The Hidden Simulation Arms Race

e-motec
April 25, 2022

The Hidden Simulation Arms Race

eMobility will only become a reality with advanced simulation tools.

Mike Dempsey

As the world grapples with the challenge of averting permanent climate change, eMobility is a topic that has risen from the fringes of academic research to mainstream technological focus. One could be forgiven for assuming eMobility focuses solely on the electric vehicle; the reality is far more complex, challenging, and exciting. It promises to not only revolutionise our vehicle powertrains, but also to revolutionise the very core of human mobility. Fittingly, simulation will be at the heart of it.

Electrification can be broken down into 3 topics:

Propulsion, covering mechanical vehicle power.

Autonomy, encompassing mechanical vehicle control.

and Connectivity, the subject of information flow. No one aspect exists in isolation, and Claytex is committed to using its experience in the motorsport sector to develop tools and methods to address each part of the electrification problem.

A modular approach to simulation models enables the concepts of model reuse and scalable detail to be fully realised. In this Fuel Cell Electric vehicle example, the fuel cell, electric motor and battery models are highlighted; like all components of this vehicle, they can be swapped in and out, meaning components can be reused much like in real life, reducing modelling workload.
Exploded Diagram:
A modular approach to simulation models enables the concepts of model reuse and scalable detail to be fully realised. In this Fuel Cell Electric vehicle example, the fuel cell, electric motor and battery models are highlighted; like all components of this vehicle, they can be swapped in and out, meaning components can be reused much like in real life, reducing modelling workload.

Much has been made about the advancement of the fully electric vehicle (EV) in the past decade, maturing into a legitimate consumer option. Fresh investment is required by OEMs as they transition from mature technologies to comparatively immature ones. New expectations of the customer in terms of performance and sustainability add further pressure.

Hybridising the internal combustion engine (ICE) powertrain has been the first step towards eMobility. Primarily, this is a challenge of system control. How should energy be harvested? What is optimal deployment? How efficiently can the ICE be run? With so many variables available to work with engineers in F1, when faced with this precise problem, immediately turned to modelling the vehicle as a combined electro-mechanical system. Application of advanced control logic in software-in-the-loop (SiL) and hardware-in-the-loop (HiL) enables engineers to exercise control systems in representative environments from the start of development, including key control actuator response non-linearities. Evolution happens rapidly, with virtual development enabling engineers to take a cross-disciplinary approach, applying advanced techniques like optimisation algorithms and AI to the control problem.

To correctly capture dynamic behaviour of a vehicle, all elements much be considered. Accurate multibody models enables each component's contribution to dynamic performance to be understood and quantified.
Slalom: To correctly capture dynamic behaviour of a vehicle, all elements much be considered. Accurate multibody models enables each component’s contribution to dynamic performance to be understood and quantified.

Whilst sunsetting from commercial sale, hybrid technologies are migrating to EVs. Smart management of energy resources can provide the same benefits of increased range and improved performance in an EV as a hybrid, something vital for consumer confidence. Technologies such as kinetic energy recovery, will be carried forwards. Simulation plays a key role in understanding the impact and efficiency for such systems in various driving conditions, circumventing expensive and time-consuming physical testing. Variable testing costs, dependent upon climatic conditions, can be reduced and controlled with the singular fixed cost investment of simulation.

Thermal management represents one area where the EV departs from the ICE hybrid. Without ICE heat rejection, electrical drive energy must be deployed to generate heat, either pre-warming the battery to the optimal temperature or keeping the occupants comfortable. Studying the efficiency of various cabin configurations, material choices and heating strategies therefore directly impact the vehicle range. Similarly, so can the requirements to cool the cabin in hot climates. Both situations highlight shortcomings of traditional development processes, requiring travel of personnel and material to areas of specific climatic interest. Inherently time consuming and expensive, the virtual laboratory of simulation removes logistical challenges, not to mention it being cheaper than using climatic test chambers. An estimated investment of £25,000 for a single conceptual study must be weighed against £30,000 (and £1500 in training) for simulation tools, available for an unlimited number of conceptual studies.

Beyond the powertrain of the vehicle, EVs represent a wholesale revolution in the vehicle dynamics. Packaging battery packs and associated control electronics is a significant challenge, primarily because battery packs tend to be solid state rather than a liquid fuel as in ICEs, imposing more design constraints. Higher material density has driven OEMs to rethink traditional vehicle architecture, distributing battery cells in the vehicle. A trade-off between packaging and safety must be made; are more crash structures required? Fundamentally, the unibody chassis concept has evolved to accept this, impacting driving performance and road safety. Simulation tool sets, comprising offline detailed vehicle chassis simulations and driver-in-the-loop (DiL), have been deployed by OEMs in response. Simulation enables designs to be tested quicker; ethical concerns about test driver safety in potentially unstable prototype vehicles evaporate. Earlier design decisions increase productivity directly, increasing throughput of development ideas.

Advanced Driver Assistance Systems (ADAS), originally conceived to improve vehicle dynamic capability, have become indispensable to the EV OEM of today. Today’s vehicle dynamitists must grapple with consumer demand for larger, more ungainly vehicles whilst enabling the average human driver to remain in safe control in all weather conditions and road types across the world. Unavoidable increases in vehicle mass from additional safety features and batteries compound the challenge. Advanced simulation tools model the interaction between the human input and electronic command for such ADAS systems, without which, many EVs would not be dynamically safe. Investment and validation of a predictive simulation tool can be upscaled, returning investment across multiple vehicle platforms, from a fixed base cost.

