Skip to Main content Skip to Navigation

Développement d'une méthode connexionniste pour la détection et le diagnostic de défauts de systèmes de chauffage

Abstract : Nowadays, the Heating, Ventilating and Air Conditioning (HVAC) systems are becoming more and more complicated due to the increase of their performances. Consequently it is more difficult for the maintenance teams to understand the running states of these systems and to detect and diagnose their faults. They wish to have performant tools which can help to detect and diagnose (if necessary) the operating faults of HVAC systems. This thesis concerns the development of such a tool for Fault Detection and Diagnosis (FDD) suitable for hydraulic heating systems. The most important faults of hydraulic heating systems, for which a FDD tool must be developed, have been selected at the first stage of the thesis. And then, the modelling-simulation of 5 heating systems with and without faults allowed us to obtain a data base which will be used to develop a FDD tool. Based on pattern recognition theory, a FDD prototype has been developed by using a connectionist model (Multi-Layers neural networks) as classifier. The prototype has been tested for 7 cases in each of the 5 simulated systems. The overall results are satisfactory (a successful detection rate greater than 90% and a false alarm probability less than 2%) although only the simulation data of one heating system were used to train the FDD prototype. This study shows that the generalization of the FDD prototype to real heating systems could give interesting results. So, the main prospect, resulting from this thesis, consists of: - validating this prototype with real heating systems, - implementing it in Building Energy Management systems in cooperation with industrialists, - applying the methodology developed in this thesis to other HVAC systems.
Complete list of metadata

Cited literature [39 references]  Display  Hide  Download
Contributor : Ecole Des Ponts Paristech Connect in order to contact the contributor
Submitted on : Monday, October 25, 2010 - 4:27:40 PM
Last modification on : Friday, October 22, 2021 - 4:40:15 AM
Long-term archiving on: : Wednesday, January 26, 2011 - 3:07:31 AM


Files produced by the author(s)


  • HAL Id : tel-00529470, version 1



Xiaoming Li. Développement d'une méthode connexionniste pour la détection et le diagnostic de défauts de systèmes de chauffage. Interface homme-machine [cs.HC]. Ecole Nationale des Ponts et Chaussées, 1996. Français. ⟨tel-00529470⟩



Record views


Files downloads