Simulation Study On Dynamic Change Of Water Inflow In Underground Coal Gasification Process

With the proposal of "double carbon" strategy,the world energy is moving towards cleaner and low-carbon.Underground coal gasification,as a new force in the "double carbon" wave,can effectively alleviate the contradiction between "rich coal" and "less gas" in China.However,at present,many underground gasification experiments still have the problem of preventing gasification due to the massive influx of groundwater into the gasification chamber.In order to carry out underground coal gasification safely and smoothly,it is of practical significance to study and analyze the influencing factors of underground gasification water inflow.Taking the shallow gently inclined coal seam as the research object,this paper establishes the numerical model of underground coal gasification,analyzes the dynamic change law of underground gasification water inflow of four well layout modes under the condition of roof and floor fracture,and reveals the influencing factors of underground gasification water inflow under the condition of roof and floor fracture,as well as the influence degree and mechanism of various influencing factors on underground gasification water inflow,Combined with the actual block,the optimal well layout mode of underground coal gasification is determined.Based on BP neural network algorithm,the prediction model of peak water inflow of underground coal gasification is established.Through continuous debugging,Bayesian regularization method is selected,and the optimal prediction model is obtained by continuous training data;In order to solve the over fitting problem,the accuracy and reliability of the prediction results are verified by comparing with the real value,and the peak water inflow of the block is predicted.The paper mainly obtains the following understanding:When the roof and floor are not broken,the influencing factors positively correlated with the peak water inflow are reservoir temperature,mining depth,coal seam porosity and coal seam cleat fracture permeability;The negative correlation factors are gasification chamber width and production pressure;The injection pressure and gas flow have little effect on the peak water inflow.At the same time,the sensitivity analysis of the above influencing factors is carried out.The results show that the decisive factors affecting the peak water inflow are mining pressure and mining depth.In the case of roof and floor rupture,the aquifer permeability,aquifer thickness and bottom water energy are positively correlated with the peak water inflow.The bottom water energy is the decisive factor affecting the peak water inflow.Through the universal research on the influencing factors of underground gasification water inflow of the four well layout methods,and the numerical simulation of the actual situation in the study area,the underground gasification water inflow of the four well layout methods in this area is lower than the upper limit of the safe water inflow.Finally,the inclined well layout method is the best well layout method of underground gasification.The prediction model is used to predict the peak water inflow of the four well distribution modes in this area.The results show that the minimum mean square error MSE corresponding to the four well distribution modes is close to 0,and the square of the overall goodness of fit R is close to 1.The fitting effect is very good.The average prediction accuracy of BP neural network model is 99.17%,and the prediction model of peak water inflow of underground gasification is very accurate.This study provides theoretical guidance for the geological site selection and design parameter optimization of underground coal gasification.The peak water inflow prediction model of BP neural network also provides an important guarantee for the safe and stable operation of underground coal gasification.