Font Size: a A A

Model For Discrimination Of Water-inrush Source Based On Three-dimensional Groundwater Hydrochemistry Field

Posted on:2014-05-26Degree:MasterType:Thesis
Country:ChinaCandidate:C Y ZhouFull Text:PDF
GTID:2251330401988795Subject:Environmental Science
Abstract/Summary:PDF Full Text Request
Coal is one of the important energy sources in China, and the production has been rising in recent years. However, the water disaster accidents are also on the increase. Therefore, discriminating the water-inrush source accurately and fastly is very important for preventing the coal mine water disaster and production safty.This article is on the base of a large number of literature and hydrology data. Using the water samples data with spatial attributes, the three-dimensional visualization of groundwater hydrochemistry field in Gubei mine was realized and the groundwater hydrochemistry field was analysed with the three-dimensional spatial interpolation techniques; The existing discriminating model based on the water chemistry information of water-inrush source was researched and analysed. The shortcomings of the method was pointed out. New discriminating model based on the water quality information of the groundwater hydrochemistry field was proposed and was proved to improve the accuracy;"Discrimination System Mine Water-inrush Source of Based on Groundwater Hydrochemistry Field" was designed and developed.The main conclusions and achievements of this article are as follows:(1) The real3D visualization of groundwater hydrochemistry field of main aquifer in Gubei mine was realized, and the hydrochemical characteristics and hydrodynamic characteristics of the mine’s main aquifer was revealed:for example, the concentration of Ca2+and Mg2+in the Cenozoic bottom aquifer are higher than the coal measures water, but the concentration of HCO3-in the coal measure water is higher; The water-type of the Cenozoic bottom aquifer is Cl-Na+K, most of the water samples from Coal Measures water are Cl·HCO3-Na+K type of water, and a small amount of water samples are HCO3-Na+K type of water in southeastern part of this aquifer; The Cenozoic bottom aquifer has some supply for the the coal measures aquifer; the regularity of spatial distribution of groundwater quality of coal measures aquifer is less obvious than the Cenozoic bottom aquifer, and the interpolation is a better method to calculate the water quality of any point in the Cenozoic bottom aquifer.(2)The existing discriminating model based on the water chemistry information of Water-inrush source was introduced, and tests show that this method has a higher accuracy but a lower kappa coefficient in Gubei mine, it illustrates that the effect of this method is not good enough.(3)This article shows that the discrimination of water-inrush source is a class imbalance problem generally. Virtual samples can be added for the rare class from the true three-dimensional groundwater hydrochemistry field. This method was proved to be credible,and it can be used as training samples for the training of the classifier. Compared with the original method, this method improves not only the total accuracy but also the kappa coefficient. Therefore, the improved method has effectively improve the efficiency of Bayes classifier.(4) Based on this study,"Discrimination System Mine Water-inrush Source of Based on Groundwater Hydrochemistry Field" was designed and developed.The innovation of this article are as follows:(1)Introducing the method of three-dimensional spatial interpolation, realize the visualization of three-dimensional groundwater hydrochemistry field;(2)Showing that the discrimination of water-inrush Source is class imbalance problem, and proposing a improvement solution;(3)Designing and developing "Discrimination System Mine Water-inrush Source of Based on Groundwater Hydrochemistry Field’...
Keywords/Search Tags:3D spatial interpolation, discrimination for source ofwater-inrush, hydrochemi stry field, class imbalance problem, Bayesclassifier
PDF Full Text Request
Related items