检索结果-内蒙古大学图书馆

arXiv 2024年

作者： Leiser, Florian Eckhardt, Sven Leuthe, Valentin Knaeble, Merlin Maedche, Alexander Schwabe, Gerhard Sunyaev, Ali Institute of Applied Informatics and Formal Description Methods Karlsruhe Institute of Technology Karlsruhe Germany Department of Informatics University of Zurich Zurich Switzerland Human-Centered Systems Lab. Karlsruhe Institute of Technology Karlsruhe Germany

Large language models (LLMs) are prone to hallucinations, i.e., nonsensical, unfaithful, and undesirable text. Users tend to overrely on LLMs and corresponding hallucinations which can lead to misinterpretations and errors. To tackle the problem of overreliance, we propose HILL, the "Hallucination Identifier for Large Language Models". First, we identified design features for HILL with a Wizard of Oz approach with nine participants. Subsequently, we implemented HILL based on the identified design features and evaluated HILL’s interface design by surveying 17 participants. Further, we investigated HILL’s functionality to identify hallucinations based on an existing question-answering dataset and five user interviews. We find that HILL can correctly identify and highlight hallucinations in LLM responses which enables users to handle LLM responses with more caution. With that, we propose an easy-to-implement adaptation to existing LLMs and demonstrate the relevance of user-centered designs of AI artifacts. © 2024, CC BY-SA.

关键词： Computational linguistics

来源：评论

学校读者我要写书评

暂无评论

Tuning of discrete-time angular droop controllers

Tuning of discrete-time angular droop controllers

引用

Control technology and Applications (CCTA),

作者： Taouba Kaouther Jouini Zhiyong Sun Veit Hagenmeyer Institute of Automation and Applied Informatics (IAI) at Karlsruhe Institute of Technology Germany Department of Electrical Engineering Eindhoven University of Technology Netherlands

We study discrete-time angular-droop control in power inverter networks and leverage second-order information for the tuning. We thereby increase the rate of convergence of the closed-loop angle trajectories towards an induced steady state angle. In particular, we select the input penalty matrix, representing a tuning knob to be a sufficiently small diagonal matrix so that the closed-loop system approximately yields Newton’s method. In this way, we ensure quadratic convergence, a feat that is otherwise not achievable with a constant, sufficiently large diagonal input penalty matrix. This draws a link between second-order methods known for their quadratic rate of convergence and the tuning of inverse optimal stabilizing controllers for discrete-time integrator dynamics.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Nonlinear Analysis of the Synchronous Reference Frame Phase-Locked Loop Under Unbalanced Grid Voltage

TechRxiv

引用

TechRxiv 2024年

作者： Ponomarev, Anton Hagenmeyer, Veit Gröll, Lutz Institute for Automation and Applied Informatics Karlsruhe Institute of Technology Hermann-von-Helmholtz-Platz 1 Baden-Württemberg Eggenstein-Leopoldshafen76344 Germany

The synchronous reference frame phase-locked loop (SRF-PLL), also called dqPLL, is an electric circuit commonly used in power electronics to estimate the phase angle of a three-phase AC grid. If the voltage is unbalanced, the PLL is modeled as a periodically forced nonlinear oscillator and is known in practice to converge to a steady oscillation. In the existing literature, the oscillation has been studied via linearization assuming a low level of unbalance. Aiming for stronger nonlocal statements, we present nonlinear analysis. We apply the method of autonomous comparison systems and incremental stability to show that the steady oscillation is unique and attractive in a wider neighborhood. Its lock-in domain is estimated using numerical phase portrait analysis. The oscillation is further approximated up to the terms of the second order in the unbalance factor – it yields an estimation of the time average of the PLL’s phase error which is not visible by linearization only. The results provide stability guarantees and can guide the tuning of SRF-PLL. © 2024, CC BY.

关键词： Phase locked loops

来源：评论

学校读者我要写书评

暂无评论

Ensuring Data Privacy in AC Optimal Power Flow with a Distributed Co-Simulation Framework

arXiv

引用

arXiv 2024年

作者： Dai, Xinliang Kocher, Alexander Kovačević, Jovana Dindar, Burak Jiang, Yuning Jones, Colin Çakmak, Hüseyin Hagenmeyer, Veit Institute for Automation and Applied Informatics Karlsruhe Institute of Technology Germany Automatic Control Laboratory École Polytechnique Fédérale de Lausanne Switzerland

During the energy transition, the significance of collaborative management among institutions is rising, confronting challenges posed by data privacy concerns. Prevailing research on distributed approaches, as an alternative to centralized management, often lacks numerical convergence guarantees or is limited to single-machine numerical simulation. To address this, we present a distributed approach for solving AC Optimal Power Flow (OPF) problems within a geographically distributed environment. This involves integrating the energy system Co-Simulation (eCoSim) module in the eASiMOV framework with the convergence-guaranteed distributed optimization algorithm, i.e., the Augmented Lagrangian based Alternating Direction Inexact Newton method (ALADIN). Comprehensive evaluations across multiple system scenarios reveal a marginal performance slowdown compared to the centralized approach and the distributed approach executed on single machines—a justified trade-off for enhanced data privacy. This investigation serves as empirical validation of the successful execution of distributed AC OPF within a geographically distributed environment, highlighting potential directions for future research. Copyright © 2024, The Authors. All rights reserved.

