Examensarbeten för masterexamen

Länka till denna samling:

https://odr.chalmers.se/handle/20.500.12380/32

Browse

Visar 1 - 5 av 2093

A Model for Simulating FCR Prequalification Tests for Kaplan Turbines
(2024) Eriksson, Eddie; Chalmers tekniska högskola / Institutionen för elektroteknik; Chen, Peiyuan; Ekstrand, Christian
Abstract A Kaplan turbine was modeled in Simulink with the purpose of simulating prequalification tests for Frequency Containment Reserve (FCR) provision, based on technical requirements for FCR introduced in 2023. The model did not successfully predict fulfillment of steady-state requirements, and was too optimistic in its prediction of fulfillment of dynamic requirements. However, all tests predicted to fail also failed in real-world tests, indicating a potential use case for ruling out Kaplan turbines not suitable for FCR provision. A sensitivity analysis showed that simulation results were mostly unaffected by runner servo parameters, indicating a flaw in the model. Compensation with a first order low-pass filter significantly improved the accuracy of simulated dynamics. However, several model parameters, including the time constant for the low-pass filter, needed real-world tests for estimation. This reduced the likelihood of achieving high simulation accuracy ahead of conducting real prequalification tests. Using the simulated prequalification tests, methods for improving performance were explored. Both tuning the FCR controllers and the servos controlling guide vanes and runner blades were shown to improve dynamic performance, but tuning the servos improved the stability margins more. An approach to active power feedback was introduced and resulted in perfect fulfillment of the steady-state requirements. The dynamic performance requirements were not affected much, but the stability margins were severely worsened due to larger control movements at higher frequencies. The results suggested that fulfilling all technical requirements for FCR will be very challenging for Kaplan turbines. However, with increased market participation from other sources, this will likely not lead to insufficient FCR capacity in the near future.
Design and analysis of a low speed and high power permanent magnet synchronous motor for a ship’s pod
(2024) Umami, Muhammad Irsyadul; Yang, Jinzhe; Chalmers tekniska högskola / Institutionen för elektroteknik; Thiringer, Torbjörn
Abstract The route between Gothenburg Port and Frederikshavn Port requires a ship propulsion system capable of delivering 200 kNm torque at 75 RPM. A Permanent Magnet Synchronous Motor (PMSM) is designed for this purpose, with Motor-CAD employed for comprehensive simulations, including geometric design and parameter optimization. The motor, rated at 1568 kW with a maximum current of 800 A, incorporates advanced design considerations to enhance performance. Magnet material selection plays a crucial role in determining the motor’s efficiency and losses. A comparison between the ferrite and the neodymium magnets reveals that the neodymium achieves a higher efficiency of 95.9% compared to 94.1% for the ferrite. Additionally, the neodymium results in significantly lower total losses of 64.19 kW, compared to 95.34 kW for the ferrite, demonstrating superior electromagnetic performance. To address thermal challenges, seawater heat dissipation is implemented as an effective cooling strategy, ensuring optimal system performance. Mechanical stress simulations further guarantee the safety and reliability of the vessel. A life cycle assessment (LCA) evaluates the environmental sustainability of the propulsion system, covering every phase from material selection to decommissioning. This work integrates thermal management, structural analysis, and sustainability to deliver an efficient and environmentally friendly marine propulsion system, offering valuable insights for the development of next-generation ship technologies.
Safety Assessment for Precautionary Collision Avoidance - A Collision Avoidance System Based on Safety-Set
(2024) Teng, Lizi; Lyu, Peilin; Chalmers tekniska högskola / Institutionen för elektroteknik; Murgovski, Nikolce; Rundstedt, Karl
Abstract This thesis presents the implementation of a precautionary collision avoidance system, centered on obtaining a safety set and employing Model Predictive Control (MPC) based on the safety set. The safety set, which defines the boundaries within which the vehicle can safely operate, is derived through optimization problems or analytical calculations. By applying the approaches, the system proactively anticipates and mitigates potential collision risks, such as hidden pedestrians emerging from blind spots. Two critical scenarios are examined: the hidden pedestrian scenario, where the vehicle must react to unforeseen obstacles, and the lane change scenario, which requires safe merging while considering nearby traffic. The integration of the precautionary safety-set with MPC allows the system to maintain a high success rate in collision avoidance while ensuring smooth and efficient vehicle operation. Experimental results confirm the system’s reliability and demonstrate its ability to maintain a high success rate in safely navigating complex scenarios, thus highlighting its potential to significantly enhance the safety and performance of autonomous driving systems.
