Engineers working in today’s vehicle-development environment are faced with a unique challenge in relation to prototype vehicles: the cost of constructing prototypes and assemblies limits their accessibility—and contemporary development cycles often outpace the opportunity to build physical prototypes. And when a prototype vehicle or assembly is produced, it quickly may no longer be representative of the most current vision.
While the most-thorough method for evaluating tire suitability for a given vehicle obviously is to drive it with the specific tires installed, the increasing limitations of access to driveable prototypes can be an obstacle. But there have been significant advances in virtual tools that can assist with the process of developing an optimized tire and vehicle system.
Contemporary computational resources now are readily available and hardware and software capabilities are exceptionally powerful. These tools can be utilized to create simulation models to predict the performance of one iteration of a vehicle’s design with a specific tire—and also expedite multi-variate and multi-physics optimization processes.
The subsystems and components which represent the vehicle model must reflect an actual or viable part, of course, and there lies the potential conflict: the pneumatic tire is a complicated, non-linear component and modeling its complex properties requires careful attention to detail. Simply ordering “a tire model” will generally not produce final results as accurate as those that can be obtained through a cooperative process.
In the past, the tire-development process typically entailed the creation of a specification document, program kickoff, several rounds of experimental physical builds and an eventual winnowing of that data to a suitable tire specification produced by one or two vendors.
The process requires access to prototype vehicles at different stages of development, submission tires and expert evaluation by a driver. The process does identify suitable tires for the vehicle but is somewhat limited in scope. And, not the least, it can require a timeline of several years.
Nothing will completely replace the expert human evaluator. The goal of every vehicle program is to create a vehicle that delivers a satisfying and safe driving experience for the customer; the expert evaluator is the best option to rate a given tire design against that target.
But there are several advantages to augmenting the classic approach with computational methods. Simulation permits investigation of components which cannot be physically installed onto a prototype vehicle. Designed experiment methods can incorporate hundreds or thousands of computational variations that show how vehicle responses change with respect to component properties.
Moreover, virtual models can be constructed rapidly and inexpensively, especially when compared to the cost of testing with hand-made prototype vehicles. The computer modeler has the significant advantage of knowing exactly what is in the model, whereas the details of components in a physical prototype may be uncertain or even unknown.
A computational tire model is required for any type of full-vehicle simulation modeling. These mathematical representations of tire data serve as the numeric interface between the virtual vehicle and virtual proving ground. Because tires are the foundation of vehicle handling and performance, vehicle models require the data produced by tire models to reliably predict how the vehicle will handle under different driving situations.
The ensuing computer simulation offers many benefits to the design process, such as eliminating the need for high-cost testing methods, providing a rapid objective evaluation of vehicle performance and safely investigating design alternatives.
Generating tire models
The tire-modeling process starts with data. A variety of tire measurements may be required, including force and moment, spring rates, impact response and geometry. Traditionally, tire models have been generated with little or no knowledge of how the tire was tested or its end use. Typically, data in these cases was collected by a test organization using a series of standard procedures.
In a new and more-integrated approach centered around the needs of the modeler or tire analyst, the choice of measurements and test conditions will depend on the ways the tire model will be used and the conditions under which the tire will operate. More-advanced tire models then are developed and certified for fit within the design parameters of intended use. In this approach, direct feedback on model performance is key to adjusting test procedures and improving future testing.
Identifying the right tire model involves a detailed understanding of how the model will be used. Commercial tire models are available and typically adequate for most general applications. Custom tire models are required only with specialized simulations or targeted analysis. Tire models can be broken into three classes: performance-based, component-level and finite-element models.
Performance-based models generally are employed to evaluate vehicle handling characteristics such as load, slip angle, speed and driving/braking torque. They rely on force and moment testing—without physical construction being characterized—and use either tables, splines or equation-based models. The Pacejka or, “magic formulas” model is one of the most widely used models in this class because of its accuracy, ease in programming and speed. The Pacejka equations were developed to fit the data gathered from experimental tests with real tires—and predict behavior with great precision.
Component-level tire models test the structure and physical properties of the tire, with a rigid or flexible ring and are most commonly used for vehicle ride and durability analysis. While complex in design, component-level tire models are the most appropriate tool for this analysis. A strong example is the FTire, flexible-ring tire model, a full 3D in-plane and out-of-plane tire simulation designed to test vehicle comfort and performance in relation to road irregularities. Other popular component models include MF-Swift 6.0 and 6.1 and CD-Tire.
Finite-element models require a detailed knowledge of the tire’s component materials and internal structure. These models are most useful for tire manufacturers, since specific details of elastomer behavior and geometry can be difficult or expensive for an outsider to acquire.
Once a tire model is selected and produced, the customer should have access to the raw data collected for testing, along with a report illustrating the model’s performance.
Testing and data-fitting
For automakers, finding the right partner to assist in making the best vehicle-design decisions is crucial. Tire and modeling experts who work with manufacturers at every stage of the supply chain, from initial ingredient evaluation to end use, offer a unique advantage in choosing the right tests and “fitting” the data. It can be beneficial to work with a partner with the expertise and capability to combine modeling, subjective handling and traditional laboratory tire testing.
Instead of relying on generic procedures, robust tire models are designed to meet the needs of the vehicle OEM and require a detailed understanding of the tire’s properties, design and intended use. By choosing a partner that offers a customer-centered approach to tire modeling, manufacturers can count on advanced, tailored testing and modeling techniques and a commitment to improving the future of tire and vehicle safety.
As the industry continues to evolve toward the use of more virtual testing and development, the need will grow for more precise modeling techniques. The tire is the foundation of the vehicle’s interaction with the road and continued work on higher-fidelity tire models will go a long way toward meeting the industry’s future needs.
Dean Tener is Technical Manager at Smithers Rapra Ravenna (OH) Laboratory, Smithers Rapra’s main tire testing center in North America. He has held positions at General Motors, Honda R&D Americas and Bridgestone/Firestone.
Michael Stackpole is the founder and President at Stackpole Engineering, a technical services provider specializing in tire testing, data-fitting and advanced tire model development. He has more than 30 years’ experience in tire and vehicle simulation, testing and modeling that includes positions at Goodyear and Firestone.