Search Results

Knock control is one of the most vital functions for safe and fuel-efficient operation of gasoline engines. However, all knock control strategies rely on accurate knock detection to operate the engine close to the optimal set point. Knock detection is usually calibrated on the engine test bench, requiring the engine to run with knocking combustion in a time-consuming multi-stage campaign. Model-based calibration significantly reduces calibration loops on the test bench. However, this method requires a large effort in building and validating the model, which is often limited by the lack of function documentation, available measurements or hardware representation. As the software models are often not available, function structures vary between manufacturers and sub model functions are often documented as black boxes. Hence, using the model-based approach is not always possible.

The calibration of Engine Control Units (ECUs) for road vehicles is challenged by stringent legal and environmental regulations, coupled with short development cycles. The growing number of vehicle variants, although sharing similar engines and control algorithms, requires different calibrations. Additionally, modern engines feature increasingly number of adjustment variables, along with complex parallel and nested conditions within the software, demanding a significant amount of measurement data during development. The current state-of-the-art (White Box) model-based ECU calibration proves effective but involves considerable effort for model construction and validation. This is often hindered by limited function documentation, available measurements, and hardware representation capabilities. This article introduces a model-based calibration approach using Neural Networks (Black Box) for two distinct ECU functional structures with minimal software documentation.