Expanding NVH prediction applications fivefold while eliminating physical prototypes

Leaders in automotive comfort

Renault S.A. (Renault) has been a pioneer in the automotive industry since the company was founded in 1899. In addition to focusing on performance and reliability, it has always strived to provide the most comfortable and enjoyable driving experience.

Delivering the highest levels of comfort requires optimizing the noise, vibration and harshness (NVH) of vehicles. With the advent of electrification, sounds are much less
masked by the engine, so NVH analysis and optimization has become even more
important.

Twenty years ago, Renault identified how crucial it was for teams developing different
systems within vehicles to work more closely to achieve the best results. The earlier in the
design process that interactions between systems can be understood and changes made, the better. Otherwise, they become too expensive and time-consuming.

To facilitate this, Renault instigated the Virtual Synthesis in Acoustic (VISA) project. “The ambition was to be able to make synthesis for forecasting global vehicle performances by assembling separate components,” explains Philippe Mordillat, NVH simulation expert at Renault.

Bringing teams together

The VISA tool was designed to assess the NVH behavior of a vehicle at any stage of the development process. The aim was to implement a modular and system-driven design
methodology to bridge the gap between the organization’s teams and external suppliers.

There were many challenges because the software available then was less advanced than it is now. For instance, technologies for coupling different subsystems based on frequency
response functions (FRFs) were available but lacked the correct boundary conditions to ensure accurate results. Additionally, methodologies for characterizing noise and vibration sources in an invariant way simply did not exist.

There was also a clear separation between testing and CAE activities: exchange and re-use of data between departments was limited, making it difficult to connect the databases and different component models to the synthesis tool in a flexible way. Methods to address complex hybrid-electric vehicle (HEV) architecture, where models of the different subsystems must be linked to boundary conditions that come from the control logic, were also yet to be created.

According to Mordillat, this limited the usability of the synthesis and prediction method Renault was looking for to broaden the NVH spectrum. “In the end, 90 percent of the user cases we could analyze were linked to pure components weakly coupling; for instance, softly mounted and applications for low or mid-frequency booming noise. We wanted to extend the applicability of the methods to higher frequencies such as road noise and pass-by noise cases.”

Expert partnership

Renault’s desire to further develop this tool and their internal processes led to the strategic
collaboration with Siemens Digital Industries Software. Since 2016, the companies have worked together to improve the VISA platform by leveraging Simcenter™ Testlab™ Virtual Prototype Assembly (VPA) software, which is part of the Siemens Xcelerator business platform of software, hardware and services.

The goal of this partnership was to implement a holistic, company-wide process to reliably assess NVH performance of an assembled vehicle virtually without a physical prototype. This would require new methods with improved reliability and accuracy of prediction, such as the interaction between strongly coupled components. The process needed to be available at any stage of development, supporting any test or computer-aided engineering (CAE) based data source either internally or from suppliers. It also needed to be able to be used by all Renault engineers, not just simulation experts, so all teams could make full use of the new tool.

To meet these challenges, two teams, Simcenter Engineering and Consulting services and Simcenter Test and Engineering services, collaborated with Renault engineers to define the use cases and apply the methods to relevant situations.

Component-based transfer path analysis

With road noise no longer masked by a combustion engine, noise sources such as tires need more attention in electric vehicles. Noise coming from the tire/road contact typically is delivered in a broad frequent spectrum, combines structural and acoustic excitation and involves strongly coupled subsystems (the wheel rim is bolted to the spindle). So having the right technologies for source decomposition is crucial to efficiently and synergistically working with tire suppliers.

Simcenter specialists supported Renault in deploying a component-based transfer path
analysis (TPA) method to identify the noise sources coming from the interaction between
tire and road in an invariant way. The wheel is characterized as an independent component
by means of a set of wheel-center loads, often referred to as blocked forces. Impedance matrices, in the form of 6 degrees-of-freedom (DOF) FRFs, are acquired to describe the properties of the connection between the rim and spindle.

This data is the technical backbone of the synthesis method, describing both the source
component generating the noise and vibration and the way that component is connected and interacts with the other subsystems.

These new capabilities are just one building block of the complete toolset now provided
with Simcenter Testlab to acquire and validate the invariant component models ready for
future use in assemblies.

