This paper describes the development of the whole xEV drive system using a multibody dynamics systems modelling approach supported by vehicle and rig testing. This allows all of the subsystems such as motor, transmission and installation, and their interactions, to be considered and their contributions to the full system response to be reviewed.
Excitation sources are described and forced response levels are predicted allowing the relative importance of the electromagnetic and gear forcing to be reviewed.
Methods for improving NVH performance are shown together with assessing any impact on gear and bearing durability.
Ideas for further developments of E-Drives and the multibody modelling approach are presented.

Emission legislation over the short- and medium-term will create very challenging requirements for the development of new powertrains. Increased complexity due to vehicle customization, different hybridization architectures and the market requirement for a vast product portfolio, has generated an astonishing explosion of variants to be calibrated and optimized.
Simulation capabilities are helping test engineers by making it possible to anticipate an ever-widening range of activities through the use of optimization loops using real-time simulators. Real Time simulators need to include human beings into the analysis loops using realistic simulators so that human reactions to vehicle, road and powertrain responses can also be considered. All these techniques are grouped under the term “virtual calibration”. This article provides some insight into these methodologies.

Rising automotive electrification requires major efforts to design the most efficient storage systems whose purpose is managing larger and larger energy loads. Batteries life and reliability depend on the development of a thermal management system to prevent energy losses or damage, whatever the environment. Moreover, OEMs equip their vehicles with high power/efficient electronic components whose energy consumption is increasing day by day, in order to meet customer expectations. As a consequence, engineers have to double their workload to take care of battery pack thermal comfort and grant battery efficiency all vehicle life long. This efficiency is affected by several parameters such as chemical components applied (internal resistance), cells electrical configuration (XsYp), pack thermal conditioning system type and charging power rate.
Although pollutant emission constrained, ICEVs show several times higher energy density compared to BEVs one. For this reason BEVs are much more sensitive about power usage, a process that has to be monitored carefully since the first step of energy conversion: charging phase. During this operation heat losses are generated due to electrical/chemical phenomena occuring inside battery cells: these side effects are negligible in case of refueling conventional ICEVs.
This paper uses 1d simplified model to perform a thermal analysis of a battery pack under development, evaluating the impact on energy efficiency conversion during charging process in different working conditions, taking into account ambient temperatures and main parameters involved.

Stringent pollutant emissions regulations and the urgent need to reduce CO2 production, are pushing engine manufacturers to strongly increase the technological content of automotive powertrain.
Modern Diesel engines coordinate several actuators related to air management, injection control and aftertreatment involving dozens independent variables. This translates in a very complex control optimization problem which should consider several competing targets as performance, efficiency, emissions and NVH.
While traditional calibration procedures rely only on expensive and extremely time-consuming experimental campaigns to optimize engine operation, CAE tools may consistently contribute in reducing the experimental testing. In fact, using a relatively limited amount of experimental data, thanks to the currently widely available computational power, it is possible to develop physical predictive models with the aim of transferring part of the calibration optimization work from experimental testing to virtual platforms.
Powertech Engineering, taking advantage of its long-standing experience in powertrain modelling and simulation, developed a Real Time platform for Virtual Calibration using 0D/1D physical engine models, featuring predictive combustion and emissions models, together with a fully physical aftertreatment model. These high-fidelity sub-domain models are coupled together in a system-wide virtual engine test rig. This virtual engine may be used in Hardware-in-the-Loop (HiL) configuration, working with the real-world Electronic Control Unit (ECU) exploiting well established optimization and calibration procedures, or offline on a common desktop PC if coupled with a virtual controller.

During the development of crash safety features in car developments, thousands of simulations are submitted and analysed for each new platform and its variants. Using Model Reduction methods the output generated by these simulation can be reduced by a factor of 100 without loss of essential information. Model Reduction methods can also be used to improve the robustness of the simulation design and flag new simulation results, which behave different from the results achieved so far.
This presentation will show, how these model reduction methods can be used to understand, whether physical test results are in the span of the simulation results of the same model. We use laser scan results of the crashed model, generate a simulation-like result using automatic orientation of the test results and perform a combined model reduction method on the node positions of the simulation and test results. Scatter plots of the contributions of the domination modes to each result indicate the relative position of the test results to the simulation results.

