Refine
Year of publication
- 2023 (6) (remove)
Document Type
- Conference Proceeding (3)
- Report (2)
- Working Paper (1)
Language
- English (6) (remove)
Has Fulltext
- yes (6) (remove)
Is part of the Bibliography
- no (6)
Keywords
- Artificial Intelligence in Education (2)
- Learning Analytics Dashboard (2)
- Natural Language Processing (2)
- Ontologies (2)
- e-Portfolios (2)
- Battery electric vehicle (1)
- Cooling (1)
- Electrical Machine (1)
- Fault Injection testing (1)
- HIL (1)
Institute
The power density of electric machines is a critical factor in various applications, i.e. like the power train. A major factor to improve the power density is boosting the electric current density, which increases the losses in the limited volume of the electric machine. This results in a need for an optimized thermal design and efficient cooling. The dissipation of heat can be achieved in a multitude of ways, ranging from air cooling to highly integrated cooling solutions. In this paper, this variety is shown and analyzed with a focus on water cooling. Further various structures in electric machines are presented.
A planar testbench is built to systematically analyze water cooling geometries. The focus lies in providing different power loss distributions along cooling channels, accurate temperature readings in a multitude of locations, as well as the pressure drop across the channel. The test bench results are aligned with simulations and simplified analytical evaluation to support the development process.
The main goal in this paper is to determine temperature gradients in the material close to the stator to quantize the potential for future cooling jacket designs. One question ,to answer is: How large the gradient is considering a realistic power loss distribution. Another sensible point are the different thermal expansions of aluminum used in cooling jackets and the steel core of the stator. This can be bypassed by using a steel cooling jacket. In this case, the performance of a steel cooling jacket compared to an aluminum version is investigated and also if light weight construction can compensate the lower thermal conductivity of steel.
After the analysis, an outlook about future changes of the measurement methods are given and first potentials for future cooling jackets are proposed.
Battery electric vehicle (BEV) adoption and complex powertrains
pose new challenges to automotive industries, requiring
comprehensive testing and validation strategies for reliability and
safety. Hardware-in-the-loop (HIL) based real-time simulation is
important, with cooperative simulation (co-simulation) being an
effective way to verify system functionality across domains. Fault
injection testing (FIT) is crucial for standards like ISO 26262.
This study proposes a HIL-based real-time co-simulation
environment that enables fault injection tests in BEVs to allow
evaluation of their effects on the safety of the vehicle. A Typhoon
HIL system is used in combination with the IPG CarMaker
environment. A four-wheel drive BEV model is built, considering
high-fidelity electrical models of the powertrain components
(inverter, electric machine, traction battery) and the battery
management system (BMS). Additionally, it enables validation of
driving dynamics, routes and environmental influences and provides
a precise analysis of the effect of powertrain system faults on driving
behavior. A possible case for a fault injection is to introduce a shootthrough fault in the inverter. Through the co-simulation, it is possible
to analyze the effects on the powertrain and the vehicle dynamics in
different driving situations (e.g. snow). This work demonstrates that
co-simulation is a valuable tool for the development and validation of
BEVs, and presents specific fault cases introduced into the
powertrain and the resulting effects tested under different driving
conditions. In addition, the study discusses the system's limitations
and future possibilities such as controller hardware integration
(Controller-HIL) and autonomous driving system validation.
Bicycle-drawn cargo trailers with an electric drive to enable the transportation of high cargo loads are used as part of the last-mile logistics. Depending on the load, the total mass of a trailer can vary between approx. 50 and 250 kg, potentially more than the mass of the towing bicycle. This can result in major changes in acceleration and braking behavior of the overall system. While existing systems are designed primarily to provide sufficient power, improvements are needed in the powertrain control system in terms of driver safety and comfort. Hence, we propose a novel prototype that allows measurement of the tensile force in the drawbar which can subsequently be used to design a superior control system. In this context, a sinusoidal force input from the cyclist to the trailer according to the cadence of the cyclist is observed. The novelty of this research is to analyze whether torque impulses of the cyclist can be reduced with the help of Model Predictive Control (MPC). In addition, the powertrain of the trailer is intended to support the braking process of the system with regenerative braking. In the context of this research, a first MPC controller design is carried out and analyzed with the help of a Hardware-in-the-Loop (HIL) approach where the microcontroller of the power electronics is included as hardware to ensure the vehicle dynamics control interacts properly with the lower-level field-oriented control. The battery and motor subsystems are simulated in a Typhoon HIL 604, which is supplemented by a vehicle dynamics model of the trailer that is integrated as a Functional Mock-Up Unit (FMU). First results indicate that the MPC longitudinal dynamics controller supports the driver during acceleration, attenuates the sinusoidal oscillations and reduces the force with which the trailer pushes the bicycle during braking.
In today's AI-driven era, computer vision, including autonomous driving, robotics, and healthcare, is prevalent. How-ever, acquiring ample data while managing resources and privacy constraints is challenging. This article proposes a solution: synthetic data generation. We use CAD software to craft intricate 3D models, process them in Blender, and evaluate quality using metrics like Structural Similarity and PSNR (Peak Signal to Noise Ratio). Synthetic data achieves up to 90% similarity with real data and an average PSNR of 21dB. Our method offers a streamlined, dependable ap-proach for enhancing computer vision, especially in object detection, addressing data acquisition challenges.
In this report, we present a system architecture and technical infrastructure that allows for a seamless integration of a standard e-portfolio platform, suitable AI-based tool chains, and interactive dashboard applications for students and teachers. This is followed by a description of how the architecture's components interact in a typical analysis workflow. Furthermore, we examine the development of an integrated knowledge architecture for guiding the semantic analysis of e-portfolios considering two different knowledge resources. The proposed architecture is the first draft of an AI-supported e-portfolio analysis system to be used in real-life scenarios at the University of Education in Weingarten.
This report provides an in-depth analysis of the state-of-the-art research and technologies of e-portfolio analysis, and elicit the current situation and the needs of students and teachers in the context of e-portfolios in higher education. It summarizes the results of this problem and context analysis within the AISOP project and builds the foundation for developing a user-centred tool infrastructure for AI-based e-portfolio analysis.