Advanced simulation for calibration for rail pressure governor

Abstract

The CRS (Common Rail System) is an accumulator fuel-injection system where the pressure generation and the injection are decoupled in comparison to other injection systems. Nowadays CRS calibration process is performed manually: a good knowledge of calibrating the vehicle is needed and numerous steps needed to be repeated when an error is made. In addition, it’s not possible to define a process that leads to an optimum calibration. Hence, the main goal of this project is to develop a complete method to reach the best possible calibration by saving costs and time for the Rail Pressure Governor. In order to develop this system it is necessary to conduct a feasibility study (to be sure about the measurements reproducibility) and to get a deep knowledge about the governor theory first. The next step applies the Design of Experiment (DoE), a method for data-based modeling of unknown systems, with the goal to evaluate and optimize the system’s behavior. After that, measurements with the vehicle using the Engine Control Unit (ECU) and INCA 6.2 will be performed. Then, it is necessary to implement software to develop the criteria for an optimization in Matlab. The results of the criteria calculation will be the feed of ASCMO (Advanced Simulation for Calibration, Modeling and Optimization) – a tool for modeling the input/output behavior of unknown systems based on measuring data. With the model generated with ASCMO it is possible to determine the best parameters for the controller and, finally, find a good calibration for the vehicle. The system will consist of an ECU model EDC17 which will be used in a Citroen C4 - platform demonstrator, connected by a Controller Area Network and an Emulator Test Probe. A code needs to be developed in order to calculate the criteria and create a proper strategy. A complete test program and a calibration have to be done in another similar vehicle after the calibrating parameters in order to validate the results and conclude about the benefit of this new method. The expected results should save time through automation and DoE, as well as prevent reworks and reduce costs. At last, it is important to emphasize the fact that the work in optimization and measurement can be divided, so the measurement can be done by someone without a deep knowledge in calibration and just the optimization will need qualified knowledge.