Research Of Fault Diagnosis Of Single Coal Bed Methane Well Based On Fuzzy Support Vector Machine

With the rapid development of our economy and the advancement of industrialization, our country has a growing demand for energy resources. And in the context of developing ecological economy, as a new type of environmental energy, coal bed methane (CBM) gets more and more attention and development. As the basic node of CBM gathering system, the normal gas production of single well is of great significance for the stable operation of the whole system. Therefore, how to diagnose the faults occurred in gas production of single well quickly and accurately is becoming an important research subject in CBM project.First, this paper briefly introduces the principle and process of single well’s gas production by drainage and the key parameters that affect its drainage. Based on the above, this paper makes a detailed analysis of several faults occurred in single well’s gas production.Second, this paper researches on the fault diagnosis method of support vector machine (SVM) in the view of small sample characteristics. And moreover, this paper gives a detailed introduction of several parameter optimization methods which include grid search method, genetic algorithm and particle swarm optimization (PSO). Then in order to eliminate noise effects on SVM modeling, we put forward an improved SVM algorithm which is fuzzy support vector machine (FSVM), and then use fuzzy C-means clustering algorithm (FCM) to calculate the fuzzy membership of training samples. Through experimental simulation in MATLAB, we establish the FSVM fault diagnosis model based on PSO to achieve the intelligent fault diagnosis of single CBM well’s gas production. Compared with the BP neural network model and SVM model, FSVM model has a better ability of learning and generalization. As a result, it has achieved good predictive diagnostics.Finally, we developed a fault diagnosis system based on FSVM for CBM wells in the platform of Visual Studio2008combined with SQL Server2005database. This system provides a concise and practical software management platform and achieve the intelligent fault diagnosis of single CBM well’s gas production by embedding FSVM fault diagnosis model successfully.