Implementation Of Forward Chaining In Intelligent Systems For Automatic Car Engine Fault Detection

Abstract

Intelligent systems are one application of artificial intelligence technology that can assist in automatic decision-making based on a knowledge base. In this study, a smart system based on the Forward Chaining method was designed and implemented to detect damage to petrol-powered car engines. This system employs the Forward Chaining method, which systematically matches the user-inputted symptoms with the rules in the knowledge base to determine the most likely cause of the damage. The knowledge base was developed through in-depth interviews with three expert technicians from official repair shops and validated by a head supervisor, resulting in 28 specific diagnostic rules. Diagnostic data, including symptoms and corresponding damage, was collected from service records of 50 cases involving Toyota Avanza and Honda cars manufactured between 2019 and 2022. This system is capable of detecting various symptoms, such as sudden engine failure, difficulty starting, excessive vibration, abnormal exhaust smoke colour, and unusual engine noise. The system development process used a waterfall approach that included needs analysis, knowledge base design, interface creation, and system testing. The test results showed that the system was able to detect 94% of damage types with high accuracy based on the combination of symptoms provided by the user. Thus, this system can be an effective tool in engine maintenance activities and improve production process efficiency in industrial environments.