Skip to main content

2024 | Buch

2024 Stuttgart International Symposium on Automotive and Engine Technology

Teil 1

herausgegeben von: André Casal Kulzer, Hans-Christian Reuss, Andreas Wagner

Verlag: Springer Fachmedien Wiesbaden

Buchreihe : Proceedings

insite
SUCHEN

Über dieses Buch

In einer sich rasant verändernden Welt sieht sich die Automobilindustrie fast täglich mit neuen Herausforderungen konfrontiert: Der problematischer werdende Ruf des Dieselmotors, verunsicherte Verbraucher durch die in der Berichterstattung vermischte Thematik der Stickoxid- und Feinstaubemissionen, zunehmende Konkurrenz bei Elektroantrieben durch neue Wettbewerber, die immer schwieriger werdende öffentlichkeitswirksame Darstellung, dass ein großer Unterschied zwischen Prototypen, Kleinserien und einer wirklichen Großserienproduktion besteht. Dazu kommen noch die Fragen, wann die mit viel finanziellem Einsatz entwickelten alternativen Antriebsformen tatsächlich einen Return of Invest erbringen, wer die notwendige Ladeinfrastruktur für eine Massenmarkttauglichkeit der Elektromobilität bauen und finanzieren wird und wie sich das alles auf die Arbeitsplätze auswirken wird. Für die Automobilindustrie ist es jetzt wichtiger denn je, sich den Herausforderungen aktiv zu stellen und innovative Lösungen unter Beibehaltung des hohen Qualitätsanspruchs der OEMs in Serie zu bringen. Die Hauptthemen sind hierbei, die Elektromobilität mit höheren Energiedichten und niedrigeren Kosten der Batterien voranzutreiben und eine wirklich ausreichende standardisierte und zukunftssichere Ladeinfrastruktur darzustellen, aber auch den Entwicklungspfad zum schadstofffreien und CO2-neutralen Verbrennungsmotor konsequent weiter zu gehen. Auch das automatisierte Fahren kann hier hilfreich sein, weil das Fahrzeugverhalten dann – im wahrsten Sinne des Wortes - kalkulierbarer wird. Dabei ist es für die etablierten Automobilhersteller strukturell nicht immer einfach, mit der rasanten Veränderungsgeschwindigkeit mitzuhalten. Hier haben Start-ups einen großen Vorteil: Ihre Organisationsstruktur erlaubt es, frische, unkonventionelle Ideen zügig umzusetzen und sehr flexibel zu reagieren. Schon heute werden Start-ups gezielt gefördert, um neue Lösungen im Bereich von Komfort, Sicherheit, Effizienz undneuen Kundenschnittstellen zu finden. Neue Lösungsansätze, gepaart mit Investitionskraft und Erfahrungen, bieten neue Chancen auf dem Weg der Elektromobilität, der Zukunft des Verbrennungsmotors und ganz allgemein für das Auto der Zukunft.