Additionally, this coupling technology allows the original equipment manufacturer (OEM) and supplier to exchange data related to product definitions and knowledge going beyond the performance data that is typically shared. This means they can properly set and work on
common criteria and targets for individual components that will then be integrated into
the full system.

Crucially, there is no risk to confidentiality as no detailed specifications or geometry are
shared, only curves such as FRFs or loads. As a result, suppliers can build virtual assemblies
and check the impact of their component’s design on the full vehicle NVH performance.

Broadening user base and growing data

Renault envisions deploying this technology broadly across teams and sites. As it is used more widely over time, Renault will build a database of components with vast amounts of historic data that can be used in future designs.

Simcenter Testlab data management has been deployed to make sure these components and data are stored centrally, can be easily searched and are documented for correct re-use in assemblies. Components can be used to build virtual assemblies and run calculations for different use cases and combinations, even by users that did not author the models of those components and do not know which data they contain.

As Mordillat notes, “The reason for using Simcenter Testlab VPA is to have an easy and
standardized process that users can follow in their everyday activity, no matter which
engineering site or vehicle model they are working on.”

Initially, around 50 users at Renault will be using the platform, 25 of whom are CAE users who will have an active role in generating component input. The number of users will grow as performance analytics are customized to focus specifically on the needs of the organization.

Customized analysis

For instance, when analyzing pass-by noise (PBN), engineers need to have a reliable predication to meet the regulations for exterior noise. Renault already had the knowledge and technologies for performing a noise synthesis during a PBN event, but these were not smoothly implemented in the original VISA platform to streamline the process.

Siemens created a dedicated customization of the processing environment in the Simcenter Testlab VPA tool, supporting seamless PBN analysis. This also gives Renault engineers the ability to perform source-transfer decomposition analysis, easily compare configurations and see the effect of design modifications.

As a result, it’s now easier to set targets for different subsystems, predict exterior noise
throughout development and determine what improvements need to be made to meet regulations. Simcenter Testlab VPA also enables better cooperation between NVH engineers and software development so the impact of vehicle controls on NVH performance can be better balanced.

Combining innovations for an exciting future

Mordillat is thrilled with the results so far, as well as excited about the future. “Using Simcenter Testlab VPA expanded our applications for NVH prediction by a factor of five,” he says. “This has drastically increased productivity.”

Improvements are clear throughout the development process, starting with the concept phase where historical vehicle data can be used as a starting point for initial target cascading to the subsystems. This will continue through to the validation stage where new component data is available for the prototype performance assessment. Throughout development, Simcenter Testlab VPA supports step-by-step replacement of historical and reference data with new designs that are added to the libraries for future use.

Renault also hopes to combine its use of Simcenter Testlab VPA with its GREEN platform – the result of another successful collaboration with Siemens. This will be particularly useful for hybrid-electric vehicles where both hardware and software controls need to be evaluated. CAE engineers will build models in the GREEN software environment powered by Simcenter Amesim™ software, while measurement specialists create test-based components using
Simcenter Testlab. These will be combined in the Simcenter Testlab VPA platform to build up the assemblies for the noise predictions.

Ultimately, using Simcenter Testlab VPA will enable Renault to predict the effect of design modifications using virtual models and reduce the amount of physical testing required. As Mordillat says, “By using Simcenter Testlab VPA, we have the ability to reduce the number of prototypes used by increasing the quality of our forecasting capabilities in the virtual phase.”

As Renault builds up its library of components and data, the value of the Simcenter Testlab VPA platform will continue to grow and the data management system will enable the introduction of future processes such as using artificial intelligence (AI) for analyzing vast amounts of data.

Mordillat believes the link between Simcenter Testlab VPA and the Simcenter Testlab Vehicle NVH simulator will be powerful, making it possible to subjectively evaluate the sound of the vehicle being developed. “As the digital twin improves, it will be possible to verify how a design modification changes the subjective sound perception. The NVH simulator will be part
of a multi-attribute evaluation where other elements such as active noise cancellation
algorithms and driving styles are evaluated simultaneously.”

When it comes to autonomous vehicles, reducing cabin noise will create better synergy with the car audio and infotainment systems. With users no longer driving vehicles themselves, the most important thing will be their comfort. Thanks to this partnership with Siemens, Renault will be in a position to deliver the very highest level of comfort without exceeding project timelines and budgets.