BP�������ڴ��������е�Ӧ��

BP神经网络不同于传统的CRA（碳酸盐岩复杂岩性处理程序）测井解释方法，具有强的抗干扰能力和非线性映射能力。通过在四川盆地东北部地区铁山坡含气构造飞仙关组气藏储量计算中的应用证实，用BP神经网络计算的孔隙度与岩心分析结果有很好的一致性，能满足储量计算要求。在气水判别方面，BP神经网络计算气层段的含水饱和度与岩心分析结果相比，误差在10％以内，能指示气层特征；在水层段，BP神经网络计算的含水饱和度大于50％，能指示水层特征。表明BP神经网络可以应用于确定储量参数和判别储层流体性质。

Application of BP Neural Network Method in Estimation of Reserves

BP neural network method is different from the traditional log interpretation method of CRA (complex reservoir analysis) and it is of a strong antijamming ability and nonlinear mapping ability.Through the application in the estimation of reserves in Feixianguan Formation gas reservoir in Teishanpo gas bearing structure in Northeast Sichuan Basin,it is proved that the porosities calculated by BP neural network method are conformable with those taken from the core analysis results,therefore they can meet a demand for the estimation of reserves.The gas reservoir can be distinguished from water bearing formation:as the water saturations calculated by BP neural network method and those taken from the core analysis results are less than 20%,it is the characteristic of gas reservoir;as the water saturations calculated by BP neural network method are more than 50%,it is the characteristic of water bearing formation.Therefore BP neural network method may be used for determining the parameters of the estimation of reserves and for identifying the properties of fluids in reservoir.