论文部分内容阅读
针对目前沉积微相自动识别模型和算法存在的某些不适应性 ,提出了一种基于遗传—BP算法与图像处理技术相结合的方法。根据取心井分析资料和专家解释结果确定区块微相类型 ,采用最小决策规则对模式特征指标和典型样本进行筛选 ,建立各类沉积微相标准模式库。利用图像处理技术将测井曲线和地质参数转化为图像模式 ,由神经网络自动提取和记忆曲线所表征的小层模式特征。用遗传和BP算法相结合的方法训练多层前馈神经网络 ,所得的神经网络稳定 ,学习收敛速度快 ,同时有很强的记忆能力和推广能力。对于过渡性微相在识别中存在的多解性 ,在小层对比基础上 ,参照邻井同层微相识别结果 ,在大环境下依据区块地质规律采用模糊逻辑推理方法确认和修正微相识别类型 ,保证平面沉积相和小层单井相的一致性。此模型对解决沉积微相自动识别问题具有良好的适应性。对大庆萨北油田 15口井进行了资料处理 ,取得了较好的效果。
Aiming at the incommensurate existing models and algorithms of automatic recognition of sedimentary microfacies, a method based on genetic-BP algorithm and image processing technology is proposed. According to the data of cored well analysis and expert interpretation results, the types of microfacies in the block are determined, and the model characteristic indexes and typical samples are screened by using the minimum decision rules to establish various standard models library of sedimentary microfacies. Image processing techniques are used to convert well logs and geologic parameters into image modes. The neural network automatically extracts and memorizes the small-scale model features represented by the curves. The multi-layer feedforward neural network is trained by the combination of genetic algorithm and BP algorithm. The neural network obtained by the method is stable, learning convergence speed is fast, and it has strong memory and promotion ability. In the recognition of transitional microfacies, the recognition of microfacies based on the block geology rule and the correction of microfacies Identify the type, to ensure the consistency of planar sedimentary facies and small single-well phase. This model has a good adaptability to solve the problem of automatic identification of sedimentary microfacies. Data processing of 15 wells in Daqing Sa Bei Oilfield has achieved good results.