关键词： Data privacy

来源：评论

学校读者我要写书评

暂无评论

Dream to Drive: Learning Conditional Driving Policies in Imagination

Dream to Drive: Learning Conditional Driving Policies in Ima...

引用

International Conference on Intelligent Transportation

作者： Tim Joseph Marcus Fechner Ahmed Abouelazm Marius J. Zöllner FZI Research Center for Information Technology Karlsruhe Germany Institute of Applied Informatics and Formal Description Methods (AIFB) Karlsruhe Institute of Technology Karlsruhe Germany

ISBN: (数字)9798331505929

ISBN: (纸本)9798331505936

Learning driving policies to control autonomous vehicles via reinforcement learning (RL) offers a solution to learn optimal driving behavior directly from sensor data. However, designing a reward function that leads to a driving policy that works in any situation has not yet been achieved. Instead, one has to use different reward functions for different situation. While possible with model predictive control (MPC), approaches based on RL must be re-trained any time the reward function changes. We suggest a different direction: we propose a model-based RL agent that learns a conditional driving policy by simulating behavior for many different reward functions in imagination using a world model. We do so by randomly sampling parameters that shape the reward function and optimizing an actor-critic policy that is conditioned on these parameters. We evaluate our approach in CARLA and demonstrate that our approach combines the flexibility of MPC with the long-term capabilities and execution speed of RL.

关键词： Training Runtime Shape Decision making Reinforcement learning Planning Reliability Driver behavior Intelligent transportation systems Predictive control

来源：评论

学校读者我要写书评

暂无评论

Grid supporting nonlinear control for AC-coupled DC Microgrids

Grid supporting nonlinear control for AC-coupled DC Microgri...

引用

IEEE International Conference on DC Microgrids (ICDCM)

作者： Ömer Ekin Filipe Perez Friedrich Wiegel Veit Hagenmeyer Gilney Damm Institute for Automation and Applied Informatics Karlsruhe Institute of Technology (KIT) Karlsruhe Germany Supergrid Institute Lyon France COSYS-IMSE Laboratory University Gustave Eiffel Marne-la-Vallée France

ISBN: (数字)9798350363630

ISBN: (纸本)9798350363647

Direct Current (DC) Microgrids have risen as a good option for better integrating distributed renewable energies and various highly dynamic DC loads, such as electrical vehicle supply equipment or industrial grids. However, for the large-scale deployment of such DC Microgrids, developing enhanced control strategies for these inverter-based, highly dynamic grids is crucial. The present paper introduces an innovative nonlinear control tailored for DC Microgrids supporting Alternative Current (AC) grids. Building upon existing methods, we extend the nonlinear control paradigm by incorporating multiple batteries and an AC link. This integration not only bolsters support for the AC grid, but also enhances the stability of DC voltage. We validate the effectiveness of the proposed algorithm through experimental evaluations conducted on a Power Hardware-in-the-Loop (PHIL) DC Microgrid configuration and compare it to a cascaded PI controller based on a linearized model.

关键词： Power system dynamics Microgrids Power system stability Stability analysis Batteries Maintenance Voltage control

来源：评论

学校读者我要写书评

暂无评论

Dynamic Mapping for Evolutionary Algorithm Based Optimization of Energy Hub Gas Scheduling

Dynamic Mapping for Evolutionary Algorithm Based Optimizatio...

