--> Abstract: Forecasting Coalbed Gas Resources Amount by Artificial Neural Network, by Yang, Yuping; #90163 (2013)

Datapages, Inc.Print this page

Forecasting Coalbed Gas Resources Amount by Artificial Neural Network

Yang, Yuping

The accuracy of counting coal-bed gas resources directly influences the economic benefit of coal-bed gas exploitation. Some common computing methods are usually linear and difficult to resolve coal-bed gas resources prediction of complex geological condition areas. Nevertheless, the artificial neural network figures out this problem finely. In this work, we use the method to forecast coal-bed gas resources of central and southern of Qinshui basin. On basis of the study of coal-bed gas resources influencing factors and control mechanism, we extract four main impact factors that can be quantified as input parameters, which are coal distribution area, thickness, density and gas content. We train the samples which are the two strata of Shanxi Formation and Taiyuan Formation in nine zones for hundreds of thousands times, and establish the prediction model of the studied region's coal-bed gas resources amount using artificial neural network, with an global error of 2.21×10-4. We also use this model to conduct prediction of other zones in the studied region, and the result show that there are tremendous coal-bed gas resources in Yangcheng-Changzi, Qinshui, Anze-Qinyuan and Tunliu-Xiangyuan of central and southern Qinshui basin, and exist potential available resources. In addition, we conduct statistics to the total coal-bed gas resources of the area of 6239.2 km2 in Qinyang basin, which are 9685.63×108 m3, and the result has few difference with the result of coalbed gas resources survey results by Zhang Jiancheng et al. of Zhongyuan Oilfield. This confirm that artificial neural network could be used to the prediction of coalbed gas resources amount, and show the superiority of the method.


AAPG Search and Discovery Article #90163©2013AAPG 2013 Annual Convention and Exhibition, Pittsburgh, Pennsylvania, May 19-22, 2013