Search Results

In vehicle NVH development, vibroacoustic simulations with Finite Element (FE) models are a common technique. The computational costs for these calculations are steadily rising due to more detailed modelling and higher frequency ranges. At the same time, the need for multiple evaluations of the same model with different input parameters, e.g., for uncertainty quantification, optimization, or robustness investigations, is also increasing. Therefore, it is crucial to reduce the computational costs in these cases. A common technique is to use surrogate models that replace the computationally intensive FE model to perform repeated evaluations. Several different methods in this area are well established, but with the continuous advancements in the field of machine learning, interesting new methods like the Gaussian Process (GP) regression arises as a promising approach.

This paper presents the application of statistical process control (SPC) methods to Windshear, a 180-mph motorsports and automotive wind tunnel equipped with a wide-belt rolling road system. The SPC approach captures the complete variability of the facility and offers useful process performance metrics that are based on a sound statistical framework. Traditional control charts are explored, emphasizing the uniqueness of variability experienced in wind tunnels which includes significant, unexplained short-term and long-term variation compared to typical manufacturing processes. This unique variation is elegantly captured by the three-way control chart, which is applied to estimate the complete process reproducibility with different levels of repeatability of vehicle drag coefficient. The sensitivity of three-way control charts is explored including the evaluation of an alternate group assignment within the same dataset.

Vehicle thermal management (VTM) simulations are becoming increasingly important in the development phase of a vehicle. These simulations help in predicting the thermal profiles of critical components over a drive cycle. They are usually done using two methodologies: (1) Solving every aspect of the heat transfer, i.e., convection, radiation, and conduction, in a single solver (Conjugate Heat Transfer) or (2) Simulating convection using a fluid solver and computing the other two mechanisms using a separate thermal solver (Co-simulation). The first method is usually computationally intensive, while the second one isn’t. This is because Co-simulation reduces the load of simulating all heat transfer mechanisms in a single code. This is one of the reasons why the Co-simulation method is widely used in the automotive industry. Traditionally, the methods developed for Co-simulation processes are load case specific.

The continuous encouragement of lightweight design in modern vehicles demands a reliable and efficient method to predict and ameliorate the interior acoustic comfort for passengers. Due to considerable psychological effects on stress and concentration, the low frequency contribution plays a vital rule regarding interior noise perception. Apart other contributors, low frequency noise can be induced by transient aerodynamic excitation and the related structural vibrations. Assessing this disturbance requires the reliable simulation of the complex multi-physical mechanisms involved, such as transient aerodynamics, structural dynamics and acoustics. The domain of structural dynamics is particularly sensitive regarding the modelling of attachments restraining the vibrational behaviour of incorporated membrane-like structures. In a later development stage, when prototypes are available, it is therefore desirable to replace or update purely numerical models with experimental data.

It is particularly easy to get tunnel vision as a domain expert, and focus only on the improvements one could provide in their area of expertise. To make matters worse, many Original Equipment Manufacturers (OEMs) are silo-ed by domain of expertise, unconsciously promoting this single mindedness in design. Unfortunately, the successful and profitable development of a vehicle is dependent on the delicate balance of performance across many domains, involving multiple physics and departments. Taking for instance the design of a Heating, Ventilation & Air Conditioning (HVAC) system, the device’s primary function is to control the climate system in vehicle cabins, and more importantly to make sure that critical areas on the windshield can be defrosted in cold weather conditions within regulation time. With the advent of electric and autonomous vehicles, further importance is now also placed on the energy efficiency of the HVAC, and its noise.

It is essential to include uncertainties in the simulation process in order to perform reliable vibroacoustic predictions in the early design phase. In this contribution, uncertainties are quantified using the generalized Polynomial Chaos (gPC) expansion in combination with a Finite Element (FE) model of a vehicle body in white. It is the objective to particularly investigate the applicability of the gPC method in the industrial context with a high number of uncertain parameters and computationally expensive models. A non-intrusive gPC expansion of first and second order is implemented and the approximation of a stochastic response process is compared to a Latin Hypercube sampling based reference solution with special regard to accuracy and computational efficiency. Furthermore, the method is examined for other input distributions and transferred to another FE model in order to verify the applicability of the gPC method in practical applications.

During the last decades, big steps have been taken towards a realistic simulation of NVH (Noise Vibration Harshness) behavior of vehicles using the Finite Element (FE) method. The quality of these computation models has been substantially increased and the accessible frequency range has been widened. Nevertheless, to perform a reliable prediction of the vehicle vibroacoustic behavior, the consideration of uncertainties is crucial. With this approach there are many challenges on the way to valid and useful simulation models and they can be divided into three areas: the input uncertainties, the propagation of uncertainties through the FE model and finally the statistical output quantities. Each of them must be investigated to choose sufficient methods for a valid and fast prediction of vehicle body vibroacoustics. It can be shown by rough estimation that dimensionality of the corresponding random space for different types of uncertainty is tremendously high.