Inhaltsverzeichnis

Frontmatter

Aerodynamics

Frontmatter
Macan Goes BEV. Aerodynamics of the First All-Electric Porsche SUV
Abstract
How does one take the Porsche model that has been the top-seller six times since 2015 and transition it into the e-mobility era? How does one combine its classic Porsche themes – its quintessential design language and track suitability – with modern customer demands like range and consumption? How does one reduce drag by 30% while retaining the identity of the vehicle? Or to frame the question differently: what were the aerodynamic challenges of the new Porsche Macan, the brand’s first all-electric SUV with the least drag?
The aerodynamics of the electric Macan were designed using state-of-the-art tools. The basic shape of the vehicle was defined and optimised in a scale model wind tunnel. In a state-of-the-art, full-scale aeroacoustic wind tunnel, further refinements to the vehicle shell and other technical details were made right down to the start of production in collaboration with the designers.
Simultaneous to the measurements in the wind tunnels, the entire aerodynamic development process was supported by detailed flow simulations (CFD).
The aerodynamic improvements and technical measures implemented reduce the drag coefficient by delta cD = −0.10 from the previous value of cD = 0.35 to cD = 0.25. This corresponds to an increase in range of around 85 km in the standardised fuel consumption cycle for Europe (WLTP).
Development of the basic form focused on the roofline, which drops off coupé-like by way of the rear window. The rear section is capped by the active rear spoiler which, in combination with the flaps of the fully closable air intake in the front of the vehicle, makes it possible to have both efficient, range-orientated travel as well as sporty, performance-focused driving.
Other major contributors to reducing drag include the aerodynamically optimised wheels and tyres and the largely closed underbody with its flexible suspension fairings.
Andreas Burk
The New Porsche Taycan: Augmented Range or Enhanced Performance? The Choice is Yours
Abstract
The Taycan, Porsche’s first fully electric sedan, in production since 2019, has undoubtedly been a huge success.
The first generation of the car was produced in five versions, Base, 4S, GTS, Turbo and Turbo S comprising two shapes, Sport Sedan and Sport Turismo respectively Cross Turismo.
Aerodynamically, the Taycan Turbo Sport Sedan achieved the best performance, starting with a drag value of 0.22 [1], and hence best in class for 2019. Concurrently the Taycan Base had a drag value of 0.24, the Taycan 4S of 0.23 and the Taycan Turbo S of 0.26.
Thanks to the new Porsche Wheel Strategy and our acquired know-how on rims and tyres, the all new Taycan, which was launched on the market at the beginning of 2024, can achieve a minimum drag value of 0.22 in all Sport Sedan versions (Base, 4S, Turbo and Turbo S). The spread in the range has thus been extremely reduced between all wheel configurations.
Concurrently performance remains a crucial, strong selling point for Porsche. The achieved improvements are not therefore only limited to the wheels.
Using the new Taycan Turbo S as the baseline, itself boasting a vmax of 260 km/h and the following areo-values:
cD = 0.22; cLF = cLR = 0.10 in range mode
(Aero-Wheels, Active Grille Shutters closed and Rear Spoiler in Eco Mode)
Porsche developed on a new derivative, by combing a more powerful battery, a new rear electric motor, a new pulse inverter, downforce optimised extensions, lightweight wheels and a PCCB, which enabled the Taycan Turbo S to achieve GT Car Level performance.
The Taycan Turbo GT version will thus achieve a vmax of up to 305 km/h, and downforce coefficients of up to 0.07 on the front and 0.13 on the rear axle.
Porsche has given the Taycan a particularly extensive update.
The new versions have more power, more range, accelerate more quickly and charge faster with greater stability.
The new Taycan is now on the market and will allow the true driving enthusiasts the joy of a greater range combined the excitement of an enhanced performance package!
Francesca Cogotti
Impact of Wheel Drive Unit Secondary Flows on the Aerodynamics of Passenger Cars
Zusammenfassung
State of the art wind tunnels use single belt or five-belt systems, the latter consisting of four wheel drive units (WDU) and a large center belt, to simulate the moving ground below the car [1]. Small gaps between the stationary floor and the WDUs are required to measure the forces acting upon them. Through these gaps a pressure induced secondary flow develops between the plenum and the space below, which interacts with the vortex structures around the wheels and tires [2]. Design related leakage from the air bearings of the WDU belt causes additional secondary flows which further differentiate the measuring conditions in the wind tunnel from driving on the street. These effects and their interaction with the flow-field around the car are investigated in this paper. For this purpose, unsteady computational fluid dynamics (CFD) simulations of a full scale DrivAer model and a detailed wind tunnel including a five-belt system were conducted. The computations were carried out using the commercial Lattice-Boltzmann software PowerFLOW. Based on these investigations a simplified simulation model was created. In addition, results of full scale wind tunnel measurements using particle image velocimetry (PIV) are used as a reference to analyze and review the results from the CFD simulations.
Dennis Möllenbeck, Axel Fischer, Hardy Schmidt

Automotive Electronics

Frontmatter
OpTC – A Toolchain for Deployment of Neural Networks on AURIX TC3xx Microcontrollers
Zusammenfassung
The AURIX 2xx and 3xx families of TriCore microcontrollers are widely used in the automotive industry and, recently, also in applications that involve machine learning tasks. Yet, these applications are mainly engineered manually, and only little tool support exists for bringing neural networks to TriCore microcontrollers. Thus, we propose OpTC, an end-to-end toolchain for automatic compression, conversion, code generation, and deployment of neural networks on TC3xx microcontrollers. OpTC supports various types of neural networks and provides compression using layer-wise pruning based on sensitivity analysis for a given neural network. The flexibility in supporting different types of neural networks, such as multi-layer perceptrons (MLP), convolutional neural networks (CNN), and recurrent neural networks (RNN), is shown in case studies for a TC387 microcontroller. Automotive applications for predicting the temperature in electric motors and detecting anomalies are thereby used to demonstrate the effectiveness and the wide range of applications supported by OpTC.
Christian Heidorn, Frank Hannig, Dominik Riedelbauch, Christoph Strohmeyer, Jürgen Teich

