This paper, based on thin sections of coal-bearing formation of Ordos basin, studies the reservoir features and diagenesis of compact sandstone of coal-bearing formation of southern Ordos basin on cases of Memer 1 of Shanxi formation and Member 8 of Shihezi formation, Sulige area, and concludes that diagenesis largely influences reservoir. In the study area, the rock is mainly debris quartz sandstone with diagenesis of compaction, cementation and dissolution. It is concluded that the members has reached stage B of medium diagenesis combined with regional diagenesis evolution series. Comparison shows reservoir varies with diagenesis that compaction, cementation and dissolution destroy the pore structure in the early diagenesis, leading to a poor reservoir feature, and dissolution starts to produce secondary pores to improve the reservoir in the medium diagenesis.  

This paper gives an introduction to the wavelet neural network optimized by genetic algorithm which is used to predict the geothermal resources,selects volcano,earth quake,Bouguer gravity,magnetic anomaly,distance to fault and SiO2 in underground water as factors,applies excellent time-frequency transformation of wavelet neural network, network self-study and genetic algorithm to optimize the network parameter.This prediction data coincides with the real values verifying this method to be applicable in predicting geothermal resources.