Detection of malfunction in ignition system for an internal combustion engine via artificial intelligence model

Engine failure is a significant issue for drivers, often requiring substantial experience to identify and troubleshoot effectively. Repairing the engine based on probable causes and uncertainties can be time-consuming and costly. Recently, AI models, particu+
larly those based on Artificial Neural Networks (ANN), have been developed and gained popularity in fault diagnosis. This paper considers two common faults in internal combustion engines - cylinder misfire and complete cylinder failure - caused by ignition system issues. An Artificial Neural Network fed by Statistical features (SANN) is employed to distinguish these two faults. The SANN was trained on statistical features derived from vibration signals and achieved an accuracy of over 90%. Thus, SANN can classify the fault generated by the ignition system. This model was further validated using a different engine as a second case, demonstrating its ability to predict fault types with acceptable accuracy. In fact, the SANN could find a malfunction of the engine mounted on a car perfectly. This capability enables operators to accurately identify the type of fault, allowing for more precise and efficient repairs. Therefore, the proposed method is well-suited for troubleshooting ignition system malfunctions and diagnosing related issues via a reliable fault detection model