术，提出有界信赖域优化分析方法求解脉冲试井模型中的储能系数、流动系数、表皮因子、井筒存储系数等6项参数。用该方法拟合某油藏脉冲试井生产数据，得到拟合值的最大相对误差为6.49%，平均误差为2.98%，同时该方法是自动拟合技术，不受人为因素影响，并具有超线性拟合速度。

Nonlinear optimization of model multi-parameters method for pulse well test

Pulse well test is one of interference well tests and conventional methods used to explain parameters can be summarized as tangent method,secant method,plate fitting and parameter matching method.The above methods are all using some special points selected in the production data to match the standard plates when explain the parameters.These methods are characterized by low utilization rate of data,slow explaining rate and poor fitting accuracy.Aiming at these disadvantages,pulse well test wellbore storage and skin factor model was established in this paper,combined with nonlinear regression techniques,an optimal analysis method using bounded trust region was proposed to solve 6 parameters in pulse well test model which include energy storage coefficient,flow coefficient,skin factor,wellbore storage coefficient.Pulse well test production data of one reservoir was fit with this method,and the maximum relative error of fitted values is 6.49% and the average error is only 2.98%.Meanwhile,the method belongs to automatic fitting technique,not affect by human factors,and has super-linear fitting speed.