Autonomous Driving

Frontmatter
Towards a Data-Based Interface Definition to Support a Modular Safety Approval of Highly Automated Vehicles
Zusammenfassung
The safety validation has proven to be one of the most obstructive challenges in the pursuit of highly automated driving. Exhaustive field tests have been deemed infeasible and novel approaches such as scenario-based testing yet require to be proven viable. The challenge becomes even more serious when considering changing architectures due to learning software components and over-the-air updates. A modular approach to safety approval, focusing on assuring the safe operation of individual modules in their respective environments, promises to reduce the overall effort. In particular, it provides an argument for preserving the approval for future updates and upgrades, reducing the need for intensive retesting of the whole system. However, established knowledge-based methods for decomposition, specification, risk analysis and module test case generation struggle to argue completeness. Data-based methods used in other applications provide the opportunity to support this issue. Inspired by the design-by-contract paradigm, we combine selected methods into a framework to feed a data-driven interface definition to support modular safety approval. The framework is practically applied in a simulation environment to a highly automated vehicle with a disruptive modular architecture. A reduced ODD setting serves as a proof of concept and provides insights on the limitations and applicability of the applied methods with respect to the use cases of a modular safety approval.
Alexander Blödel, Björn Klamann, Steven Peters
Occurrence Estimation for the Classification and Prioritization of Concrete Scenarios in the Context of Virtual Scenario-Based Validation of Vehicles
Zusammenfassung
This article deals with the creation and concretization of a scenario catalogue in the context of scenario-based validation of highly automated driving functions. Due to the possible combinatorics of the various driving situations and influencing parameters in the open-world context, it is not possible to simply execute all possible combinations. Instead, new methods are needed that can make a representative statement about the respective logical scenario, but at the same time attempt to reduce the degree of execution. This article aims to examine the usefulness of using (top-view) recordings of real world traffic to extract the occurrence of appearance of certain, predetermined scenario parameters of a logical scenario. This information on the occurrence is used to prioritize simulation efforts for more efficient and better evaluation of driving functions in the context of autonomous driving. To this end, a methodology for exposure estimation is to be developed as part of this work. This approach uses analogies to existing methodologies (e.g. Automotive Safety Integrity Level (ASIL) defined in ISO 26262 [1]) and transfers them to the scenario-based context to attach an exposure level to each concrete scenario. An existing public dataset of top-view camera recordings of real traffic scenarios was analysed and evaluated with respect to pre-defined scenario-specific parameters. Finally, the results will be evaluated and validated with an exemplary logical scenario.
Julian Fuchs, Lennart Ries, Eric Sax
Synthetic Object Placement with Statistical Representations Regarding Real Data Sets
Abstract
Collecting automotive sensor data for AI-applications is time and cost expensive. One solution is the usage of simulated data. However, the quality of the applications results highly depends on how close the simulated scenery is to reality. One main challenge is the definition of the properties of dynamic objects. A simulation of their behaviour over time is possible, but for this, extra software e.g. SUMO or CARLA is needed. This paper therefore gives a proposal for defining those properties without further simulation software. The proposal uses statistical representation from real automotive data sets. This gives the advantage of direct object placements. In addition the objects properties are as near as possible to reality. Another advantage is the creation of more data frames as a data set can provide. First, several parameters are collected from real data sets. Then the parameters will be analysed regarding their statistical appearance. This analysis is then the basis for mathematical representations. Further metadata will be collected separately. Here, standards as openDrive are important tools. For validation, the objects are placed in a simulation environment (CARLA). For the Definition of Experiment (DoE) here, sampling methods as Latin Hypercube Sampling are fed with the previously determined analysis. In the end, the synthetic data is validated.
Lukas Lang, Hans-Christian Reuss
Rapid Assessment of EV Battery Without Driving the Car
Abstract
The high-voltage (HV) battery of an electric vehicle (EV) is the most expensive component in a vehicle. These should generally be designed for the lifespan of the vehicle, but mobile use in road vehicles is a particular challenge. In addition to the environmental conditions, the remaining usable capacity depends very much on user behavior. The batteries are therefore subject to a certain amount of wear and tear. For users, this is associated with restrictions in range and for owners – be it the original owner or the used car buyer – this represents a financial risk. Once the vehicle reaches a certain age, repairing or replacing the battery is no longer economical. The condition of the battery in terms of the actual remaining capacity has a significant influence on the used car market. This information is very important for both sellers and buyers. In recent years, various battery test procedures have been introduced in order to be able to use them for assessing the vehicle HV battery. DEKRA has developed and launched an assessment system with a special feature: a quick assessment. It is therefore tailored to the specific use case of used vehicle valuation. The challenge of an independent evaluation is sufficient powerful electric stimulation of the battery, which is implemented by a test drive in the form of an acceleration. However, for technical, organizational, or infrastructural reasons an acceleration drive is not in every situation possible. Accordingly, alternative methods are being investigated – instead of stimulation by discharging by means of a short test drive, stimulation by charging was also investigated. Two basic charging approaches – DC and AC charging – integrated into a new M2M architecture and the results are presented according to their accuracy.
A. Richter, S. Tilgner, T. Ost, C. Nolte