Nowadays, drag reduction is one of the most important targets established by vehicle manufacturers, when designing D, E, and F-segment passenger cars. This explains the growing popularity of fastback automobiles. However this styling feature entails disadvantages in terms of lift, hence most of these vehicles make use of aerodynamic devices to counteract the lightening of the rear axle at high speed. In this work the aerodynamic development of a fastback car was investigated by means of the open-source software OpenFOAM®. The focus is on the DrivAer flat-bottomed model, designed at the Technische Universität München (TUM) to meet the demands of the local automotive industries. Different aero-packages suitable for aftermarket installation have been conceived and mounted on the vehicle to achieve increasingly higher levels of downforce. Particular attention was given to the balance of the vertical load between front and rear axle, while the chassis ride height remained consistent with the road-legal designated use.

ADAS Systems are a reality of the new generation vehicles, but experts of the Industry advise that billions of kilometers must be covered by autonomous agents to achieve the super-human driving level. Covering such distance to train and to prove the functional safety of the agent may require centuries, and alternative approaches, such as virtual testing, are a necessity.
OEMs are not new to simulation and strive to integrate the different third-party software to complete Simulation System: graphical engines, traffic simulators, vehicle dynamics solvers, etc., must cooperate in a distributed computational environment.
The exhibited framework tackles the problem of shared communication and serialization, automatic configuration and content distribution, and software orchestration in a purely distributed environment where HIL systems integration (e.g. car ECUs, nVidia DRIVE AGX) is fundamental. In addition, the procedural scenario generation pipeline is discussed.

Thermal systems integrated inside new electric vehicles are more and more complex. The main reason is that these thermal systems are no longer dedicated to one unique function or one vehicle performance like in the past. 10 years ago, the refrigerant loop in the vehicle had only one purpose: cooling the cabin for thermal comfort performance. The water circuit was also mainly designed to cool the engine, with a connection to the HVAC unit through the heater-core to heat the cabin. That’s all. But today, thermal management of innovative EV and HEV car models is designed with much more challenges to address. The refrigerant loop becomes a Heat Pump, with many evaporation branches, to cool the cabin ranks but also to cool the battery, water loops are many and interconnected with 3 or 4-way-valves, some new exchangers are created in order to optimize the use of thermal energy and limit the wastes. Fuel Cells are developed again and need also to be cooled, with specific requirements. Thermal Management becomes complex and more broadly, Global Energy Management of the vehicle becomes a key to offer a better autonomy but also safety and comfort to the customer of these new “smart vehicles”. To face this increase of complexity, engineers need new methodologies and new tools. MBSE (Model Based System Engineering) is one of them and ensure, as long as it is applied correctly, a robust design of your innovative functions and new performances. And in this context, Model Based Design is more essential than ever. Essential to design your thermal physical architecture, to size your components and to design the associated control software, keeping an eye on the various performances that the engineer has to achieve at the vehicle level (comfort, autonomy, speed, …). The aim of the paper and the speech will be to explain the new structured methodologies guided by MBD and MBSE, necessary to design with a high level of robustness and efficiency the Energy Management and the Thermal Management of Electric Vehicles.

High-performance computing (HPC) was once the sole domain of big car OEMs like BMW, FCA, and VW. Nowadays also small and medium enterprises make use of HPC. First, hardware has become better affordable, more powerful and better accessible. This is enabled by more performant servers and clusters both on premises and in the cloud.
Secondly, the software, and more in particular the engineering simulation software has significantly improved in terms of applicability, user-friendliness, robustness and performance.
This presentation will address the advances made in the area of applicability and performance for automotive applications.

