Damage to the Air Handling Unit (AHU) system is often difficult to detect early, especially by less experienced technicians. This study aims to design and evaluate an artificial intelligence-based expert system to assist in diagnosing faults in AHU. The method used is a rule-based expert system with forward chaining inference, allowing the system to conclude based on a combination of user-selected symptoms. The development process began with knowledge acquisition through consultations with experienced AHU technicians, which was compiled into a knowledge base consisting of 9 common symptoms, 10 diagnostic rules, and 10 corrective recommendations. The system was implemented as an Android-based mobile application to facilitate direct field use by technicians. Testing with ten real-world cases demonstrated that the system produced accurate diagnoses consistent with expert technician analysis. With an accuracy rate of 100%, the system has proven effective as a decision-support tool in AHU maintenance and repair processes. Furthermore, it holds potential for future development through sensor integration and broader symptom coverage.
Keywords: expert system, forward chaining, AHU, fault diagnosis, artificial intelligence.