As auto manufacturers shift towards entirely new types of vehicles, they face a number of engineering challenges. Paul Schreier looks at how design/simulation software is helping them reach their goals more quickly and efficiently
We all have a general idea of what the car of the future might look like: above all, it will likely rely more on electricity to cut emissions, and the innovative use of electronics for everything from the drivetrain to convenience factors will help various manufacturers differentiate themselves. These uses of electricity will make thermal management more important, and that’s just one area where modelling is making large contributions to the development of the car of the future.
Seven years of engine modelling
While much attention is being paid to electric/hybrid cars, potentially significant breakthroughs are being made in internal-combustion engines. For instance, the Scuderi Group has been developing a split-cycle engine that should reduce NOx emissions by up to 80 per cent and potentially increase efficiency by 10 to 15 per cent under full-load conditions over conventional engines; after adding turbocharging and an air-hybrid energy storage system, the company expects to see efficiency increases of 25 to 50 per cent.
Stephen Scuderi, VP and patent attorney for the firm, explains: ‘Although this technology has been known for many years, it’s only advances in simulation capabilities over the last decade that have allowed us to overcome the thermal and efficiency problems long associated with this type of engine. Today it’s impossible to design an engine without both simulation and empirical data; you can’t measure everything going on inside the engine, but simulation results must be validated in the real world. It’s taken us seven years of modelling work to reach the point where we are finally building a working prototype. Part of this is due to the fact that this engine concept is so new that we don’t have any baseline working engine against which we can compare the simulated results.’ The company is starting to license this technology and expects to see vehicles with a Scuderi engine on the market as early as 2012.
The Scuderi split-cycle engine takes the four strokes of the Otto cycle and divides them between two sets of paired cylinders: intake/compression and combustion/exhaust. In essence, it’s a compressor on one side and a two-stroke engine on the other; this scheme, in its air-hybrid form, allows the engine to recapture and store energy that is normally lost. Ignition is timed to occur between 11 and 15 degrees after top-dead centre under highly turbulent conditions − which leads to increased thermodynamic efficiency and very rapid atomisation of the fuel/air mixture for a more complete combustion process and a cleaner burn.
While the ideas and IP come from Scuderi, most of the design effort took place at Powertrain Technologies in Shoreham-by-Sea; that company acts as an exclusive subcontractor to Southwest Research Institute (SwRI), an independent consultant and test laboratory located in Texas. For its simulations, SwRI is working with tools that are standard in the auto industry. It turns the CFD analysis over to Convergent Science in Wisconsin; that firm develops and sells the Converge CFD software. The fundamental principle behind the uniqueness of Converge is that the user does not directly supply a grid as an input; instead, the user supplies a triangulated surface and a series of guidelines from which Converge creates the grid at runtime. For simulations with moving boundaries or changing embedding, the grid is recreated at each time step. While a 3D CFD analysis is used to study combustion in the cylinders, SwRI works with the GT Power package from Gamma Technologies in Illinois, to perform 1D cross-section thermodynamic analysis of the cylinders.
Heat management in batteries
Meanwhile, electric/hybrid vehicles are already on the road, and for them aerodynamic efficiency is very important due to the limited driving range of batteries, says Richard Johns, director for the automotive industry at CD-adapco. He adds that the thermal management of battery packs is also critical, especially since there is a move towards lithium-ion batteries, which can generate considerable heat.
Figure 2: This simulation from Ansys shows the current density for six IGBTs; the higher the value, the higher the temperature and thermal stresses.
Some manufacturers place batteries in the car’s centre away from the crash zone, but then it’s more difficult to remove their heat and can also affect passenger comfort. Thus this area is ripe for modelling. Some batteries are cooled with dedicated systems, but air or fluid flow also impacts structural integrity. So, while it has traditionally focused on flow, CD-adapco in recent years has been adding multiphysics analysis to its software. In that way it can now evaluate all aspects – flow, structural mechanics, heat transfer – at one time without any external passing of boundary conditions from one model to another.
CD-adapco is also working with manufacturers to reduce carbon emissions. This company has formed the ATOMIC project; based on guidance from participating companies, it focuses on using CFD methods to model the atomisation of sprays generated by fuel-injection systems. The project is being supported financially by companies including Daimler AG, Delphi Diesel Systems, Lotus, Honda, MAN Diesel, Porsche and Wärtsilä.
LEDs for lights
Heat is also a concern of the LED headlamps and tail lights that have started to become standard equipment, because they offer an operating lifetime as long as that of the car, use less power and weigh less. Headlamps will automatically adjust their output depending on the time of day and the distance of oncoming traffic, and they even shift beam location as you drive around corners. Brake lights no longer simply turn fully on; with LEDs, their response time decreases and their intensity can vary depending on the amount of braking.
And while LEDs are more efficient than traditional lamps, they still generate considerable heat, especially in environmentally-sealed enclosures, and this heat can dramatically reduce the LEDs’ operating life or change the light colour. Thermal issues also arise, because such headlamps hold complex arrays – the headlamps in the Audi R8 uses 54 LEDs – yet must fit into a small odd-spaced space.
Figure 3: Simulating dynamic hysteresis loss in a power-steering motor using Vector Fields’ hysteresis solver.
To a large extent, performing a CFD analysis on LED lights for heat transfer, radiation or convection cooling is nothing unusual and involves everyday physics that software such as from his firm has been handling for a long time, comments Ivo Weinhold, product manager for engineering fluid dynamics products at the Mentor Graphics Mechanical Analysis Division. However, complex packaging geometries complicate the design process. Once a geometry is created in a CAD package, much time must be spent cleaning it up and generating the mesh, and here FloEFD, which integrates into popular CAD packages, can speed up the design process. Engineers can experiment with new geometries and materials to quickly find out what effects such changes have on thermal properties.
