Automotive Acoustics Conference 2021

Porsche has launched its first all-electric series. The acoustic was largely uncharted territory. An all-electric vehicle without combustion engine requires a highly differentiated evaluation of acoustic vehicle characteristics such as drivetrain, comfort, chassis comfort, body comfort as well as more stringent requirements in many areas. These include, for example, aeroacoustics, noise such as rattling, crackling, gearing noise, climate acoustics and more.

“Porsche sound” starts with an authentic sound of the engine. Therefore, the customer expects a special sound experience in his new Taycan. Here, Porsche puts a special task in the list of objectives: an emotional Porsche—typical sound with drive-dominated sound components.

The acoustic feedback is achieved through high load dynamics particularly during sporty driving.

In the initial phase of the project, high demands were placed on the sporty and dynamic sound design. At the same time disturbing noise is not noticeable. The sound is to give the Porsche Taycan an unique character.

Vehicle concept features as the two-speed transmission on the rear axle or the use of permanent magnet synchronous machines (PSM) determine the sound image and the dynamic behavior.

The innovative drive concept includes body-mounted powertrain mounts for optimum driving comfort, applicative adapted wheel distribution, targeted pairing of the gear wheels for acoustic optimization as well as a significant reduction in acoustically dominant engine orders.

Furthermore, the acoustic properties of ride, handling, body comfort, acoustic quality, brake acoustics and climate acoustics are described in the paper. Customer relevant objectives are compared with benchmark results such as interior noise and vibration comfort. This article provides an overview of the development priorities and the property profile of the new sporty electric Taycan in the C-segment.

Exemplary product formative innovations such as high-performance brake systems and the innovative air ventilation system and the share of the acoustic expression of the new Porsche Taycan are described.

Automatically detecting anomalous noise in vehicles and identifying their cause(s) is a challenge, which, as long as it exists, has never been resolved in any satisfactory and universal manner, until now. Given the complexity of the input variables, conventional methods of signal processing quickly reach their limits. This is where AI-based methods of detecting and classifying anomalous acoustic events promise significant advancements.

In this paper we show how to detect anomalous vehicle noises with an AI-based approach, we evaluate the precision of the algorithm and possible optimizations, give a concept on how to integrate such a functionality into modern cars, and, finally, give an outlook on latest advancements.

Road noise is undoubtedly the “acoustic enemy number one” for present and upcoming xHEV and EV vehicles. In the very typical customer usage of city driving for instance road noise is the major source.

This article deals with the prediction of tire exterior noise, a topic that has gained substantial relevance in the last few years for two important reasons. The first one is the update of the test procedure for the measurement of Pass-By noise, which has strongly increased the contribution that tire noise gives to the Pass-By level. The second one is the proliferation in the number of Battery Electric Vehicle models developed by car manufacturers, for which tire noise is obviously a major (if not the only) contribution to exterior noise.

A widely used “synthesis model” for the prediction of tire exterior noise relies on the so-called source-transfer-receiver scheme, where the tire is modeled as a set of monopoles whose volume velocity Q is evaluated based on near-field test data acquired in operative conditions and the transfer is correspondingly modeled in terms of p/Q transfer functions. This procedure, while being approximate (the mechanisms underlying the actual noise radiation from tires are very complex and hardly representable by a simple set of monopoles) has the big advantage of being simple and effectively applicable during vehicle development. Furthermore, making a clear distinction between source and transfer, it lends itself to the definition of separate targets for tire manufacturers on the one hand and suppliers of passive treatments on the other hand.

However, when applying this procedure, the positioning of the monopoles and the definition of the corresponding strength is normally accomplished by means of semi-empiric models and/or engineering judgement, whose justification has seldom been deeply investigated. This is what is done in this paper, where a more refined way for the definition of the position and of the strength of the monopoles is described and applied, in which these quantities are defined by means of an optimization process based on an L1-norm regularization technique.

The method is described in detail and it is applied to the prediction of the tire Pass-By noise of a Battery Electric Vehicle. The prediction is validated with outdoor Pass-By tests with different types of tires and with different passive treatments, in such a way to obtain a comprehensive assessment of the method capabilities and limitations.

Current UN ECE vehicle noise regulation requires a demanding reduction of the overall noise level emitted to the outside to the benefit of the health of citizens and road users. New limits have been implemented in the UN R51 Regulation to reduce noise. To achieve these goals, car manufacturers need to refine their vehicle set-up, and a purely experimental approach may come too late. This article shows a method for simulating pass-by noise that exploits numerical transfer functions and a library of experimentally characterized sources with the aim of reducing noise and finding a better compromise between costs and effectiveness of the modifications. In addition, a simple software tool was created for data processing and to facilitate the workflow, used for the evaluation of the rank of the different paths.

