Comparison of multiple methods for identifying water sources of mine water inrush and quantitative analysis of mixed water sources based on isotope theory

Nowadays, shallow coal resources are increasingly depleted, mining depth and mining intensity to the deeper level development is inevitable, and at the same time with the increase of the depth of the complexity and variability of the underground space triggered by water accidents are also more frequent and serious. Accurate and rapid identification of the type of water source is the key to preventing and controlling water disasters, which can provide a scientific basis for rescue and management work. At present, there are many different types of mine water source identification methods, their identification effect and limitations are not clear, and there are drawbacks that cannot identify mixed water sources. In this paper, six conventional ions were selected as discriminative indices for the identification of sudden water sources based on the differences in the chemical composition of water in different aquifers in the mine. Grey correlation analysis, Fisher's discriminant analysis, cluster analysis and Bayesian discriminant analysis were used to establish the discriminant model of sudden water source, respectively. Twelve groups of training samples, three groups of test samples and one group of samples to be evaluated from different aquifers in the Longfeng coal mine were used to discriminate the water source of the outbreak using different models. The most reasonable identification method was screened by comparative analysis. In addition, based on the theory of stable isotopes of hydrogen and oxygen, a quantitative calculation method of the composition ratio of mixed water sources of mine water emergencies is also proposed. The research results show that the Bayesian discrimination method is suitable for the determination of few data samples, different proportions of classification samples and the existence of interfering data, and the classification efficiency and accuracy are higher. According to the "constancy" characteristic of stable isotopes in groundwater, linear algebra is used to calculate the proportion of mixed water sources, which compensates for the shortcomings of the existing methods that cannot identify the mixed water sources, and provides a new idea and method for identifying the water sources of water emergencies.