From concept to engineering, and from design to test and manufacturing, engineers from wide ranges of industries face ever increasing needs for complex, realistic models to analyze the most challenging industrial problems; Finite Element Analysis is performed in an effort to secure quality and speed up the development process. Powerful virtual development software is developed to tackle these needs for the finite element-based Computational Fluid Dynamics (CFD) simulations with superior robustness, speed, and accuracy. Those simulations are designed to carry out on large-scale computational High-Performance Computing (HPC) systems effectively.
The latest revolution in HPC is the effort around the co-design approach, a collaborative effort to reach Exascale performance by taking a holistic system-level approach to fundamental performance improvements, is In-Network Computing. The CPU-centric approach has reached the limits of its scalability in several aspects, and In-Network Computing acting as “distributed co-processor” can handle and accelerates performance of various data algorithms, such as reductions and more.
HPC-AI Advisory Council performed deep investigations on a few popular CFD software to evaluate its performance and scaling capabilities when using HDR InfiniBand interconnect. The study reveals the influence of the applications on runtime, scalability and performance of the simulations.

ELDOR, ANSYS and SAP have successfully completed a pilot project in which a digital twin based on engineering simulation has been implemented for an electric motor. Using ANSYS Twin Builder technology, the original 3D simulation model has been first compressed and then integrated into the SAP IoT platform, to finally be connected to the real asset. ELDOR’s vision is to adopt simulation-based digital twin concept into the development and testing phase to shorten the design cycle for their e-motors.

Automotive engine cooling systems doubles as heat sources to provide passenger thermal comfort in cold environments.
In order to meet customer expectations the satisfactory and reliable, classical design introduced a simple, mechanically driven thermostatic valve to prevent engine coolant circulation inside main radiator and heat losses in the environment. More demanding clear air bills and need to improve fuel economy imposes optimization of every vehicle equipment: this paper explores current available 1d CFD simulation and testing tecniques to investigate engine coolant loop valves effects for thermal and hydraulic purposes.

Metal Additive Manufacturing (AM) is widely used in Formula One, Motorsport and Racing to manufacture complex parts in a short time. Power Bed Fusion (PBF) technologies, such as Selective Laser Melting and Direct Metal Laser Sintering, are currently applied to manufacture parts in aluminium, titanium, inconel and other highperformance superalloys (e.g. exhaust system, aerodynamics inserts and wings, pipes, roll hoop, etc). The maximum freedom in the designing phase, the opportunity to manufacture lightweight parts, complex geometries, even applying lattice structures at variable controlled densities, are the main success factors of metal AM growth in Motorsport. Nevertheless, metal AM is not synonymous of perfection, it has its own limits and constraints. One critical issue of metal AM consists of distortions which occur to the part during the laser melting process. In particular, it happens with thin wall titanium components, which are frequently deviating from the nominal 3D CAD geometry, despite stress-relieving treatments. The application of simulation tools to limit and compensate the distortions could dramatically reduce the risk of scraps, delays in delivery and the related costs. This paper concerns the AM process of an aerodynamic insert by RENAULT F1 Team in titanium Ti6Al4V: production has been optimized identifying the best part orientation and support structures positioning into the melting chamber, as well as using the Additive Print module by ANSYS, a simulation software useful to predict the distortion of the part and develop a new compensated 3D model which assures the best “as-built” quality.

New standards for the reduction of vehicles emission have been recently applied even to vehicle when in off conditions. Main actors of this new task for the emission reduction were at first USA and China. Later on, even Europe has started to focus attention on this topic.
The standard is based on the necessity to avoid emissions even when the vehicle is parked. The topic of this article is the EVAP (Electric VApor Pump), that is an high speed centrifugal pump used to send the gasoline vapors from the gasoline tank to the intake manifold, in order to avoid the vapor escape during the oil station tank filling as the cap tank is opened.
Vapors will be burned later on during the engine combustion.
The first generation of the EVAP has been developed by Pierburg and it was a completely new product for the company. A compact design with high rotational speed.
During the development of the product, modifications were done in order to get longer the durability tests.
A new matter was introduced in the company, the Rotordynamics, as a powerful help to the development of the design and to the management of modifications.
The numerical and diagnostic tools developed during the project, will be presented in this paper.

