Research On Operation Parameters Fault Diagnosis Approachusing Decentralized HVAC Systems

Since abnormal operation parameter may cause problems like high energy consumption, occupant complaints and degraded system operation property, FDD(fault detection and diagnosis) for operation parameter is very important. Abundant researches have provided various FDD methods which have been proved to be effective by simulation. While few FDD systems have been applied in practice, there are two reasons classified into modeling and installing difficulty and complexity. FDD model for certain HVAC system is unique and not adaptive to other HVAC systems, which means FDD modeling for new HVAC systems is a time-consuming and labor-consuming work. And installing problem refers to the present centralized control systems leads demonstration of FDD systems to high work cost and labor cost. Aiming at improving the accessibility for FDD systems in real HVAC systems, this paper proposed a decentralized HVAC parameters FDD method, based on which researches in terms of modular parameter FDD modeling and decentralized optimization method are to be conducted. For modular parameter FDD modeling, HVAC systems are divided into typical blockings, based on which, selecting SCOM as the modeling method to establish a standard FDD model base. For decentralized parameter FDD, decentralization method for converting a centralized problem to a decentralized problem is proposed and a corresponding decentralized optimization method which based on genetic algorithm is developed.The main task and research content of this paper can be classified as follows:1)By literature review, analyzed the mathematical model for FDD problem, the nature of which is a constrained optimization problem. And select SCOM as the FDD modeling method.2)General introduction of modularized FFD is given by analyzing the significance and feasibility of modularized FDD modeling. HVAC system blockings division and their connection methods have been studied.3)A set of FDD model for 11 blockings in HVAC systems are to be established to construct a standard model base.4) Developing a decentralized optimization method. Based on decentralized network, original FDD model is to be divided into a series of subsystems model. A corresponding decentralized optimization algorithm is developed to solve the decentralized constrained optimization5)Verifying the effectiveness of decentralized optimization method by simulation on MATLAB platform6)Anaslysising the characteristics and applied limits for decentralized data FDD.In the end a conclusion and evaluation for the proposed method is carried out.