Prediction of performance degradation in aircraft engines with fuel flow parameter

• Published in: Neural computing & applications 2024-02, Vol.36 (6), p.2973-2982
• Main Author: Kurt, Bulent
• Format: Article
• Language: English
• Subjects: Accuracy; Aging aircraft; Aircraft; Aircraft engines; Aircraft maintenance; Aircraft performance; Airplane engines; Artificial Intelligence; Artificial neural networks; Computational Biology/Bioinformatics; Computational Science and Engineering; Computer Science; Confidence intervals; Data Mining and Knowledge Discovery; Fuel flow; Image Processing and Computer Vision; Mathematical models; Neural networks; Original Article; Parameter estimation; Performance degradation; Performance evaluation; Probability and Statistics in Computer Science; Statistical analysis
• ISSN: 0941-0643; 1433-3058
• DOI: 10.1007/s00521-023-09174-9

Planned maintenance is required by licensed maintenance organizations to detect and prevent performance degradation in aircraft engines. In the literature, engine performance is evaluated with parameters that show engine performance. Fuel flow parameter is one of the important parameters that shows engine performance. In the models developed earlier, no engine performance evaluation was made with the fuel flow parameter at all stages from the take-off to the landing of the aircraft. In this study, fuel flow parameter is estimated with over 99.9% accuracy by using artificial neural network in MATLAB ® software. In order to detect the engine performance deterioration of the aircraft, the fuel flow values obtained from the artificial neural network and confidence intervals with 99% confidence level were established. Each value taken from the fuel flow sensor is evaluated by the model in all flight phases. In the model, engine performance is considered normal if the fuel flow value is within the confidence interval, and abnormal (anomaly) if it is outside the confidence interval. An accuracy of over 99.9% was achieved and results of this study showed that fuel flow rate of the engine of interest was within the confidence interval (no performance deterioration).