3ENERGY ENGINEERING ISSUE 69

CLAYTEX

2 ISSUE 69 ENERGY ENGINEERING

CLAYTEX

Getting physicalStimulating vehicle sensors is only part of the equation when it comes to the virtual development and testing of today’s vehicles

One of the main topics of discussion throughout 2016 has been autonomous

vehicle development. Claytex have introduced the latest developments in rFpro and Dymola which now enable the virtual testing of the complete vehicle including the autonomous control systems. As distributors and partners to both products Claytex is uniquely placed to help develop the complete immersive virtual test environment for both vehicle and driver.

In early September 2016, rFpro announced the availability of a new product called the rFpro Sensor IG. This new extension to the rFpro portfolio enables data feeds to be provided to sensor models in real-time, accessing the same high

fidelity road and track data used to stimulate the vehicle physics model and human driver.

rFpro was developed as a Driver-in-the-Loop simulator environment, initially in Motorsport, and provides high fidelity graphics and audio built using LiDAR scanned road data with sub millimetre precision. It’s highly extendible, plugin architecture enables models from many different simulation tools to be connected into the environment to define the vehicle behaviour. Together with the latest generation motion platforms, rFpro has progressed from Formula 1 into other motorsport series and into Automotive OEM’s and Tier 1 suppliers to support the virtual development of vehicle dynamics.

The Sensor IG product allows

the same high fidelity data to be manipulated and presented to the vehicle sensor models in real-time. For camera sensors, we can apply lens distortion effects and simulate dirt on the lens. For range based sensors such as LiDAR, Radar and Ultrasound the road data is interpreted into distance information with the same possibility to impose lens distortion or dirt effects at the front of the sensor. The data is then transferred to the clients own sensor models to correctly process the data just as the real device would do in the vehicle.

Other features found in rFpro allow us to introduce traffic and pedestrians into the virtual environment and to connect multiple systems together to see how they will interact.

Stimulating these vehicle sensors is only part of the equation when it comes to the virtual development and testing of today’s vehicles. We also need to make sure that the physical behaviour of the vehicle is accurately represented and that covers many different physical domains as well as all their associated control systems.

Dymola is a multi-domain modelling and simulation tool that enables the physical behaviour of the complete vehicle system to be predicted. It enables the modelling of engines, gearboxes and drivelines, electric motors, battery packs, cooling systems, vehicle dynamics and every other aspect of the vehicle in one simulation environment. It uses the open standardised modelling language, Modelica, to define the physical behaviour and supports the FMI open standard to allow models

and control systems developed in other tools to be easily integrated.

Again, originating in Motorsport, the real-time vehicle dynamics solutions for Dymola enable detailed MultiBody suspension models to be used. These have been applied in Formula 1, NASCAR and IndyCar and for many of these customers the focus was initially vehicle dynamics but the multi-domain capabilities of Dymola and Modelica have allowed them to extend this to cover every aspect of the car. These same capabilities are also being utilised by many Automotive OEM’s and their suppliers to support the development and testing of new vehicles.

One of our collaborative research projects that has been developing this capability for powertrain control systems was the MORSE project. This was a 2 year Innovate UK funded project where we worked with Ford and AVL Powertrain UK to develop

engine and powertrain models to support the virtual calibration of driveability and OBD validation using SiL and HiL environments. This project has led to the development of new combustion models and real-time capable, crank angle resolved engine models as well as many other enhancements to the powertrain models to include thermal effects and support fault injection.

The final pieces in the puzzle are support for the latest generation motion platforms as well as the possibility to integrate HiL systems into the virtual environment. This enables testing with human drivers and passengers immersed in the same environment as the vehicle, including all its sensors and controllers. The controllers could be integrated as SiL or the real controller could be connected via a HiL system.

www.claytex.com

Image: Hybrid Vehicle Model Diagram