中文摘要：

以BP神经网络理论和测井解释为基础,测井数据作为输入,构建神经网络模型.对川中地区下中三叠统杂卤石层做精细识别,将识别结果与录井资料对比,正确率达到86.3% ,在改变约束条件的情况下正确率达到97.7% ,识别效果好;以杂卤石含量髙低对测井响应值的影响程度不同为依据,构建杂卤石层分类识别模型,模型识别正确率达到82.51% ,能较为准确且快速地识别出杂卤石层、石膏质杂卤石层和杂卤石膏岩层,与常规测井解释方法相比具有明显优势.结果表明,将BP神经网络运用到钾矿勘探中具有良好前景.

英文摘要：

Based on the theory of back propagation neural network and logging interpretation methods,a neural network mod elwith logging curyes as input was built,and applied to the polyhalite reseryoirs in the lower-middle Triassic strata. The discriminationresults were compared with logging data. The accuracy rate of the model reaches 86. 3% , and achieves 9 7 % ifchanging the constraint conditions, suggesting that the discrimination ability of the new model is good. The new model showsthe accuracy rate reaches 82. 51 % to classify the polyhalite reservoirs. The model can efficiently discriminate pure polyhalitereservoirs, gypsiferous polyhalite reservoirs and polyhalite-gypsum reservoirs, thus is more advanced than regular logging interpretationmethods. This study demonstrates the great potential applying the B P neural network in potash exploration.