The reduction of vehicle exhaust particle emissions is a success story of European legislation. Various particle number (PN) counters and calibration procedures serve as tools to enforce PN emission limits during vehicle type approval (VTA) or periodical technical inspection (PTI) of in-use vehicles. Although all devices and procedures apply to the same PN-metric, they were developed for different purposes, by different stakeholder groups and for different target costs and technical scopes. Furthermore, their calibration procedures were independently defined by different stakeholder communities. This frequently leads to comparability and interpretation issues. Systematic differences of stationary and mobile PN counters (PN-PEMS) are well-documented. New, low-cost PTI PN counters will aggravate this problem. Today, tools to directly compare different instruments are scarce.

This paper describes a novel approach for modeling an automotive HVAC unit. The model consists of black-box models trained with experimental data from a self-developed measurement setup. It is capable of predicting the temperature and mass flow of the air entering the vehicle cabin at the various air vents. A combination of temperature and velocity sensors is the basis of the measurement setup. A measurement fault analysis is conducted to validate the accuracy of the measurement system. As the data collection is done under fluctuating ambient conditions, a review of the impact of various ambient conditions on the HVAC unit is performed. Correction models that account for the different ambient conditions incorporate these results. Numerous types of black-box models are compared to identify the best-suited type for this approach. Moreover, the accuracy of the model is validated using test drive data.

Vehicle thermal management (VTM) simulations constitute an important step in the early development phase of a vehicle. They help in predicting the temperature profiles of critical components over a drive cycle and identify components which are exceeding temperature design limits. Parts with the highest temperatures in a vehicle with an internal combustion engine are concentrated in the engine bay area. As packaging constraints grow tighter, the components in the engine bay are packed closer together. This makes the thermal protection in the engine bay even more crucial. The fan influences the airflow into the engine bay and plays an important role in deciding flow distribution in this region. This makes modelling of the fan an important aspect of VTM simulations. The challenge associated with modelling the fan is the accurate simulation of the rotation imparted by the fan to the incoming flow. Currently, two modelling approaches are prevalent in the industry.

From an acoustical point of view, the integration of seatbelt retractors in a vehicle is a real challenge that has to be met early in the vehicle development process. The buzz and rattle noise of seat belt retractors is a weak yet disturbing interior noise. Street irregularities excite the wheels and this excitation is transferred via the car body to the mounting location of the retractor. Ultimately, the inertia sensor of the locking mechanism is also excited. This excitation can be amplified by structural resonances and generate a characteristic impact noise. The objective of this paper is to describe a simulation method for an early development phase that predicts the noise-relevant low frequency local modes and consequently the contact of the retractor with the mounting panel of the car body via the finite element method.

Over the past decades, interior noise from wind noise or engine noise have been significantly reduced by leveraging improvements of both the overall vehicle design and of sound package. Consequently, noise sources originating from HVAC systems (Heat Ventilation and Air Conditioning), fans or exhaust systems are becoming more relevant for perceived quality and passenger comfort. This study focuses on HVAC systems and discusses a Flow-Induced Noise Detection Contributions (FIND Contributions) numerical method enabling the identification of the flow-induced noise sources inside and around HVAC systems. This methodology is based on the post-processing of unsteady flow results obtained using Lattice Boltzmann based Method (LBM) Computational Fluid Dynamics (CFD) simulations combined with LBM-simulated Acoustic Transfer Functions (ATF) between the position of the sources inside the system and the passenger’s ears.

Highly automated driving (HAD) is under rapid development and will be available for customers within the next years. However the evidence that HAD is at least as safe as human driving has still not been produced. The challenge is to drive hundreds of millions of test kilometers without incidents to show that statistically HAD is significantly safer. One approach is to let a HAD function run in parallel with human drivers in customer cars to utilize a fraction of the billions of kilometers driven every year. To guarantee safety, the function under test (FUT) has access to sensors but its output is not executed, which results in an open loop problem. To overcome this shortcoming, the proposed method consists of four steps to close the loop for the FUT. First, sensor data from real driving scenarios is fused in a world model and enhanced by incorporating future time steps into original measurements.

