基于RBF神经网络的页岩气丰度预测

Prediction of shale gas abundance based on RBF artificial neutral network

对页岩气成藏条件进行分析认为,机质含量、类型、成熟度、储层厚度、埋藏深度和孔隙度等为影响页岩气成藏的主要因素;依据四川盆地和北美页岩气形成的地质条件的相似性和所收集的大量页岩气数据,利用RBF神经网络构建了各主要因素与资源量丰度的网络模型,并用MATLAB对网络模型进行训练。对Ohio和龙马溪组页岩资源丰度进行了预测,预测误差分别为27.5%和3.5%,表明RBF神经网络模型可较好预测页岩气产量,对页岩气开采和勘探具有参考价值。

According to the analysis of shale gas accumulation conditions,it is held that the content,type and maturity of organic matter,and the thickness,buried depth and porosity of reservoir are the main factors of influencing shale gas accumulation. The relationship model between shale gas abundance and the factors is established using RBF neural network according to the similarity of the geological conditions of Sichuan Basin to the forming conditions of North American shale gas and a large number of shale gas data,and the model is trained with MATLAB. The error between the prediction results and the actual values of the shale resource abundance of Ohio and Longmaxi formation in Sichuan Basin is 27. 5% and 3. 5% separately,which shows that the prediction of shale gas production based on RBF neural network model can provide practical value and theoretical guidance for the shale gas development and exploration.