Method And Application Research On Online Fault Detection And Diagnosis For Variable-air-volume Air-conditioning Systems

Variable air volume (VAV) air-conditioning system is a popular type of heating, ventilation and air-conditioning (HVAC) system in high-grade office buildings for energy saving, good heat comfort, flexible control and most adapting to space with variable load conditions. However, VAV air-conditioning systems tend to have more faults due to the complexity of VAV air-conditioning systems and higher requirements of their control systems. Faults will result in the increase of energy consumption, the deterioration of heat comfort, the damage of components and the increase of maintain costs. Energy management and control systems (EMCS) are widely employed in modern buildings. The huge amount of data available on EMCS systems provides rich information for online fault diagnosis of HVAC systems. Online fault detection and diagnosis for HVAC systems has already received increasing attention recently.Therefore, research on online fault detection and diagnosis (FDD) of VAV air-conditioning systems is of great importance in economy and engineering. Online fault detection and diagnosis methods for VAV air-conditioning systems are deeply investigated in this study from the perspective of main components.The results of model-based FDD are strongly dependent on the accuracy of HVAC component models. The accuracy of HVAC models can be improved if model parameters are tuned by using the collected operating data of HVAC systems. To this end, component models with self-tuning parameters for air handling units are presented in this study. Model parameters are tuned by using a genetic algorithm (GA) which minimizes the error between measured and estimated performance data. The propsed models were validated against real data gathered from existing HVAC systems. The validation results show that component models with self-tuning parameters have higher prediction precisions and can be used to detect faults in air handling units. An online adaptive scheme is also developed to update the fault detection thresholds, which vary with system operating conditions. Fault detection thresholds are determined by using a statistical method. Adaptive fault detection thresholds can improve the accuracy of fault detection and facilitate the practical implementations of model-based FDD methods.As for fault detection and diagnosis of VAV air handling units, each single method is flawed and ineffective in real implementations. An online fault detection and diagnosis strategy for VAV air handling units is presented based on component models with self-tuning parameters and expert rules. Component models with self-tuning parameters are used to detect faults in air handling units. Three rule-based fault classifiers are developed to find fault sources. The proposed fault detection and diagnosis strategy was online validated and tested on real VAV air-conditioning systems.Proper design parameters of CUSUM control chart are very essential to design CUSUM control charts. Improper parameter values will deteriorate the performances of CUSUM control charts in practice. To this end, this study introduces the methods by which the essential design parameters of CUSUM control chart and the required parameters of the control process can be identified. Mathematical analysis method is used to analyze the effect of serial correlation on the performance of control charts using data from the first order autoregressive process (AR(1)) and the first order moving average process(MA(1)). Based on the above study, residual-based CUSUM control charts are used to detect faults in VAV terminals. The residual-based CUSUM control chart can improve the accuracy of fault detection through eliminating the effects of serial correlation on the performance of control charts. Also, the residual-based CUSUM control chart can enhance the robustness and reliability of fault detection through reducing the impacts of normal transient changes.An online fault detection and diagnosis strategy for VAV terminals is presented based on residual-based CUSUM control charts and expert rules. Residual-based CUSUM control charts are used to detect faults in VAV terminals. A rule-based fault classifier consisting of expert rules and fault isolation algorithms is developed to find fault sources. The proposed fault detection and diagnosis strategy for VAV terminals was online validated and tested on real VAV air-conditioning systems.Based on theoretical study, an online fault detection and diagnosis software for VAV air-conditioning systems is developed, which consists of an online fault detection and diagnosis software for VAV air handling units and an online fault detection and diagnosis software for VAV terminals. The new-developed FDD softwares were online implemented and validated on real VAV air-conditioning systems. The validation results show that the new-developed FDD softwares can depict operational condition of HVAC equipments and find fault sources precisely.