神经网络方法烃类预测中的问题探讨

Issues of hydrocarbon prediction with neural network methods

石油大学地球资源与信息学院,山东东营257061

Faculty of Earth Resources and Information, University of Petroleum, Dongying 257061, China

自神经网络应用于烃类预测以来, 很多专家学者一直在神经网络和烃类预测两方面钻研, 加快了神经网络与不同数学模型相结合的步伐, 并分别取得了较好的应用效果。根据神经网络进行烃类预测的基本原理和应用条件, 简述了神经网络应用于烃类预测时的不同模式、特点、相应要求及改进思路, 并简要分析了利用神经网络进行烃类预测所存在的问题, 最后展望了神经网络烃类预测的发展趋势。

Neural network technology has been applied in the hydrocarbon prediction since 90’s. Many scholars endeavored to study the neural network technology itself and its application in the hydrocarbon prediction. Their efforts bring forth the combination of neural network with different mathematical models, and yield significant application results. According to the basic principles of neural network techniques for the hydrocarbon prediction, this article discusses briefly the different mode, characteristics, and application conditions of neural network techniques in hydrocarbon prediction, and then analyzes some typical problems encountered when neural network techniques are applied to the hydrocarbon prediction, and finally put forward a outlook for the development trends of neural network technology in the hydrocarbon prediction.