Several trends could currently be observed in the automotive industry which are affecting Noise, Vibration and Harshness (NVH) development for new vehicles or platforms. Firstly, many axle systems are being adjusted for new electrified architectures. Secondly, e-axles are creating comfort relevant structure borne noise at higher frequencies compared to non-electrified axles. Furthermore, virtualization and “front-loading” of the NVH tuning can be seen to avoid intensive hardware testing. The combination of these trends is challenging for Original Equipment Manufacturers (OEMs). As a leading global automotive NVH development supplier, Vibracoustic has devised unique NVH axle test rigs and virtual approaches to support OEMs on their way to superior NVH performance. Vibracoustic employs proprietary NVH test rigs to analyze and optimize entire axles without the surrounding vehicle. Load cells at all interface points, acceleration sensors and other signal sources permit a comprehensive characterization of an axle. In our publication [1] we focused on the frequency band up to 250 Hz while the axle is either subjected to hydraulic shake inputs at the wheel hubs or running on a dyno. In this presentation we want to extend the method for e-axle NVH up to 2 kHz. Vibracoustic is collaborating with OEMs to enrich their virtual NVH development and optimization processes. In this paper we introduce two new approaches to investigate the NVH performance experimentally on full e-axle level and how these approaches are used to foster “front-loading” of NVH developments.

The increasing availability of Electric Vehicle (EV) models in the market raises questions about customers’ expectations regarding interior acoustic comfort. Uncertainty exists regarding reactions to the tonal components of electric powertrain noise and how expectations may adapt to lower interior noise levels. Motivated by these considerations, Autoneum and HEAD acoustics collaborated on a study to develop Sound Quality metrics that assess perceived NVH quality in EVs, with special emphasis on tonal noises. The study involved 14 vehicles under different operating conditions, representing various contributions from electric powertrains and tire noise. This article presents the study’s methodology, results, and a comparison between the newly developed Sound Quality metrics and traditional metrics (SPL, AI) for evaluating NVH quality in EVs. Furthermore, the application of these metrics to assess the impact of passive treatments on acoustic quality in a Battery Electric Vehicle is discussed with one example.

One of the key advantages of Operational Transfer Path Analysis (OTPA) as an NVH refinement framework is the maintenance of boundary conditions of the system under study since transfer functions are obtained using operational data instead of modal and acoustic excitations. Consequently, its results are sensitive to instrumentation and hence it demands a greater care for the inclusion of all coherent propagation paths within the vehicle. Application of OTPA in the automotive industry for vehicles with combustion engines is well studied, however there remains a vast scope for its implementation in modern electric vehicles due to their propulsion system’s higher operating frequency as well as the increase in directivity and mobility of high frequency noise. The presented research paper aims at utilizing OTPA as a comprehensive case study on high frequency noise propagation in an early prototype Battery Electric Vehicle (BEV) with particular focus on the distinction between air-borne and structure-borne noise. Additionally, this work investigates the instrumentation of air-borne indicators with respect to the identified acoustically sensitive paths into the passenger cabin as opposed to the near-field instrumentation commonly utilized in various TPA implementations. Moreover, the presented research employs an indirect approach in the instrumentation of parasitic paths indicators which are often challenging to identify despite their prevalence in BEVs. The outcome of this study investigates the utilization of OTPA for higher frequencies and pinpoints some of the critical paths responsible for their propagation inside the vehicle followed by a validation study that was conducted to verify the estimated path contributions. Eventually, this investigation touches upon some further challenges in frequency distinction between the air-borne and structure-borne contributions.

As autonomous vehicles (AVs) are introduced to the market, it is expected that the customers’ focus while commuting will shift from driver-related to non-driver-related tasks and, as a result, more isolation from road noise, component noise and the environment may be required. Furthermore, with a fully symmetric AV, one must consider that there is no front or back of the vehicle and that there is no seat position that is considered “the most important”; in other words, all seating positions are equally important. As a result, a major focus on airborne noise isolation must be considered early in the vehicle development phase to avoid either costly changes late in the program or dissatisfied customers in the market. To achieve this, a Statistical Energy Analysis (SEA) model was utilized to support the development of a “new” vehicle platform for an autonomous electric vehicle. It is well known that traditional SEA models are developed from existing prototypes or mule vehicles; however, as a new incumbent original equipment manufacturer (OEM) this is not an option. This paper will highlight how existing and well-established computer-aided engineering (CAE) methodologies, such as SEA and Finite Element Method (FEM), can support the NVH design and development of a new vehicle architecture. The initial application of loads in the SEA model based solely on virtual models will be described, the challenges of unique load paths for an AV architecture, and the limitations of surface areas to be utilized for sound absorption within the cabin of an AV will be presented. Finally, the simulation results are validated against test results. The utilization of the SEA model in the virtual development phase in the program enabled the right targets to be set early on, avoid surprises with such a unique vehicle, and a reduction of turnaround time for countermeasure development.