Calibration Methods at FEV

The target of modern powertrain development is to meet the challenging and to some degree contradictory development goals within a short time frame. Examination of the time required for development, it becomes apparent that design loops have to be avoided. Recently, methods have been successfully applied for an efficient design process. Testing and vehicle application work can significantly benefit from methods empowered by model-based approaches. Nowadays, models with different levels of detail are able to significantly improve nearly every development phase. Standardized and automated test benches and vehicle procedures can support an efficient and comprehensive development process, which is also necessary to tackle growing complexity.

To support all of the major calibration steps in the development process (see Fig. 1), it is necessary to deliver methods that are capable of decoupling calibration tasks as far as possible from the hardware. It is also essential that they provide possible solutions as fast as possible with as little experimentation as possible.

Fig. 1 : Support of Different Phases of Calibration Projects with Appropriate Methods
The calibration methods themselves are divided into three major fields presented by the three columns in Fig. 2. The first column representing the Standards delivers procedures for simple calibration tasks covering easy technical descriptions and the definitions of several processes. These standards ensure quality and reproducibility issues as well as they provide a well structured knowledge base. To enforce this, these standards are centrally available within the FEV-Intranet and are mandatory for use in all projects, unless a customer specifically requires a different approach. In those situations, deviations from these FEV-standards will be documented to ensure the reproducibility of the results.

Fig. 2 : Main Columns of Calibration Methods
The second column represents more complex calibration tasks, which are covered most of the time by model-based approaches. The third column covers vehicle tools to support calibration engineers in their daily vehicle calibration work.

These tools are embedded into a standardized environment (Fig. 3) to provide a comfortable graphical user interface with several plausibility checks and calibration support tools.

Fig. 3 : TOPexpert Environment for Vehicle Tools
Models that are supported numerical solvers can be utilized to solve calibration-specific optimization tasks (Fig. 4). The optimization problem is defined with the help of a cost function with boundaries punishing every deviation from the desired target value with a weighted scalar value. Additionally, restrictions like hard constraints (NOx emissions within a cycle or fixed full load points) or restrictions to gradients (Ensuring drivability) can be introduced by means of boundary constraints. Using these cost functions as an example, it is possible to minimize fuel consumption while keeping the emission levels within the legislative boundaries. Additional criteria also guarantee the smoothness of the maps delivering ready-to-use maps for the whole operation area.

Fig. 4 : Base Points for Emission Map Optimization (Diesel)
As an example for vehicle tool The Thermal Diagnostics vehicle tool is provided as an example in Fig. 5. This diagnostic is used in the event there is an increase of engine out emissions if either the engine coolant thermostat is blocked open or the engine coolant water temperature sensor is malfunctioning. Both diagnostics consume a great deal of time and resources. The whole engine coolant and ambient temperature combination, which can occur during customer relevant driving conditions, must be mapped, analyzed and verified in a very complex manner. After every calibration step, the result must be verified and analyzed. These calibration loops can be reduced by using the TOPexpert vehicle tool for simulation.

 

Fig. 5 : First Calibration Step TOPexpert Tool for Thermal Diagnostics
To detect a blocked thermostat, an engine coolant water temperature model (2) needs to be mapped as close to the behavior of the real engine coolant temperature as possible (1). This can be achieved using the offline simulation capability of TOPexpert. The original model (3) is still shown in the figure, to illustrate the improvement due to the data change. Even previous taken data can be reviewed, and the result of the change can be evaluated with regard to selectivity and robustness.

The FEV TOP Expert Tool supports engineering by guiding them through the complete calibration process to guarantee high quality standards, which are necessary during engine and vehicle development. Starting with the description of the methods and the documentation of calibration procedures for all existing tools, the tool guides the engineer through the necessary steps. Values that are not plausible and missed steps will be shown directly on the screen and warn the engineer regarding the improbable results.

The use of modern calibration methods has an impact on the calibration process. Formerly, processes started with measurements followed by the evaluation of the data, including the calibration. To evaluate the effect of the changed calibration, multiple reiterations are necessary where the same boundary conditions need to be reproduced. If the result of this measurement does not meet the expected quality, the loop needs to be repeated until the expected result is achieved. By using calibration tools, the process can be further improved (Fig. 6). FEV developed a measurement program, which defines the types of topics to be measured in one measurement loop.

Fig. 6 : Time Schedule for Calibration Processes
After performing a defined number of measurements, the data will be read in the vehicle tool and will be evaluated. In the next step, an offline simulation is executed after changing the maps. The influence with regard to the relevant output signals can then be directly analyzed. The final robustness check will be done under real world conditions, such as on the test bench or during test trips.