Li-ion battery aging is driven by a number of factors such like battery technology, discharge and recharge cycle characteristics as well as the control strategies and the capability to thermally condition the battery pack. The large number of interconnected parameters and their dynamic nature as well as the number of cycles required to age a battery cell makes physical testing of battery systems a time consuming and costly task, especially when scaling from single cells to full battery packs. This paper proposes a method of linking electrochemical Li-ion models of battery systems and physics based aging models, with multi-domain (electrical, mechanical, thermal, and flow domains) system-level models at minor additional computational cost, allowing for virtual testing of large parameter sets under realistic boundary conditions. The use of battery models that simulate electrochemical mechanisms empowers engineers with the ability to predict the performance and aging of the cell beyond the conditions at which the battery has been tested and to extrapolate the results to full battery packs used in the electric powertrain. It allows to take critical decisions on battery technology, cooling concept or control algorithm selection faster in the design process and helps to reduce battery testing and characterization effort without compromising results accuracy.

The purpose of this work is to carry out a comparative analysis of two different approaches related to engineering: a more conventional one, based on computation, against a natural one. The idea is to look at them taking advantage of several point of views and application fields and along the design chain (design & optimization, manufacturing processes, quality assessment, use of light materials), in order to achieve a super engineering model in the end.

Nowadays small hybrid, BEV or simply gasoline vehicle are becoming demanding in terms of efficiency, modularity, weight, costs and appeal. We are asked to trade-off all these performances in the best manner to meet customer expectation. Optimization of engine coolant module in terms of efficiency , weight, fuel saving is a key factor to achieve competitiveness.

Structural component commonality occurs when one component is common to more than one structure. Component commonality is desirable for mass production. In particular, it is highly desired in car body designs due to cost saving. A common design require less tooling, less replacing inventory and provides simplicity. Conversely, having the ability to do component commonality opens the door for more economic customizations, especially when customization is used for smaller parts and larger parts are kept common. In this paper, we discuss the use of multi-model optimization (MMO) to achieve optimal component commonality. MMO allows multiple optimization models to run simultaneously via a master optimization process. The individual sub-models may share some or all of their design variables with other sub-models. Each sub-model may include its own private constraints and objectives. Using multi-model optimization, large-scale optimization problems may be solved on HPC clusters, with each sub model analysis solver running on a separate compute node. The primary benefits of multi-model optimization are: a) the ability to optimize large-scale models where designable regions are common to several model assemblies, but remaining parts of the models differ. For example, optimization of components common to several configurations of a vehicle, such as sedan and convertible versions and b) Reducing the wall clock time of large scale problems by splitting the solution of different boundary conditions onto separate compute nodes. To demonstrate the use of MMO a body-in-white automotive model with multiple load cases is considered for the analysis. In one model, the automotive body is considered with the full roof structure and in the second model, the same automotive body is considered with a sunroof. The top of each car is designed independently but the floor of both cars is design for commonality. The main example is solved using topology optimization, however, other types of optimization are possible. The problem is solved using the GENESIS structural optimization software.

Real objects need to be introduced into the CAE world for quality control, design evaluation, FE simulations and other. Computed Tomography (CT) provides a powerful method for achieving this in a nondestructive way even for complex assemblies. In this work, a CT scan of a shock absorber is virtually disassembled with the aid of RETOMO in order to see its anatomy. Innovative machine learning algorithms introduced in RETOMO greatly reduce required user time while improving segmentation quality. The complete process from CT data to CAE model is demonstrated through a realistic reverse engineering scenario. Eventually the seamless interaction of the BETA suite is exemplified through the creation of a CAE model for a durability load case.

For Advanced Driver Assistance Systems (ADAS) and Autonomous Vehicules (AV) there is a crucial need for the application of reliability methods to calculate with the help of scenario simulations in an efficient way the probabilities of failures (PoF). Usually the probability of a crash / km is derived by the PoF analysis of different logical scenarios and building a sum of the PoF's from the logical scenarios. Also it is important to take into account the effect of steering mechanism from electronic control units and how this translates into the quality of metamodels, that may further be used for the analysis. We describe appropriate workflows and reliability methods and their applications in different projects.

