This report is the documentation for the work done by two master students at the IRS (Intelligent Reliable Systems) line at Aalborg University Esbjerg. The project is developed based on a project-proposal from Danfoss A/S, which also supplied a base-model of a refrigeration system. The focus of the report is the development and testing of different methods for detecting and isolating faults in a supermarket display case. The Kalman Filter, Extended Kalman Filter and Unknown Input Observer have all been tested and compared to each other. A simple online parameter estimation is also tested with good initial results, but due to time constraints it has not been completed. Based on the comparisons, a suggestion for a fault detection and isolation system is given, which will be able to both detect and isolate the most often occurring faults in a display case.