Virtual Real Driving Emissions (vRDE) Part 2: Simulation Solutions

Written by Damian Sokolowski

May 31, 2017

In Part 1 of our discussion about Real Driving Emissions, we covered the shortcomings of traditional fuel economy and emissions testing methods using chassis dynamometers and fixed driving cycles. In Part 2, we will talk about ways in which simulation can replace the need for costly on-road testing, especially upstream in the product line development cycle.

The Traditional Approach

The way in which a driving cycle is performed on a chassis dynamometer already aims to simulate on-road behavior by imposing a load onto driven axles using rollers, typically connected to electric machines. Road loads (aerodynamics, inertia, etc.) have to be simulated by the dynamometer since the vehicle is actually stationary during the entire test. The test therefore only approximates on-road operation of these vehicles.

Numerical simulation offers a cost- and time-effective alternative to chassis dyno testing given that we have access to highly predictive and fast running engine and vehicle models. The preferred approach for this type of simulation is to dynamically target a chosen driving cycle via a controller or driver template. This method limits simulation vehicle performance to the maximum engine output and will reveal shortcomings of the powertrain or control strategies, in contrast to kinematically imposing the vehicle speed and back-calculating the engine load required.

Figure 8: Typical Dynamic Driving Cycle Vehicle Model

The unique advantage offered for these types of simulations by GT-SUITE is the ability to vary the fidelity level of each of the subassemblies representing the various subsystems. For example, we may want to consider fully-predictive GT-POWER engine models with turbocharging and predictive combustion, or use a fast-running map-based lookup resulting from dyno data. For the transmission, we may use a simple kinematic gear ratio or a discretized dual clutch transmission with individual gear meshes, synchronizers, and internal clutches. The driver can be one of our model-based controllers or a custom-built Simulink model with advanced features and production-level controls. Additionally, the model can easily be expanded to include other physical domains that might include thermal management, lubrication, cabin comfort, or, most importantly for Real Driving Emissions, aftertreatment.

The Case for an Integrated Approach

Since RDE driving cycles have inherent variability by design, it’s much less practical to calculate boundary conditions of independent subsystems and impose them in standalone design simulations. A modular modeling approach allows users to create subsystem models that can be run by themselves or coupled with one another in an integrated effort.

Figure 9: Example of Integrated Vehicle Model

This is especially true for thermal management and emissions aftertreatment systems, which strongly depend on the dynamic loading of the vehicle and transient control strategies. Options such as active grill shutters and electrically-heated catalysts can only be accurately studied during transient operation. A system architect may begin by making technology choices using very coarse models and steady-state simulations, but the task of validation and certification can only be considered through integration where the engine, vehicle, thermal management, aftertretament, and all relevant controls are simultaneously represented.

The GT-SUITE aftertreatment library features a fast and accurate advanced adaptive chemistry solver with a flexible interface that can model all catalysts and diesel particulate filters on the market today. Additionally, the library features state-of-the-art mechanisms and validation results in the form of example models. Even when coupled to a vehicle model, the aftertreatment system can be simulated tens to hundreds of times faster than real time.

Figure 10: Aftertreatment Model Example in GT-SUITE

With accurate measurement of engine out emissions, we can then correctly predict the tailpipe out emissions that a PEMS system will read during a test without the need for any of the components to be on hand. Adding a vehicle and driver to this system gives us everything we need for a proper RDE simulation.

RDE Driving Cycle Generation and Simulation

One of the more unusual features of the transition to real driving for emissions measurements is that the cycle is random by design. This means that adding it to a simulation is a lot more complicated than simply subjecting a virtual vehicle to another WLTP or FTP-75 cycle, which is a prescribed speed function of time. The RDE cycle requires adherence to a set of rules, but to be certain of meeting certification levels, we must cover a broad range of permutations that might be experienced during the actual test.

To that end, Gamma Technologies has developed a RDE cycle generator that draws on a database of real driving segments. The’ProfileRealDriving’ template gives the user control over the most important shaping parameters of the driving cycle and randomly generates one at the start of each simulation. For example, the user has the ability to determine the driving cycle distribution between urban, rural, and motorway portions, or to include elevation variability for additional load variation. Additionally, the cycle generator can be populated with user’s own measurement data for generation of RDE-compliant cycles with an OEM’s own source.

Figure 11: Example of Random RDE Cycle from 'ProfileRealDriving'

This new driving cycle template can be referenced directly by any of the GT-SUITE driver templates and targeting controllers. In addition to the new ability to randomly generate a driving cycle, GT-SUITE’s recently revised ‘ProfileGPS’ template also simplifies replication of testing routes in the simulation environment. Users can reconstruct a recorded drive by either extracting latitude, longitude, and elevation data, or by linking to a GPX file directly. GT-SUITE will then automatically extract the relevant information from the file and reconstruct that driving cycle, along with powerful, built-in filters for smoothing any noisy data.

GT-SUITE – A Comprehensive vRDE Tool

In addition to being able to simulate each of the relevant systems with a high level of detail, GT-SUITE offers a mature platform in which to integrate all vehicle systems as a replacement to costly physical testing of prototypes. The new methods called for by RDE legislation are supported in the current implementation and will be continuously monitored for any revisions and reflected by the software.

Beyond the core capability, GT-SUITE also includes a fully featured post processor in GT-POST, which allows for assembly of reports and compliance validation directly within the native environment. A suite of cosimulation harnesses for most leading commercial tools allows for direct linking with existing models, regardless of the platform they were developed on. All this, coupled with the ability to reuse the model in Model in the Loop (MiL), Software in the Loop (SiL), and Hardware in the Loop (HiL) tests, means that the effort put forth for vRDE can be reused throughout the entire development process, up until hardware becomes available.

Virtual Real Driving Emissions are an area that will dramatically challenge status quo approaches to emissions and fuel economy compliance. An integrated simulation approach with GT-SUITE provides the necessary support to process changes and help OEMs meet the new standards at every step of the way, from concept to production.

For more information about GT-SUITE’s vRDE capabilities, contact Damian Sokolowski at [email protected]

Written By: Damian Sokolowski