Research Of Fault Diagnosis Based On HHT-WNN In Air-conditioning System

With the advances in modern science and technology and the improvement of people’s living standards, air conditioning system, which is short for HVAC (Heating, Ventilation and Air Conditioning) system, and its control strategy have become more and more complex. Thus the requirements of system control are increasing for protecting the normal operation of the air conditioning system.In modern commercial buildings, HVAC system energy consumption accounts for about50%of the total energy consumption, even up to70%in some areas. However, indoor air pollution has a direct impact on people’s physical and mental health, even their lives in recent years. The research shows that air-conditioning systems can achieve a20%to30%energy saving effect when adding the fault detection and diagnosis (Fault Detection and Diagnosis, FDD) in commercial buildings. Therefore, for the purpose of achieving energy efficiency and improving indoor air quality, it is necessary to join fault diagnosis link in the design of the air-conditioning system.Sensor is the bottom element in the system,which is widely used for the air conditioning system. There are many sensors such as temperature sensors, humidity sensors, flow sensors, pressure sensors and so on. The performance of their work affect seriously the stability of the air-conditioning system. Once the sensor fails, it will not only lead to the reduction of indoor comfort and the decline of indoor air quality, but also cause the shortening of the life of the system and the increasing in energy consumption, as well as not achieve advanced control strategies.With the combination of HHT(Hilbert-Huang Transform) and WNN(Wavelet Neural Network), this paper proposes the fault diagnosis method based on HHT-WNN fault diagnosis, which can not only achieve a single sensor fault diagnosis and recovery, but also realize the trouble diagnoses and recovery of all the sensors (sensor systems) in the entire air conditioning system, including the four fault such as the complete failure, deviation failure, drift failure and the decreased precision failure.The simulation, in which four different fault for temperature sensor is conducted, verifies that this method detects accurately whether the occurrence of failure in the sensor or not, can also identify the type of fault and then recovery the failure.