Development and Validation of a MATLAB Tool for Replicating the Knock Index Calculation of a Marelli Motorsport ECU

Knock is one of the most damaging and anomalous forms of combustion in spark-ignition internal combustion engines, with the potential to cause significant damage to critical components, such as the piston head, while also limiting overall engine performance. Various methods have been developed to provide feedback signals for knock detection, with the most common approach being the monitoring of engine vibrations via accelerometers. Although traditional techniques perform adequately for standard commercial engines, they are often insufficient for high-performance engines, which operate at much higher rotational speeds. At elevated RPMs, these engines generate substantial vibrations, making it challenging to distinguish knock events from background noise in the accelerometer output. In such contexts, accurate knock detection is essential, not only for performance optimization but also for ensuring the longevity and reliability of engine components. The present project, developed during my internship at the Testing Department of Autotecnica Motori, aims to create a MATLAB-based software tool capable of replicating the knock index calculation performed by a Marelli Electronic Control Unit (ECU) in motorsport applications, starting from the high-frequency accelerometer signal acquired during test bench activities. Advanced signal processing techniques are implemented to isolate knock-related frequency components and to identify knock events with high accuracy across a wide range of operating conditions. Once validated, the developed code is also used to optimize the selection of ECU parameter values, enabling a better correlation between the knock index calculated by the ECU and the MAPO (Maximum Amplitude of Pressure Oscillations) index, obtained from in-cylinder pressure measurements using measuring spark plugs. The proposed off-line computational tool significantly reduces the time required on the test bench to determine optimal knock detection parameters, while minimizing human error and associated costs. Beyond its immediate application, the methodology provides a structured and repeatable approach to knock calibration in high-performance engines, offering potential adaptability to different engine architectures and control strategies. The results demonstrate that the proposed tool effectively replicates the knock index computed by the Marelli ECU, representing a reliable and efficient support instrument for advanced engine calibration in motorsport contexts.