Multiphysics in hybrids
Multiphysics is also key for designing hybrid/electric vehicles, remarks Scott Stanton, technical director of advanced technology initiatives at Ansys. In traditional vehicles, simulation examined isolated areas; today, though, we’re no longer looking at a straightforward energy-storage system with a 12V battery; we’re now dealing with new types of batteries that go from 300V to 600V, along with their electrochemical characteristics and thermal effects. In addition, power electronics are necessary to convert the battery DC to AC to drive the electric motor.
In high-power switching, thermal losses can lead to lower efficiency. Looking into this aspect starts by examining details at a very low level, of how a chip is laid out, and it then expands to include modelling the entire powertrain. The inverter that converts battery DC to AC switches IGBTs (insulated gate bipolar transistors) at 10 to 20kHz, and the controller that steers these devices considers, among other things, vehicle speed, motor torque and driver commands. Device manufacturers typically package six IGBTs into one module that can switch currents as high as 1000A (Figure 2). When dealing with these currents, layout is obviously very important because all the paths must have the identical resistance, otherwise there will be unequal currents and exaggerated stresses in bonding wires. It’s not only important to examine how best to package these modules, it’s also important to determine the optimal way to mount them in the vehicle. Engineers are looking for software packages that can provide this level of scalability.
While at the system level it is becoming possible to model the battery/inverter/motor/control system, it is just as important to perform detailed modelling at the component level, and Ansys feels it is in a good position to provide the tools that address all levels of scalability. In fact, Ansys is active in so many areas of the automotive industry that it sets up its own conference held every two years dedicated to that subject: on 6 and 7 July of this year, the European Automotive Simulation Conference will cover road, rail, racetrack and off-highway vehicle engineering. The program will cover the entire range of physics having relevance in vehicle design, namely structural, fluids, thermal and electronic design automation.
Figure 4: A car’s value is increasingly determined by its innovative electrics/electronics features. A key factor in helping electronics contribute to added value is through the data management of mechatronics.
An auto power-steering motor must meet demanding specs, including the minimisation of torque required to turn a motor’s rotors when the coils are not energised. Employing speciality electrical steels in the motor laminations provides one means of minimising this drag torque, and the Opera simulation tool suite from Vector Fields (a business unit of ERA Technology) provides the means to quickly and accurately design and characterise this aspect before prototypes are built.
In the drive towards high efficiencies, many motor designers now employ more efficient ferromagnetic alloy materials for laminations. Vector Fields’ hysteresis solver (Figure 3) provides the means to accurately simulate the dynamic performance of these materials. With this software, developers can accurately understand the improvements that electrical steels can make as well as explore design ideas that minimise hysteresis effects.
Mechatronics plays a larger role
Two terms that come to mind for the auto of the future, says David Taylor, responsible for automotive and machinery industry marketing for the Siemens PLM Software division, are complexity and electrification. A term related to complexity is mechatronics, which concerns the integration of electrical and mechanical designs, and this aspect is becoming compounded because so much of the future auto is being directed towards electric or hybrid vehicles (Figure 4).
Development teams have increased their levels of productivity within their isolated domains, but they are struggling to integrate them. This cooperation will be critical in the car of the future because of mechatronics. One example is drive by wire, where electronic signals are replacing mechanical linkages such as the steering column. New steering mechanisms, even if electric, must give the driver familiar tactile feedback, so the control electronics must provide a natural feel. Obviously, mechatronics includes everything from software development, wire harnesses and electronics, to mechanical aspects – and no matter what the source, they must all work together.
Siemens’ Teamcenter software is pulling various aspects of mechatronics systems together into a common data backbone that recognises the relationships among these design aspects. Now auto manufacturers can work out problems that have normally only appeared at dealerships; much warrantee work, explains Taylor, is a result of various system components not working together.
Along these lines, Siemens recently announced that Nissan is using Teamcenter for product data management of its GTR, the 2009 Motor Trend car of the year. Taylor gives an additional simple example of how a Teamcenter database can help improve cars while reducing costs. About two years ago, Ford determined that it spent considerable expense replacing ECUs (engine control units). When customers had complaints, repair technicians had no way to trace what version of the software was being used and instead had to replace the entire ECU. Besides the obvious costs of parts and labour, replacing the chip could sometimes introduce vibration problems if the new device wasn’t properly installed. Ford then changed its system to manage the binaries in the ECU by the VIN (vehicle ID number), and the binaries are managed by Teamcenter. A technician looks up which software version is installed and compares it to the latest version; if desired, he can flash the new update into the ECU without removing the component. Ford estimates that this move has saved it approximately $300m in warrantee costs since the program was introduced several years ago.
In summary, Taylor notes: ‘We used to talk about complexity reduction. That’s not realistic if you look at the complex designs of next-generation autos – we now need complexity management.’
In a similar vein, BMW is using CATIA software from Dassault Systemes for the design of all engines across its fuel and diesel-powered cars, motorcycles and hybrid vehicles including what is believed the industry’s first hydrogen-powered vehicle. With this 3D virtual-design platform, engineers can consolidate design environments and create a single reference model for the design of all future BMW engines. In this way, the automaker can harmonise and consolidate all mechanical design initiatives into a single design infrastructure that provides the latest technologies to aid in the software simulation, calculation and testing of new engine models.