XILforEV

The project proposes the development of a new experimental methodology for the design of complex systems for electric vehicles.

This methodology is based on the real-time connection of test benches and platforms located in different geographical locations that will constitute a unique X-in-the-loop (XIL) experimental environment. Digital twins that are updated in real time through the use of dynamic data-driven systems (DDDAS) concepts will be used in the execution of tests. This new test methodology will make it possible to explore interdependencies between different subsystems that could hardly be addressed until full vehicle testing.

Problem

Vehicle electrification is one of the key factors that will determine the trends, challenges and progress of the automotive industry for the coming decades. Now, new car sales growth for all-electric vehicles (EVs) is forecast to start at 2 Mio. in 2020 to 44 Mio. in 2030 (combined for the EU, USA and China). Electric vehicles should be cheaper than combustion models as early as 2025.

The largest increase in the role of electric vehicles is related to automated driving because an electric propulsion system fits well with the system architecture of autonomous vehicles by making their control functions more flexible and redundant.

These and other factors make the electric vehicle segment very attractive to the industry. As a result, now not only traditional automakers, but also many newborn rivals in the IT sector, are focusing heavily on the development of electric vehicles. However, recent observations show that sustainable production of electric vehicles, regardless of the manufacturing model, requires new design procedures. The overall development process of electric vehicles consists of many stages, elements and components, which are also characterized today by uneven levels of technological development. maturity.

After performing the analysis of current EV design technologies, the XILforEV consortium has identified the following specific question, which is not sufficiently addressed neither at the industrial level nor in research: how to efficiently perform integrated development and testing of EV systems from different domains?

The problem is that here not only a proper design of the electric powertrain is required, but also a revision of the car chassis design.
EV motion control requires a combined operation of powertrain and chassis actuators (e.g., brake combination) which motivates at least the following design challenges:

• Harmonization of powertrain and EV chassis performance dynamics;
• Deliver the required user acceptance of the new EV functionalities;
• Address more complex fault tolerance and robustness requirements.

Considering these factors, the use of well-established processes in EV system design may have some sensible limitations, e.g., co-simulation issues for software-in-the-loop (SIL)/model-in-the-loop (MIL) procedures, availability of hardware-in-the-loop (HIL) test configurations for different systems on the same host, tangible extension of road test programs with additional time/cost resources to test new functionality.

Results

To address this scope of problems, the consortium proposes a new approach aimed at developing a connected and shared X-in-the-loop (XIL) experimental environment that links test platforms and configurations from different physical domains and located in different locations. The domains under discussion may cover (but are not limited to) in-circuit hardware test equipment, dynamometers, software simulators, driving simulators, and other variants of experimental infrastructures.
The real-time execution of specific test scenarios simultaneously on (i) all platforms/devices connected with (ii) the same real-time models of objects and operating environments allow exploring the interdependencies between various physical processes that can hardly be identified or even expected at the design development stage. In a long-term perspective, the plug-in concept of including multiple test platforms/devices and easy on-demand access to test programs for developers, engineers and researchers will have a major impact on the electric vehicle design community by connecting experimental environments around the world.

All of the objectives listed below address the development, design and test procedures applied to electric vehicles and their systems:

• Develop the methodology for connected and shared XIL experiments and specify the architecture of the corresponding experimental environment;
• Design the hardware and software components required to perform shared XIL experiments;
• Introduction of a machine learning layer in the XIL subsystem models for automatic real-time (RT) model accuracy and confidence improvement based on test setup results;
• Development and validation of high-confidence models suitable for time-accelerated virtual simulation, which will merge the different technologies involved and allow seamless integration and scalability while maintaining compatibility with the functional mock-up interface (FMI) co-simulation approach;
• Conducting case studies, which will demonstrate the practical implementation of the XILforEV concept and the benefits in reference cases, incl. validation of the fail-safe functionality and robustness of the developed systems;
• Develop procedures for the inclusion of users in shared experiments taking into account the Open Access and Open Science frameworks.