Research On Fault Diagnosis And Parameter Prediction Of Coal-Bed Methane Single Well System

With the development of industrialization, our economy and energy structure has changed. The change of energy structure and demand of clean and high quality energy makes the status of coal-bed methane increase. Coal-bed gas resources in China are very abundant and gas reserves are ranked third in the world, however the technology for coal-bed methane extraction is not developed enough. In the event of an accident during the extraction, serious economic expenses and personnel casualties may be incurred. Therefore, the research on the effective use and safety of the technology is the important to be focused on. The coal-bed gas transfer system is composed of many single wells, and the coal-bed methane single well safe production affects the whole coal-bed methane system function. Therefore, this paper is mainly aimed at the fault diagnosis and parameter prediction of coal-bed methane well drainage system.The research object of the present paper is the Shanxi Province JinCheng city coal-bed methane project. This paper mainly introduces the basic principle and the technological process of single well gas extraction system in the coal-bed methane. This paper analyzed the relationship between the various faults in the coal-bed methane system and the trend of key parameters in the process of single well, coal-bed methane single well gas recovery system has the characteristic of running stability and gas extraction parameters changing quite gently, this paper introduces the knowledge and modeling process of support vector machine. Select the appropriate support vector machine model parameters have important implications for the diagnosis model, this paper introduces several support vector machine parameters optimization methods, analysis both the advantages and disadvantages of the three optimization algorithms. In order to increase the diversity information of training samples, this article combined the process of selection and crossover operator of Genetic Algorithm and Particle Swarm Optimization, and improved the speed and position formula of particle swarm optimization algorithm, overcome the problem that Particle Swarm Optimization is easy to fall into local extremum. In order to achieve the fault diagnosis for coal bed methane system, this paper establish the support vector machine fault diagnosis model based on improved particle swarm optimization. Through the simulation experiments in Matlab, compared with other diagnostic methods, the proposed method has higher accuracy and stability.Grey theory ignores some uncertain factors in the process of dealing with gray system, grey theory weaken volatility of the parameters series, enhanced regularity of the time series data. Support vector machine based on the statistical learning theory, the prediction accuracy is very high when dealing with small sample system, but it's very easy to falling into over-fitting phenomenon which can affect the generalization ability of the model. In actual production, the coal-bed machine production system is influenced by many internal and external factors, some factors are transparent, but there are also have some vague and unknown factors. Therefore, according to the coal-bed methane system with grey fuzzy unknown factors and there are only a small number of samples and other issues, For this situation, this paper established support vector machine model based on improved particle swarm optimization. The simulation results show that comparison with other forecast methods, this method has higher forecasting accuracy.