The design of a specialty tractor is nowadays far more complicated than in the past: more restrictive and conflicting needs from end-users, marketing and homologations should all be packed in a sleek compact design, deteriorating in particular heat management performance.
With this in mind, CFD simulation become an important tool in design process to foresee and optimize cooling performance before prototype production and testing.
Object of the present study is a new project of a specialty narrow tractor, available in four combination: in vineyard or orchard configuration, both with Landini or McCormick brand. The goal of the analysis is to evaluate the flow rate distribution in the hood openings, the flow over cooling packs and their heat dissipation and the underhood flow in general, in order to enhance overall cooling performance as well as operator comfort and safety.

A CAE simulation was performed to assess the dynamic behavior and the structural performance of a driving mechanism of a two-wheeler. MBS, FEM and durability tools were used.

The aim of this study is to analyse the performances of centrifugal water pumps comparing the results obtained with different CFD codes (both commercial and opensource) with experimental data.
The simulations have been performed employing both steady and transient solvers after carrying out a grid independence study.
The results obtained by this work highlight the cons and pros of the different methods and allow to forecast with more accuracy the behavior of centrifugal pumps by CFD simulations.

A simulation process for the calculation of vibrational solder joint fatigue in FEMFAT spectral is presented. The process consists of three core technologies: semi-automatic generation of substitute FE-models for electronics components, automatic placement of these FE-models on FE-models of printed circuit boards (PCBs) and automatic generation of solder joint submodels, for which the geometry is calculated based on pin dimensions of the electronics component, the pad dimensions on the PCB, solder volume, surface tension and gravitation. With a single frequency response analysis (FRA) applied on an FE-model of an electronic device equipped with our substitute FE-models, and static analyses of all solder joint submodels, the vibrational solder joint fatigue can be efficiently calculated in FEMFAT spectral, where the section forces in the pins of the electronics components are mapped onto the submodels and scaled by a power spectral density to calculate the damage in all solder joints.

The manufacturing industry market has been recently invested by a huge technological development. New technologies have been introduced, starting from the design till the final creation of the object, making the whole process more fast and simple. In this framework, being competitive requires to be fast in leading clear and reliable solutions to the clients.
In this study, modern CAE tools are used to perform the virtual restyling of a motorbike. The whole design process consists in multiple steps, namely the acquisition of the existent geometry by means of 3D scanning, the definition of the mathematical model and finally the CFD simulation, validated with experimental data from wind tunnel. This approach allows to quickly verify the influence of geometrical modifications on global performance of the motorbike.
The geometries and CFD results are then combined in an interactive VR environment. This is a unique and innovative way of presenting, that allows the client to virtually visualize the final product in its real dimensions, and to interactively check the performance derived from geometrical variations.

In this paper a numerical Finite Element methodology is proposed, which aims at evaluating the cylinder liner bore distortion in an eight-cylinder V-type four stroke turbocharged engine. In particular, preliminary Finite Element models are developed and properly tuned in order to obtain the same cylinder liner distortion registered by experimental measurements, where a single engine bank is coupled with a simplified test engine head. Further Finite Element analyses, both thermal and thermo-mechanical, are then performed to evaluate the cylinder liner distortion considering the actual engine head. In order to speed up the analyses, the engine head, the gasket, and the bolt tightening are substituted by pressure distributions mimicking the actual contact interactions. The methodology reveals itself to be well correlated with the experimental evidences and with the complete Finite Element model of the engine bank thus consisting in a useful tool for the virtual component approval.

We would like to present a technology that we started developing about 15 years ago, and that radically changes the perspective and the expectations of CFD users about what can be simulated and in how much time it can be done. We believe it is a paradigm shift and that is the reason why we call it CFD 2.0.
In order to let the audience understand what we mean by paradigm shift, we will show applications to a few 3D models chosen for their extreme geometrical complexity. The models, of interest in the Automotive industry, include very many detailed representations of countless objects whose scales start from few millimeters.
The technical recipe of our CFD 2.0 is made of our latest version of the Immersed Boundary technique and an automatic Cartesian Grid Generator. This technology allows to decrease the user time needed in the preprocessing phase from hours and days to less than 10 minutes, no matter the complexity of the 3D models, with no need to clean the subgrid objects, reaching amazing and unbelievable levels of robustness and accuracy.