Battery

Frontmatter
Efficient Parametrization of Electrochemical LIB Models
Abstract
Lithium-ion batteries (LIB) play a key role in electrified mobility. Different vehicle classes, propulsion architectures and duty cycles call for tailor-made designs of the battery pack. Model-based frontloading contributes to tackling the challenges of selecting and optimizing design options early in the development phase. Electrochemical models are an established approach to model LIB. They describe the cell behavior based on balancing laws for mass and charge considering transport processes in the electrolyte and solid electrode structure. One key aspect of such models is their efficient parameterization. An advanced parametrization workflow supported by AVL CRUISE™ M is introduced which ensures data consistency and appropriate interpretation of experimental data. This is achieved by parameter classification and relating measurable quantities to relevant model parameters. The workflow minimizes the required measurement data. The major benefit of this approach is the separation of fixed parameters describing basic properties and adjustable parameters which improve the voltage response to dynamic cell operations. Focusing on adjustable parameters enables usage of optimization tools in a subsequent step while retaining minimal uncertainty in the full parameter set. The workflow is discussed based on selected examples of state-of-the-art cell assemblies and electrode materials. Furthermore, an outlook is given on the applicability of this workflow to next generation materials.
Christoph Lechner, Susanne Kutschi, Johann C. Wurzenberger

Charging & Infrastructure

Frontmatter
Fuzzy Rule-Based Coordinated EV Charging Management
Abstract
Coordinated electric vehicle (EV) charging is a concept with a growing potential. The main goal of this work is to manage the overall charging power of a group of EV’s during charging according to the available system capacity. This concept is increasingly needed for charging stations that are located at commercial or residential buildings. These buildings are observed to have large load peaks during certain times of the day.
In a previous study, a rule-based energy management method was developed. The later, allows coordinated charging according to predefined conditional rules and results in reduced load peaks. For this method, the EV’s are categorized according to specific clusters using unsupervised machine learning (ML).
In this work, the rule-based energy management system, is further developed and designed using fuzzy inference system. This approach allows coordinated energy management based on rules that interpret human reasoning and decisions. The implementation of fuzzy logic is beneficial as it introduces a great amount of flexibility and scalability to the system. In addition to the fuzzy inference system, coordinated charging has been implemented in this work based on mathematical optimization and using genetic algorithm. Then, the performance of the two aforementioned approaches are compared with each other in simulation and utilizing the same test scenarios that were implemented in the previous work.
The fuzzy inference system has the most favorable performance among all three concepts. As it was successful in providing, coordinated EV charging, flatten the load peak and provide EV demands according to their assigned cluster specifications.
Alia Salah, Omar Abu Mohareb, Frank Brosi, Hans-Christian Reuss
Energy and Automotive—Do We Have Enough Sufficient electrical energy
Abstract
The question “do we have sufficient electrical energy for charging electric vehicles” is heavily discussed in the public. We are changing for renewable energies and in addition the forecasts predict a tremendous increase for demand of energy. The energy transition is therefore a huge challenge. Studies are available to provide figures for an adequate answer for this key question. According to these studies e-Mobility will have 2045 about 17% share to the energy demand. Furthermore electric vehicles are capable to store electrical energy and feedback energy into the grid in phases of low energy availability. This use case will support energy transition as a win-win option. To support energy transition users also have to adapt their behavior consuming energy. If energy transition can be successfully performed, sufficient energy for e-mobility seems feasible.
Ursel Willrett

Chassis Systems

Frontmatter
Brake Feel—Objective Assessment and Potential for Virtual Development
Abstract
The brake feel is of great importance as it conveys information to the driver about the deceleration process in general as well as the brake performance in particular. In order to have a common basis for developing brake feel, it must first be defined with objective Key Performance Indicators. The paper presents a methodological approach for an objective assessment of the brake feel which enables virtual development. By analyzing the deceleration data of different vehicles in different braking situations, such as braking straight with varying decelerations, a foundation for further steps is established. This basis includes common diagrams such as the force-travel characteristic of brake pedals. In the next step, this data is compared with recent literature to identify a good brake feel to set target values. There is some variation in the data as the vehicles are of different vehicle classes as well as the varying pedal path. It can be observed that the compared vehicles have variable brake feel and the KPI emphasize this variation. Based on these results, an objective evaluation of brake feel can be performed. With an objective assessment of brake feel, brake systems can be tested during early development stages in combination with Software-in-the-Loop and Hardware-in-the-Loop setups, which significantly reduces development and testing flexibility. This sets the basis for a virtual development of brake systems.
Raphael Groß, Anton Tworek, Peter Pfeffer

