Study On Crystal Structure Identification And Conversion Method Of Natural Gas Hydrate

Natural Gas Hydrate is a crystalline compounds of non-stoichiometric type which is formed by water and hydrocarbon,carbon dioxide and hydrogen sulfide and other small molecule gas under certain conditions.Mainly include type I(cubic crystal structure),type II(diamond crystal structure)and type H(hexagonal crystal structure)of three kinds of structure.To study the hydrate structure and its change rule,the formation mechanism of natural gas hydrate and the microscopic dynamics is of great significance.This paper discusses to identify the gas hydrate crystal structure types and transformation,the content of the main research work is as follows:(1)To research the methods for identifying gas hydrate crystal structure and present situation of technology research at home and abroad.Conclude that the current technology is not mature,demanding conditions and the discriminant for gas hydrate crystal structure,introduces a new method of artificial neural network model;(2)In order to establish artificial neural network,through the phase equilibrium model to calculate the sample data.Through analyze and contrast the phase equilibrium constant method,graphic method,empirical formula method and thermodynamic model method in the research methods of natural gas hydrate phase equilibrium conditions.This paper use the method of thermodynamic mode to study the gas hydrate phase equilibrium condition and calculate the accuracy of Van der Waals-Platteuw model and Chen-Guo model in the method of thermodynamic model.It's found that compared with Chen-Guo model,the error of Van der Waals-Platteuw model is smaller and the result is more accurate;Using Visual Basic 6.0 language develop "the structure testing software of natural gas hydrate",based on Van der Waals-Platteuw model and use the software to simulate the different components of natural gas under different temperature pressure generated by the structure type of natural gas hydrate(NGH)as sample data;(3)Set up an artificial neural network model to predict the type of natural gas hydrate structure.Training a variety of functions and structures of network model in order to get the best neural network model.Training function is trainlm and structure is the hidden layer.Each hidden layer containing seven node of artificial neural network has the smallest MSE values and the highest accuracy.(4)The artificial neural network model is established for the validation,in the process of validation,Leave-one-out was introduced and used to training identifying gas hydrate structure types of artificial neural network.From the results of the mean square error(MSE),the artificial neural network used to train by the method of Leave-one-out has higher accuracy and it's can be very good to forecast the gas hydrate structure types;(5)In the phase equilibrium condition,the conditions of hydrate formation should be calculate by gas-liquid-solid three phase flash calculation model,analyzing the Cole-Goodwin algorithm of gas-liquid-solid three phase flash calculation model and the algorithm of Chen-Guo.Propose an algorithm which determines whether the structure type of hydrate changes from the phase equilibrium conditions to deviate from the equilibrium conditions and how these changes happen based on the Chen-Guo aigorithm;(6)Based on the structure analysis of nature gas hydrate device system,design one multifunctional experimental device which can determine the structure type of nature gas hydrate and one device which can transform the structures of nature gas hydrate.