Performance of the Bond Graph Approach for the Detection and Localization of Faults of a Refrigerator Compartment Containing an Ice Quantity

Abstract

In this paper, a robust fault diagnosis for a refrigerator compartment containing a quantity of ice using the bond graph (BG) approach is performed by linear fractional transformations (LFTs). The BG model describes heat transfers supported by the amount of ice placed in the refrigerator compartment, as well as a water container. The LFT modeling of BG elements offers advantages from the point of view of structural analysis and data processing implementation. We have introduced four faults, which consist of ice temperature rise, water leakage, insulation failure at the hot walls of the refrigerator and an increase of the internal temperature due to poor door sealing. The faults are in the form of additional heat fluxes. The simulation results show the effectiveness of the proposed method for detecting and localizing faults. In addition, the lack of door sealing has the most influence on the temperatures in the internal cooling space, water, and ice compared to the other faults.