Circularity & LCA

Frontmatter
Managing the Complexity of Circularity in Automotive Design
Abstract
Circular economy is one of the biggest challenges for the automotive industry. It promotes resource efficiency, reduces waste, and encourages the reuse and recycling of materials, leading to a more sustainable approach of mobility. However, designing an automotive system that incorporates circular economy aspects is a highly intricate task. Its complexity stems from various aspects, such as the interdependent solution space, different possibilities how material cycles can be closed and the challenge to evaluate the impact of a circular design on a multi-faceted automotive system. In this paper, we present an approach to incorporate circular economy aspects in the early stages of product development. Five development artefacts, a methodology and a MBSE software system help to effectively implement sustainability measures and thus reduce circular-related complexity. This enables the automotive developer to perform life cycle assessments in early stages of development, evaluate circularity indicators and trace sustainability throughout the vehicle’s life cycle, even if the automotive system is not yet fully defined. We present the approach and the software system, including its evaluation on the design of an automotive center console.
Lukas Block, Antonino Ardilio, Lukas Keicher
An Analysis of the Greenhouse Gas Potential of Current Powertrain Technologies
Abstract
Long-term studies have identified the increase in global CO2 concentrations as the main reason for the rise in global mean temperature. This leads to the frequent occurrence of extreme weather phenomena, rising sea levels and poses a threat to ecosystems. To reduce these impacts, efforts must be made to keep the CO2 concentration in the atmosphere constant or even reduce it. A first step in this direction is the reduction of fossil CO2 emissions, which are firmly anchored in our production, energy and transport systems. These fossil CO2 emissions are generated during production, operation and final recycling of products and processes. With the help of life cycle analysis, these CO2 emissions of products and processes are made visible and thus tangible and assessable. In the context of the transport sector, there is a diversification of powertrain technologies, some of which have very low or no tank-to-wheel emissions. However, this analysis cannot be considered as complete, as both the production of the energy or the energy carrier and the production of the vehicles are disregarded. An assessment of the total emissions (cradle-to-grave) of a vehicle over its life cycle provides a holistic approach which is essential for an appropriate evaluation of these products. This paper presents life cycle models for various current powertrain technologies. To enable an appropriate comparison of these technologies, three representative vehicle configurations are evaluated and the carbon footprint for the production (cradle-to-gate) of these vehicles is compared. This paper was first published at the 45th International Vienna Motor Symposium.
Tobias Stoll, Hans-Jürgen Berner, André Casal Kulzer

Connected Vehicle

Frontmatter
Accelerated Time-to-Market for Powertrain Development and Fast System Verification incl. OTA-Updates
Abstract
A key role at AVL for development, integration, calibration/applications & verification of the whole vehicle respectively the main systems (powertrain & energy storage systems), plays the highly flexible powertrain and vehicle in the loop (PaViL) test system. Besides the strong trend to virtual testing, there are still lots of development & testing tasks which require the actual HW system. The parallel use of virtual methods on SiL & HiL in combination with PaViL can increase the total development efficiency significantly. The paper describes a specific example, how to use this kind of special powertrain test sys-tem by multidiscipline development teams (i.e. energy consumption, efficiency, emission, feature testing incl. functional safety, diagnoses), with the same ONE UUT (unit under test) to perform the different development steps in an optimized & efficient way. The joined planning of test sequences and a high degree of automatization of performing tests & data processing allows the usage of the UUT & of the test environment almost 24/7. This reduces the need for prototype vehicles and on top of it, the whole development time is minimized to support short “time to market”.
DI Wolfgang Weisser, Peter Ebner, DI Stefan Eder

Cyber Security

Frontmatter
A New Perspective in Automotive Cybersecurity Principles and Practices of Security Chain Strategy
Abstract
This paper introduces a Security Chain strategy for enhancing cybersecurity in Vehicle Control Electronic Control Units (ECU), pivotal in today’s automotive industry. With escalating cyber threats and stringent global regulations like UNECE R155 and China’s automotive cybersecurity standards, Vehicle Control ECUs face significant security challenges. This paper addresses the conflict between the necessity for robust cybersecurity measures and their associated high costs and complexity. The proposed Security Chain strategy offers a holistic approach to manage cybersecurity risks effectively and economically.
Key components of this strategy include:
1. Theoretical Foundation of Security Chain: Establishing a comprehensive framework for identifying and managing cybersecurity risks in ECUs
2. Identification of Key Security Features: Analyzing and categorizing crucial cybersecurity features in Vehicle Control ECUs.
3. Integration with Existing Standards: Harmonizing the Security Chain strategy with current cybersecurity measures, national laws, and OEM specifications.
The solution not only addresses current cybersecurity challenges but also provides a new perspective for thinking about automotive cybersecurity. This approach ensures Vehicle Control ECUs are resilient against diverse cyber threats while adhering to industry standards and regulations.
Huiru Ma, Bernd Reh
Zonal Architecture and HW-Driven Protections for Software-Defined Vehicles
Abstract
In this paper we will discuss the crucial role of Zonal Architecture and HW-defined protections, a cornerstone of the Software-Defined Vehicles revolution.
The security paradigm transitions from Domain Architecture to Zonal Architecture will be introduced, emerging threats in virtualized environments will be presented and necessity of reducing attack surface will be discussed. Furthermore, cutting-edge hardware methods to tackle cyber threats will be explored. Finally, enhanced techniques to reinforce microcontroller security in Zonal Architecture will be suggested.
Lara N. Popova, Markus Maier
Automotive Cybersecurity: Level up Your Product Risk Management!
Abstract
2023 witnessed a surge in cybersecurity incidents within the automotive sector: major US automotive supplier falling victim to ransomware, Porsche Macan sales halting due to cybersecurity non-compliance, recurring hacker intrusions into infotainment systems at Tesla, Hyundai, Ford, and others. The trend continues into 2024, with a new focal point: the applicability of UN R155 and R156 to all vehicles, encompassing not only new models but also newly manufactured former models. Cybersecurity is evolving beyond a safety or operational concern for OEMs; it is now a compliance risk that demands attention.
One essential pillar of regulation consists of managing risks adequately, timely and efficiently.
Adequately, because the risks must be identified and assessed, including the complexity due to functions and components interactions.
Timely, the risks must be continuously re-evaluated before and after production, to follow software changes and to adapt to the threat landscape.
Efficiently, risks mitigated via security controls must be tracked down along the development process, to ensure good implementation and alignment between theoretical artefacts and real residual in-vehicle risks.
Unfortunately, many OEMs are not ready to deal with it. Poor methodologies associated with outdated tooling jeopardize the whole security chain. In this paper, we will discuss pitfalls, upcoming challenges for stakeholders in 2024 and how to build viable solutions for efficient vehicle security.
Théo Tamisier, Robin Huber

