基于最小二乘支持向量机的大孔道定量计算方法

Tianjin Branch of CNOOC Ltd., Tianjin 300452, China Luming Oil and Gas Exploration and Development Co. Ltd., Shengli Oilfield Company, SINOPEC, Dongying 257000, China No.6 Oil Production Plant, Changqing Oilfield Company, PetroChina, Yulin 718600, China No.8 Oil Production Plant, Changqing Oilfield Company, PetroChina, Yulin 718600, China

大孔道的发育造成大量注入水低效、无效循环，降低了开发效果，如何定量计算大孔道已成为高含水期油田开发中亟待解决的问题。为此文中提出利用最小二乘支持向量机方法定量计算大孔道参数。从大孔道形成前后的动、静态响应特征出发，选取了大孔道的控制因素作为最小二乘支持向量机的输入，对应的大孔道参数定量解释结果作为最小二乘支持向量机的输出，通过统计训练学习，建立了大孔道与控制因素相关的定量计算模型，进而得到大孔道的定量计算结果。结果表明，支持向量机模型的计算结果与样本值拟合精度较高，能较好且客观地反映各控制因素对大孔道的影响。在岔15断块试验区应用后发现，模型的计算结果与实际监测值误差较小，满足矿场应用要求，具有较高的应用价值。

The existence of thief zone will lead to inefficient water circulation and unstable displacement, which seriously reduces the reservoir development effect. Therefore, how to quantitatively identify thief zone has become imperative for the exploitation of high water cut oilfield to improve oil recovery. A new method based on least square support vector machine is proposed. The method takes the dynamic response characteristics as input and parameter of thief zone as output. Least square support vector machine model is established through training, which can obtain the quantitative results of thief zones. The results show that the calculation results of support vector machine model fit with sample values, which can objectively reflect the influence of control factors on thief zones. The method is applied in test area of Block Cha 15. The error between the model and actual monitoring results is small, meeting the mining requirements. This method has a high application value

10.6056/dkyqt201501016