Bus: eMobility will impact not only the personal vehicle, but also the commercial. Inbuilt scalability of simulation solutions enables them to be easily deployed to this challenge as well.
Bus: eMobility will impact not only the personal vehicle, but also the commercial. Inbuilt scalability of simulation solutions enables them to be easily deployed to this challenge as well.

Beyond this, many energy recovery systems impact the vehicle dynamics of the vehicle by their very action. Even in motorsport, this requires development to improve drivability, so the best drivers in the world can manage the interaction. For the average driver on the public highway, ease of use of a vehicle equipped with advanced energy recovery systems is paramount, due to the gamut of operating conditions. Again, DiL simulation tools enable this to be studied without risking human life.

In 2022, we are at an interesting inflection point regarding the human driver in the modern road going vehicle. Many ADAS systems now go beyond helping the driver to control the vehicle, to assisting their informational processing, reaching at the very core of the concept of human controlled highway driving. The vehicle is beginning to “think” for itself.

Autonomous vehicles (AVs) present only a marginal improvement in immediate energy efficiency; one can expect a computer algorithm to drive more efficiently. They do however open the door to a bigger revolution in mobility. Mobility as a service, with fleets of vehicles operating independently, is coming. As societal urbanisation increases, the very concept of vehicle ownership is set to be challenged. Ride hailing services like Uber or Lyft represent the conceptual genesis. Eventually, it will be more profitable, with better customer service, for a ride hailing service to utilise AVs. Avoiding relying upon the human element dictating the availability of service will improve customer experience.

Cornering: EVs present new challenges to the vehicle dynamitist. Multibody simulation enables confidence to be engineered into new design choices.
Cornering: EVs present new challenges to the vehicle dynamitist. Multibody simulation enables confidence to be engineered into new design choices.

Understandably, tool sets will be required for EV manufacturers to develop the AV; control architecture will be non-physical, so virtual development using simulation tools is logical. SiL using simulated plant hardware models can exercise the control system virtually; control actuation non-linearities can be included in the development process through HiL. Eventually, the same actual code gets redeployed on the final physical prototype. Sophisticated sensor models, replicating the fallacies of the physical sensor, are required. Such models interpret the virtual environment, accounting for physical limitations, such as dirty lenses, sun glare, rain, reflections, and other phenomena which will degrade the signal the control system receives. Additionally, simulation means autonomous AI can be trained on the actual target vehicle, regardless of if it exists or not, circumventing further costly recalibration.

Sensors: Sensor models built to interpret the virtual environment enables Autonomous Vehicle control systems to be trained virtually with all-weather challenges they face in the physical world.
Sensors: Sensor models built to interpret the virtual environment enables Autonomous Vehicle control systems to be trained virtually with all-weather challenges they face in the physical world.

Simulation in the development of AVs becomes unavoidable when the concept of certification and regulation occurs. 5 billion real world miles are estimated for AVs to drive to encounter all scenarios necessary to calibrate and validate its virtual brain. Business desire is to be first to market; driving these miles virtually is quicker than doing so physically. Considering sizable testing would be repeated in each sales market for regulatory body safety certification, it is abundantly clear that simulation tools are indispensable in enabling AVs to come to market in a timely fashion.

A more obvious and unpredictable hazard AVs will encounter, pedestrian and fellow road users, highlights a core ethical issue in AV development. Testing such a system on the public road without strong confidence in safety, presents acute danger and risk to pedestrians and road traffic. Legislative acceptance of a growing number of injuries and fatalities as the price to pay for technological advancement cannot be guaranteed. Traffic and pedestrian models integrated into the virtual environments such vehicles are tested in settles this ethical debate in the safest way possible.

Pedestrians: Ethical issues surrounding training Autonomous Vehicles around other road users can be circumvented by deploying advanced traffic and pedestrian models in virtual environments.
Pedestrians: Ethical issues surrounding training Autonomous Vehicles around other road users can be circumvented by deploying advanced traffic and pedestrian models in virtual environments.

eMobility will only become a reality with advanced simulation tools. Humanity’s timescale to avert permanent climate change demands rapid technological change. Crucially with any disruptive technological revolution, the business case must make sense for long term viability. Advanced simulation tool’s inherent speed and flexibility makes them fundamental in winning as large a stake in the eMobility market as possible. Fixed cost simulation tools reduce financial risks, reducing the potential for commissioning various unexpected tests. As tools evolve, so do the people that use them. Investment into training to use simulations tools can truly unlock their potential to transform eMobility. Put simply, they will enable OEMs to come to market before their rivals. It is in this sense, that advanced virtual systems simulations tools represent the hidden arms race of the eMobility revolution.

Mike Dempsey MD Claytex

:

2022 Spring Issue

Read & Download

e-mobility magazine spring

Engage with the Pioneers & Innovators who are making the shift to electrified transportation a reality.

e-motec

Check out my other articles

Is Giga Always Better?

Is Giga Always Better?

Is Giga always better? ~ The evolving landscape of EV manufacturing ~ Roger Brereton The competition to electrify vehicles is not simply the competition between rival manufacturers. What is now at...