Data Science & AI

Frontmatter
Mobility Enabled by AI: How to Benefit from Recent Advancements in AI
Abstract
Recently, ChatGPT from OpenAI has become the most frequently named AI innovation. It demonstrates the power of AI. This paper is about 2 benefits from this for automotive: 1.) How can large language models such as ChatGPT be used in automotive processes and 2.) how can the advancements in the development of ChatGPT be used in other automotive use cases?
1.)
Large Language Models such as ChatGPT can be used for the enhancement of engineering processes. Automotive projects require a lot of resources. Currently many positions cannot be staffed, and this leads to incomplete or delayed processes and products. Examples will be presented how large language models can be beneficially used in problem decomposition, solution design, code generation and testing.
 
2.)
An analysis of the development methodology from ChatGPT reveals important learnings about successful AI development. The development approach of ChatGPT can be summarized as follows: Openness to new ideas, very large data for training and testing, combination of supposedly competing concepts, data crunching with expensive computing power, and broadband verification. These learnings can be well transferred to automotive use cases. If these learnings are followed, automotive can be even more successful with AI.
 
Recent advancements can boost the productivity of processes and/or facilitate new features and set the expectations by our customers. GenAI can become a game changer for the automotive industry.
Ulrich Bodenhausen
Accelerate Autonomous Vehicle (AD) Development with Data-driven Automotive AI
Abstract
Developing AVs is a time-intensive and complex process that requires best-in-class data and AI training infrastructure. Companies developing software-defined vehicles need to accelerate time to market and minimize costs without sacrificing safety. Combining vehicle sensors, map data, telematics, and navigation guidance using machine learning and data fusion techniques. Data-driven development is not without its challenges. One of the biggest challenges is data collection and integrity, as data needs to be collected accurately and consistently to drive accurate decisions.
Frank Kraemer
Scalable Measurement Data Acquisition for ADAS Development
Abstract
The article discusses the role of scalable measurement data acquisition in the development of Advanced Driver Assistance Systems (ADAS). The text highlights the transition from traditional human-driven vehicles to automated driving, facilitated by the integration of radar, lidar, video, and ultrasonic sensors. It emphasizes the need for powerful control units (ECUs) to process vast amounts of data and derive driving strategies. The development of software functions is shifted to labs, utilizing virtualization to replace experiments with real hardware, necessitating reliable tools for data acquisition and cloud or backend data access. ETAS offers a modular portfolio of scalable solutions for In-Vehicle Data Acquisition (DAQ) across all development phases, addressing the challenges of prototyping, development, and post-production validation.
Thomas Schöpfner, Patrick Nickel
OPTIMALIS® – An AI Framework Tailor-Made for the Needs of the Automotive Industry
Abstract
The future success of the automotive OEMs will closely be related to the adaption of artificial intelligence (AI) in all their domains, e.g., in the engineering process, in production, and in the software-defined vehicle itself. In more detail, fast AI-surrogate models can speed up time-consuming simulations in the engineering process by a huge factor – independent of the simulation domain, being it fluid, structural, thermal, or electrical. Hence, AI-surrogate models can be used to identify the optimal configuration, something that is often practically impossible using traditional simulations. Moreover, the automotive OEMs own huge datasets, originating for example from the development and testing, from production or from the fleet. To gain knowledge from these datasets can be of upmost importance. One key is the data-driven detection of unexpected data using AI, for example, for predictive maintenance or for AI-driven data correction. In addition, one can use data to predict its future evolution, e.g., to predict driving profiles or the power consumption of the vehicle electrical system. In this article, we introduce the AI framework OPTIMALIS®, demonstrate its broad range of application and highlight its intuitive usage. The key advantages of OPTIMALIS® are its standardized data-interface, its ability to automatically adopt to data, and its autonomy. As such, OPTIMALIS® is the ideal tool to test the usage of AI in a customer’s use case and to leverage their business to the next level.
Frank Beutenmüller, Victor Fäßler, Patrick K.S. Vaudrevange, Thomas Wolf