The General Motors Reduced Scale Wind Tunnel Facility, which came into operation in the fall of 2015, is a new state-of-the-art scale model aerodynamic test facility that expands GM’s test capabilities. The new facility also increases GM’s aerodynamic testing through-put and provides the resources needed to achieve the growing demand for higher fuel economy requirements for next generation of vehicles. The wind tunnel was designed for a nominal model scale of 40%. The nozzle and test section were sized to keep wind tunnel interference effects to a minimum. Flow quality and other wind tunnel performance parameters are on par with or better than the latest industry standards. A 5-belt system with a long center belt and boundary layer suction and blowing system are used to model underbody flow conditions. An overhead probe traverse system is installed in the test section along with a model positioning robot used to move the model in an out of the test section.

A local Gaussian process regression approach is presented, which allows to model nonlinearities of internal combustion engines more accurate than global Gaussian process regression. By building smaller models, the prediction of local system behavior improves significantly. In order to predict a value, the algorithm chooses the nearest training points. The number of chosen training points depends on the intensity of estimated nonlinearity. After determining the training points, a model is built, the prediction performed and the model discarded. The approach is demonstrated with a benchmark system and air charge test bed measurements. The measurements are taken from a turbocharged SI gasoline engine with both variable inlet valve lift and variable inlet and exhaust valve opening angle. The results show how local Gaussian process regression outmatches global Gaussian process regression concerning model quality and nonlinearities in particular.

Air charge calibration of turbocharged SI gasoline engines with both variable inlet valve lift and variable inlet and exhaust valve opening angle has to be very accurate and needs a high number of measurements. In particular, the modeling of the transition area from unthrottled, inlet valve controlled resp. throttled mode to turbocharged mode, suffers from small number of measurements (e.g. when applying Design of Experiments (DoE)). This is due to the strong impact of residual gas respectively scavenging dominating locally in this area. In this article, a virtual residual gas sensor in order to enable black-box-modeling of the air charge is presented. The sensor is a multilayer perceptron artificial neural network. Amongst others, the physically calculated air mass is used as training data for the artificial neural network.

One of the remaining challenges in the simulation of the aerodynamics of ground vehicles is the modeling of the airflows around the spinning tires and wheels of the vehicle. As in most advances in the development of simulation capabilities, it is the lack of appropriately detailed and accurate experimental data with which to correlate that holds back the advance of the technology. The flow around the wheels and tires and their interfaces with the vehicle body and the ground is a critical area for the development of automobiles and trucks, not just for aerodynamic forces and moments, and their result on fuel economy and vehicle handling and performance, but also for the airflows and pressures that affect brake cooling, engine cooling airflows, water spray management etc.

In recent years, there has been renewed attention focused on open jet correction methods, in particular on the two-measurement method of E. Mercker, K. Cooper, and co-workers. This method accounts for blockage and static pressure gradient effects in automotive wind tunnels and has been shown by both computations and experiments to appropriately adjust drag coefficients towards an on-road condition, thus allowing results from different wind tunnels to be compared on a more equitable basis. However, most wind tunnels have yet to adopt the method as standard practice due to difficulties in practical application. In particular, it is necessary to measure the aerodynamic forces on every vehicle configuration in two different static pressure gradients to capture that portion of the correction. Building on earlier proof-of-concept work, this paper demonstrates a practical method for implementing the two-measurement procedure and demonstrates how it can be used for production testing.

Unsteady flow over automotive side-view mirrors may cause flow-induced vibrations of the mirror assembly which can result in blurred rear-view images, adversely affecting marketability through customer comfort and quality perception. Prior research has identified two mechanisms by which aerodynamically induced vibrations are introduced in the mirror. The first mechanism is unsteady pressure loading on the mirror face due to the unsteady wake, causing direct vibration of the mirror glass. The second mechanism, and the focus of this study, is a fluctuating loading on the mirror housing caused by an unsteady separation zone on the outer portion of the housing. A time-dependent Computational Fluid Dynamics (CFD) methodology was developed to correctly model mirror wake behavior, and thereby predict flow-induced mirror vibration to improve performance estimations.

The German funded project ARAMiS included work on several demonstrators one of which was a multicore approach on large scale software integration (LSSI) for the automotive domain. Here BMW and Audi intentionally implemented two different integration platforms to gain both experience and real life data on a Hypervisor based concept on one side as well as using only native AUTOSAR-based methods on the other side for later comparison. The idea was to obtain figures on the added overhead both for multicore as well as safety, based on practical work and close-to-production implementations. During implementation and evaluation on one hand there were a lot of valuable lessons learned about multicore in conjunction with safety. On the other hand valuable information was gathered to make it finally possible to set up a cost model for estimation of potential overhead generated by different integration approaches for safety related software functions.