Diffusion models for novel view synthesis in autonomous driving
(2024) Gasparyan, Artur; Qiu, Ruiqi; Chalmers tekniska högskola / Institutionen för elektroteknik; Svensson, Lennart; Svensson, Lennart; Hess, Georg; Lindström, Carl; Tonderski, Adam
Novel View Synthesis (NVS) generates target images from new camera poses using source images and their corresponding poses. It has gained prominence in the field of autonomous driving (AD) as a tool for generating synthetic data to improve perception systems. Current NVS implementations, such as Neural Radiance Fields (NeRFs), excel at constructing 3D scenes from sensory inputs but struggle to accurately render sparsely observed or unseen views. This thesis addresses these limitations by integrating Diffusion Models (DMs) into the NVS pipeline to enhance reconstruction quality in such cases. We propose a pipeline inspired by ReconFusion, training NeuRAD, a NeRF-based NVS method designed for dynamic AD data, on additional poses not present in the original training set. A pretrained, open-sourced DM, Stable Diffusion, provides supervision by refining NeuRAD’s outputs for these unseen views. To improve the DM’s performance on AD scenes, we finetune it using Low-Rank Adaptation (LoRA), enabling efficient adaptation to small datasets. ControlNet is incorporated to extend the diffusion model with additional conditioning signals, ensuring better alignment with scene-specific characteristics. Despite introducing these enhancements, our experiments reveal mixed results. While some metrics show improvement, others remain inconsistent, particularly in challenging scenarios. We identify weak conditional signals and limited LoRA rank as potential limitations. Future research should explore incorporating more robust conditioning signals, such as depth or temporal information, and training on diverse scenes to improve generalization and stability. These directions offer promising avenues for advancing NVS in AD applications.
Estimation of Degradation Modes for Lithium-ion Batteries Estimation of degradation modes of an aged battery using Open-Circuit Voltage curves
(2024) Balabhadra, Mukundh; Madha, Yogith; Chalmers tekniska högskola / Institutionen för elektroteknik; Wik, Torsten; Fridholm, Björn
Lithium-ion batteries (LIB) have become essential in the automotive industry due to their favorable properties of high volumetric and weight-based energy density, lowself-discharge rates, and relatively affordable costs. Due to their widespread use, comprehending the life cycle dynamics of LIB is paramount. Over time, the internal chemistry of these batteries undergoes changes, leading to variations in power and capacity output, commonly referred to as power and capacity fade. Accurate prediction of end-of-life (EoL) is pivotal as it allows for the mitigation of accelerated degradation risks, thereby Improving operational lifespan. The present study endeavors to analyze and develop a non-destructive methodology for estimating the degradation modes of LIBs, utilizing both pristine and aged Open-Circuit Voltage (OCV) curves. The primary objective is to devise methodologies that can estimate the degradation mode to high accuracy given the batteries’ OCV. The ultimate goal is to establish techniques for estimating degradation modes within LIBs on an online platform, thus facilitating a better understanding of battery degradation processes. By quantifying the extent of aging within the battery, this approach aims to empower Battery Management Systems (BMS) to proactively adapt and optimize operational strategies, consequently prolonging the lifespan of battery packs deployed in vehicles. This dissertation presents findings that compare experimental results with simulations conducted using PyBaMM. The experiments are conducted under diverse circumstances mirroring real-world scenarios, encompassing considerations such as down-sampled data points, sensor noise, and data segmentation. Through this comprehensive investigation, this research contributes to advancing the understanding of LIB degradation dynamics and lays a foundation for the development of robust predictive maintenance strategies for traction batteries in automotive applications.

Browse

Senast publicerade