Development Methods

Frontmatter
A Digital Machine-Executable V-Model for a Formula Student Racing Car
Zusammenfassung
Based on a manual formalization step of the requirements, a digital machine-executable V-Model for a Formula Student racing car is shown to automatically generate a permissible design solution of a racing car which fulfills all given requirements. The V-Model used is reduced in its full generality by several shortcuts in form of some predetermined design solution patterns such as the choice of an electric propulsion in the form of 2 or 4 wheel hub motors, a concentrated battery pack and a fixed monocoque design. In the race car design, the focus is on the braking system which is designed and optimized for maximum system performance. The design procedure of the braking system is encoded in a sequence of machine executable model transformations to adapt the designated race car design in an optimal way to a specific race track.
Julian Borowski, Stephan Rudolph
Impact of Electrification and Digitalization on the Development Methodology and Testing Technology
Abstract
Because of electrification of vehicle drives and because of digitalization, three striking trends are emerging for the development at test beds: Firstly, the importance of the classic powertrain test beds is decreasing and two extended test bed types are emerging. On the one hand, components are tested on component test beds that are significantly upgraded in terms of importance as they allow automated model parameterization. On the other hand, the electro-mechanical, thermal and electronics/software subsystems are brought together on a so-called system integration test bed and tested together. Secondly, the extensive transfer of real test drives to the system integration test bed by implementing the three realistic boundary conditions (“power, thermal and information flow”). There is still significant potential for improvement in the future. This trend will result in a reduction of the test drives, whereby the degree will depend on how far it is possible to sufficiently implement the mentioned requirements. And thirdly, the use of artificial intelligence. In addition to the increased use of data analysis methods, due to the smaller volumes of data but physical knowledge in technical developments there will be a combination of artificial intelligence and prior information in the form of so called “hybrid intelligence”. However, the aforementioned changes require corresponding organizational measures and broader skills of systems engineering personnel in order to be able to implement them.
Stefan Geneder, Günter Hohenberg
Dynamic Offset- and Drift Compensation in Powertrain Test Benches
Abstract
One of the most important aspect of a powertrain test bench is to capture measurement data. To ensure the validity of the test procedures and the quality of the measurements it is essential to use regularly calibrated measurement systems. However, since these systems operate continuously and under rough environmental conditions such as a wide range of temperatures, vibrations and strong electromagnetic fields, it is not guaranteed that the captured measurement data is free of unwanted effects such as offset and drift. In this paper, a method to compensate for the unwanted effects on measurement data in powertrain test benches is developed and validated with real measurements on a powertrain test bench. This method uses symmetries in the type of measured data, for example drag losses in electric machines, to identify and calculate the offset while the test bench is running. It can be applied to various kinds of measurement systems such as torque sensors which usually rely on strain gauges for their operation, DC current transducers, DC voltage metering devices and more. It does not require an interruption of the test procedure and can be integrated into the test benches main control system to automatically and periodically execute the compensation routine for the specified measurement equipment. These factors make it a valuable tool to improve the accuracy, quality and consistency of measurement data on powertrain test benches.
Mathias Jaksch, Hans-Christian Reuss
Measuring Internal Energy with Force
Abstract
Today’s Components are getting more and more efficient. The higher the efficiency, the smaller the steps in efficiency gain is between two development steps. This is challenging measuring techniques and equipment to prove these small gains. The calculation of efficiency can be carried out using power loss and power input. This method has advantages in measuring uncertainty when efficiency increases. One way to measure power loss of a device is to measure the rejected heat with calorimetric measurement methods. The main disadvantage is that the change of the temperature field (internal energy) has to be constant. When this state is reached, the rejected heat over the surface is equal to the internally produced power loss. This takes mostly a long time. The challenge of getting faster in measurement is hence to measure the change of the internal energy simultaneously to the rejected heat over the surface of the device. One part of the internal energy of technical devices can be a metal housing. The internal energy of metals, with respect to heat accumulation, is proportional to their temperature. When temperature of metals is changed, they tend to change their volume over the linear thermal expansion coefficient too. To suppress this expansion, a force or moment is necessary. This reaction can be measured with sensors. The investigates the potential of this Method to measure the change in internal energy with force or heat flux with moment.
Daniel Puscher, Hans Christian Reuss
Creation and Usage of Virtual Control Units (V-ECUs) in SIL and HIL for Development and Validation especially for Software-Defined-Vehicles (SDVs)
Abstract
The mindset to think about cars as Software-Defined-Vehicles (SDVs) is intensifying more and more by many automakers and suppliers. They aim to become strong software companies at the same time. This leads to an increasing interest in virtual development and validation environments.
