Research On Air-conditioning System Fault Prediction And Diagnosis

With the development of the world economy, the energy consumption are getting higher and higher, and the energy quantity demands of the world increase with about 2% growth rate annually, and this growth rate is over 5% per year in China, Thailand, Malaysia and Singapore. In recent years, due to the energy and environmental problems have become the development difficulties of every country all over the world, energy problem are getting more and more attentions to by every countries, China has launched a "11th five-year" energy saving plan and a long-term energy saving plan, and strives for decreasing the energy consumption in the possible shortest time. Along with the development of construction industry, building energy problems are mentioned constantly, meanwhile, energy consumption is also getting higher and higher. Statistics show that the energy consumption of building accounts for about 11% - 25% of total energy consumption energy, and the energy consumption of building in China accounts for 28% of total energy consumption. For China is one of the biggest energy consumption countries, so we must strengthen the energy conservation, due to the energy consumption of building accounts for larger and larger proportion of total energy consumption, the government has already taken the building energy conservation as an important task.Intelligent building mainly consists of building automation, office automation and communication network system, building automation is the main consumption object, and among it, the air-conditioning is main energy consumer, which accounts for more than 50% of total energy consumption. Due to the air conditioning system's complex composition, it is inevitable to appear various faults in operation process, these failures will cause the degradation of air conditioning system performance, also can use more energy. To accurately diagnose and predict the faults will help us to reduce maintenance costs and can also operate in high efficiency and save energy, therefore, air-conditioning system fault diagnosis and prediction research has significant social and economic benefits.Refrigeration system is the critical equipment of intelligent building air conditioning system, this article analyze several common faults during the operation of the air conditioning refrigeration system, and analyze the interaction of every intrinsic parameters of the machine unit as per the principle of refrigeration system, then combining with Anhui Institute of Architecture &Industry intelligent building the provincial level key laboratory existing air conditioning refrigeration system, establish several common faults'determine rules table. And according to this table, design the support vector machine (SVM) fault classifier, establish air conditioning system fault prediction and diagnosis system model based on support vector machine, and then obtained good effect for applying it into data samples prediction and diagnostic, which plays some reference and guidance roles in practical engineering application.