The end of 2017 is about to see the introduction of new RDE regulations for passenger cars that also prescribe emission limits in real-world road traffic over a broad vehicle operating range and demand they be reliably met. This means automobile manufacturers must cover a much wider spectrum of driving situations in engine calibration. All told, this will significantly boost the work involved in testing and validation as many more parameters will have to be optimized alongside each other than they are today.
“Compared with today’s development process, the quantity of data is exploding, making it possible to speak of big data”, says Dr. Mirko Knaak, Senior Consultant at IAV. To structure data management, IAV has developed a new system that brings together all steps of the calibration process in an integrated chain of methods and tools. This covers the entire process of recording data, simulating and testing to evaluating the results and managing knowledge in the database. “This gives us a platform we can offer to our clients for an all-embracing concept that quickly and efficiently produces robust results”, says Tobias Niewolik, development engineer at IAV, summing up the concept’s benefits.
Automated data processing and evaluation
IAV’s Mara tool forms the heart of data processing and evaluation. The software manages measurement files and analysis configurations. New measurements are automatically processed, with the results being sent to the sections or teams involved. This distinguishes the new approach from conventional calibration processes.
Although coventional processes produce a myriad of measurement values, they are
not brought together at a central point and automatically made available to other calibration engineers involved. This results in a situation whereby additional test series or simulations are carried out in spite of results already being available.
“Our new approach lets us make better use of the potential provided by available data, create synergies and drastically reduce the level of development work entailed”, Dr. Mirko Knaak explains. For instance, if an engineer sees the vehicle’s emissions getting out of hand in a specific driving situation, such as in response to a hefty step on the gas pedal after a trailing-throttle phase, he or she can look for comparable results from all measurements in the calibration project.
At the press of a button, IAV Mara automatically evaluates these specific events or even entire cycles from the database. This is done using individually tailored templates that can be generated without any programming knowledge. The program also uses steady-state models capable of providing a significant edge in examining OBD functions or the spread of results from RDE tests based on driving style. Open interfaces also ensure compatibility with established development tools.
IAV has developed a machine-learning tool to analyze problems of a more complex nature. Whereas IAV Mara requires rules written by experts, the algorithms in the machine-learning tool automatically identify relationships in data and, from this, derive patterns or parameter combinations that cause emissions to vary when problematical events occur.
A further application is the generation of worstcase scenarios for validating RDE compliance. To do this, the tool automatically checks existing datasets for situations indicating comparable behavior. The tool can then take these specific maneuvers as the basis for generating numerous stochastic RDE cycles. Remote access to prototype vehicles Prototype vehicles are often only available on
a limited scale during the calibration phase. IAV’s new methodology permits a highly efficient use of available prototype capacities. While a vehicle is on a test drive, development engineers can access the on-board measuring equipment from their desk by LTE cellular communication.
To do this, the IAV interface on the office computer logs into the vehicle system. Tobias Niewolik: “The developer can remotely change specific calibration parameters and see how they affect measurement values online. This saves time, money and makes better use of prototype vehicle capacity.”
IAV is successfully using the new methodology in engineering projects. “The next logical step in optimizing the development process will be to integrate the Internet of Things where more and more vehicle components are becoming intelligent, interconnected and communicating with other systems. This will generate additional data that can be used for minimizing consumption and emissions”, says Dr. Mirko Knaak, opening a glimpse at tomorrow’s development methodology.