Software-in-the-Loop (SIL) and Hardware-in-the-Loop (HIL) environments are both well-known and established methods which both provide meaningful advantages. To achieve highest efficiency and flexibility, the transition between SIL and HIL is key. By this, re-use of all kinds of artifacts like models and test cases between SIL and HIL becomes possible. Especially remarkable is a possible combination of virtual and real Electronic Control Units (ECUs) which leads to hybrid environments or couplings of SIL and HIL test systems. Different types of virtual control units (V-ECUs) can be used for this purpose, but the generation of V-ECUs is still an issue for many companies.
We would like to share our understanding of V-ECUs: the definition of different types especially for SDVs, the generation and usage of corresponding V-ECUs, and the SIL-HIL transition. These insights can be used for guidelines to define an overarching validation strategy at OEMs or suppliers.
Martin Ruehl, Fabian Bronner
Investigation of Digital Potentiometer Networks for Precise and Digital Calibration of Resistive Sensors in Wheatstone Bridges
Zusammenfassung
In the sensor market resistive sensors play a significant role. These types of sensors require accurate electronics for measuring. The sensors and electronics are usually calibrated individually after production. This means that recalibration during operation is not easily possible and leads to inaccurate calibrations due to aging effects over the life cycle or changing environmental influences. Measuring bridges are suitable for measuring resistive sensors. If these are equipped with digital potentiometers, the system can be readjusted via digital communication. This enables repeating calibration of the system in the field, thus eliminating ageing effects during life cycle and environmental influences. A Wheatstone bridge is used as the basis, in which the sensor is installed in one leg. A digital potentiometer network forms the other leg and is used to calibrate the bridge. Digital potentiometers are widely used components that provides a resistance value which can be change via a digital communication interface. This allows the Wheatstone bridge to be adjusted during operation using microcontrollers. One problem that arises is the relatively low resolution of a single digital potentiometer. By this the Wheatstone bridges cannot be calibrated optimally. Various digital potentiometer networks are examined and compared in this paper, achieving a better accuracy and resolution of the calibration.
Ralf Sauerwald, Lukas Brandl, Hans Christian Reuss
Reduction of the Carbon Footprint in Automotive Filtration Applications by means of Alternative Materials and Optimized Product Design
Abstract
Determining the carbon footprint of a product portfolio allows focusing on the most contributing materials and processes. The most important step is developing a meaningful CO2 database by compiling primary data with the suppliers. Secondary data from databases can complement the database but should be avoided especially for specialty materials. Conducting LCA’s for a broad product portfolio is not possible following a manual approach by using typical software. An automated process, using existing data from internal systems such as SAP and IMDS, is needed to compile the footprint for all kind of products and processes, providing a robust database for further strategy planning. Having identified the most impactful materials, developing alternative materials such as modified resins for filter media, renewable resources for sealing foams and fostering the usage of recyclate plastics offers high CO2 savings and reduction of crude-oil based materials. Alternative materials can lead to reduced emissions during production in case they need e.g., no curing of filter media or reduced temperatures in the curing ovens. Although the reduction of greenhouse gas emissions is an overarching target of the industry, it is important to maintain product performance and longevity. A reduction of CO2 by 20% is not helpful if the product needs to be replaced twice as often due to missing robustness. Only customer acceptance will lead to market penetration of sustainable alternatives, and this requires keeping the product experience on a similar level.
Lars Spelter, Sven Grebhardt, Christian Schweikl, Markus Hirt, Martin Veit, Alexander Kilian, Thilo Müller
Fiber Segment Interferometry for Automotive Strain, Shape and EV Battery Temperature Testing
Abstract
It is essential for Research and Development teams to have access to the most appropriate and capable tools and measurement technologies. Fiber Segment Interferometry (FSI) is a measurement technique with potential applications across Automotive Engineering. It combines the precision and speed of response of laser interferometry, but contained within optical fiber. The dielectric construction of optical fiber sensors also make them particularly well suited to high-voltage environments. This allows it to extend the capability of test stands to address challenging applications such as monitoring conditions inside EV battery packs and e-motors.
This approach differs from a conventional optical Fiber Bragg Grating (FBG) approach where diffraction gratings that reflect specific wavelengths of light are fabricated in optical fiber. Instead, FSI directly detects the change in length between points in the fiber providing a segment-wise distributed measurement along the entire sensor. In the case of temperature sensing, the fiber is encased in a low-friction polymer tube that is bonded to the device under test. For strain measurements, the fiber is bonded directly to the surface (as with a conventional strain gauge) to ensure transfer of mechanical strain.
This paper will evaluate the performance of the technology for distributed strain, shape, and temperature measurements for automotive applications.
Michael D. Summers, Filippo Rossi, Jack Waller, Kristopher Statham, Marian Gragert
Backmatter
Metadaten
Titel
2024 Stuttgart International Symposium on Automotive and Engine Technology
herausgegeben von
André Casal Kulzer
Hans-Christian Reuss
Andreas Wagner
Copyright-Jahr
2024
Electronic ISBN
978-3-658-45018-2
Print ISBN
978-3-658-45017-5
DOI
https://doi.org/10.1007/978-3-658-45018-2

    Premium Partner