引用

IEEE International Conference on Smart Energy Grid Engineering (SEGE)

作者： Rafael Poppenborg Katharina Beisswanger Christian Hotz Kevin Förderer Thomas Kolb Veit Hagenmeyer Karlsruhe Institute of Technology (KIT) Institute for Automation and Applied Informatics (IAI) DVGW-Research Center at Engler-Bunte-Institute of Karlsruhe Institute of Technology Karlsruhe Institute of Technology (KIT) Engler-Bunte-Institute Fuel Technology Division (EBI ceb)

To reach the goal for limiting climate change, a large amount of renewable energy sources (RESs) needs to be installed and integrated into the existing electrical grid infrastructure. This leads to an increasing demand for flexibility that can be provided by the promising approach of the Energy Hub Gas (EHG). For intelligently controlling the different components of an EHG enhanced optimization methods are required. Due to the complexity of this optimization task a concept and its evaluation for enhancing the ability of an evolutionary algorithm (EA) to track several objectives is presented. With this concept a dynamic mapping of the objective functions to the fitness value of an EA by calculating a degree of fulfillment (DOF) in relation to the maximum and minimum possible boundary value is enabled. The evaluation of this concept shows a significant improvement of optimization results with increasing the DOF by 10.2%.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Net load forecasting using different aggregation levels

引用

Energy informatics 2022年第1期5卷 1-21页

作者： Beichter, Maximilian Phipps, Kaleb Frysztacki, Martha Maria Mikut, Ralf Hagenmeyer, Veit Ludwig, Nicole Karlsruhe Institute of Technology Institute for Automation and Applied Informatics Hermann-von-Helmholtz-Platz 1 Eggenstein-Leopoldshafen76344 Germany Cluster of Excellence Machine Learning University of Tübingen AI Research Building Maria-von-Linden-Str. 6 Tübingen72076 Germany

In the electricity grid, constantly balancing the supply and demand is critical for the network’s stability and any expected deviations require balancing efforts. This balancing becomes more challenging in future energy systems characterised by a high proportion of renewable generation due to the increased volatility of these renewables. In order to know when any balancing efforts are required, it is essential to predict the so-called net load, the difference between forecast energy demand and renewable supply. Although various forecasting approaches exist for both the individual components of the net load and the net load itself, it is unclear if it is more beneficial to aggregate several specialised forecasts to obtain the net load or to aggregate the input data to forecast the net load with one approach directly. Therefore, the present paper compares three net load forecasting approaches that exploit different levels of aggregation. We compare an aggregated strategy that directly forecasts the net load, a partially aggregated strategy that forecasts demand and supply separately, and a disaggregated strategy that forecasts demand and supply from each generator separately. We evaluate the forecast performance of all strategies with a simple and a complex forecasting model, both for deterministic and probabilistic forecasts, using one year of data from a simulated realistic future energy system characterised by a high share of renewable energy sources. We find that the partially aggregated strategy performs best, suggesting that a balance between specifically tailored forecasting models and aggregation is advantageous. © 2022, The Author(s).

关键词： Aggregates

来源：评论

学校读者我要写书评

暂无评论

From Stationary to Flexible Plant Operation: Extension of a Co-Current Rotary Dryer Model for Energy Demand Flexibility

From Stationary to Flexible Plant Operation: Extension of a ...

引用

International Conference on Process Control (PC)

作者： Jan M. Schaßberger Veit Hagenmeyer Lutz Gröll Institute for Automation and Applied Informatics Karlsruhe Institute of Technology Hermann-von-Helmholtz-Platz 1 Eggenstein-Leopoldshafen Germany

Industrial greenhouse gas emissions must be reduced to meet the climate goals. One suitable measure is the electrification of industrial processes, which poses great challenges in conjunction with the necessary transition of the electrical grid towards renewable energy sources. In this context demand side flexibility of energy-intensive industrial processes is considered to take a key role, which requires suitable models that can predict the system behavior over a large operating range. For this reason, we developed in the past a new model of a co-current rotary dryer as an example of this energy-intensive process group with the focus on energy demand flexibility. In the present paper, this model is extended by the consideration of the dryer shell, which is usually neglected in modeling. By simulation it is shown that the shell has a small influence on the temperature and humidity profiles in a stationary state. However, transient simulations indicate that the heat capacity of the shell has a strong effect on the dynamic system response.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Digital Twin Architecture to Use for Optimizing an AoD-Printing Process

Digital Twin Architecture to Use for Optimizing an AoD-Print...

引用

Design, Test, Integration and Packaging of MEMS/MOEMS

作者： Hanna Pfannenstiel Martin Ungerer Ingo Sieber Institute for Automation and Applied Informatics (IAI) Karlsruher Institute for Technology (KIT) Leopoldshafen-Eggenstein Germany

ISBN: (数字)9798350378269

ISBN: (纸本)9798350378276

Digital Twins have been gaining relevancy over the past decade as computing power becomes increasingly more powerful and cheaper. However, most digital twins require an already profound knowledge of their physical counterpart in order to mimic it. This paper proposes a data architecture which enables a digital twin to learn from its physical counterpart and help to build models in applications where little established models are available. The digital twin in this paper focuses on accompanying an additive manufacturing process.

关键词： Computational modeling Memory Computer architecture Packaging Three-dimensional printing Data models Digital twins

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：