A novel sensors fault detection and self-correction method for HVAC systems using decentralized swarm intelligence algorithm

• Published in: International journal of refrigeration Vol. 106; pp. 54 - 65
• Main Authors: Wang, Shiqiang, Xing, Jianchun, Jiang, Ziyan, Dai, Yunchuang
• Format: Journal Article
• Language: English
• Published: Paris Elsevier Ltd 01.10.2019; Elsevier Science Ltd
• Subjects: Algorithms; Capteurs intelligents; Computer simulation; Convergence; CVC; Decentralized swarm intelligence optimization; Détection et auto-correction des défaillances; Error correction; Fault detection; Fault detection and self-correction; Fault diagnosis; HVAC; HVAC equipment; Maintenance costs; Nodes; Optimisation décentralisée d’intelligence en essaim; Optimization; Sensors; Smart sensors; Swarm intelligence; Fault detection and self-correction; CVC; HVAC; Decentralized swarm intelligence optimization; Optimisation décentralisée d’intelligence en essaim; Détection et auto-correction des défaillances; Smart sensors; Capteurs intelligents
• ISSN: 0140-7007; 1879-2081
• DOI: 10.1016/j.ijrefrig.2019.06.007

•To realize precise location and self-correction of a fault sensor in HVAC, an exponential function method is proposed.•A novel decentralized flat sensor diagnosis network for HVAC system is established.•A decentralized swarm intelligence optimization algorithm for sensors fault detection and self-correction is developed.•The convergence of the algorithm is proved theoretically and the feasibility is demonstrated by hardware test. Under a self-organizing and flat sensors network structure, a novel decentralized sensors fault detection and self-correction method is proposed. In the self-organizing sensors network, each sensor is updated as smart agent and is connected based on basic physical relations and can communicate with adjacent nodes. This network possess the advantages of plug-and-play and can reduce the high labor and maintenance cost, without having to build central monitor. For the proposed decentralized method, each smart sensor is treated as an individual which has its own utility with local coupling between adjacent nodes. And a decentralized swarm intelligence optimization algorithm is designed and executed in all the smart nodes in parallel to achieve the sensors fault detection and self-correction. Moreover, convergence property of the presented method is analyzed theoretically. Simulation results and hardware platform test on an actual HVAC system illustrate the effectiveness of the proposed method.