arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2304.10702eess.SYcs.LGcs.SY
收藏

应用机器学习中的电网行为模式与泛化风险

Power Grid Behavioral Patterns and Risks of Generalization in Applied Machine Learning

Carnegie Mellon University(卡内基梅隆大学);Pacific Northwest National Laboratory(太平洋西北国家实验室)

Shimiao Li, Jan Drgona, Shrirang Abhyankar, Larry Pileggi

AI总结:

本文通过分析真实电网运行数据揭示时变拓扑、负荷与发电的时空模式,并评估现有机器学习模型因忽视这些领域特征而产生的泛化风险。

摘要:

近年来,面向电网应用的数据驱动方法已有大量文献。然而,对领域知识的考虑不足会给这些方法的实用性带来很高风险。具体而言,忽视电网特有的时空模式(如负荷、发电和拓扑等方面)可能导致模型在新输入上输出不可行、无法实现或完全无意义的预测。为解决这一问题,本文研究真实运行数据,以提供对电网行为模式的洞察,包括时变拓扑、负荷和发电,以及各负荷与发电之间的空间差异(高峰时段、多样风格)。随后,基于这些观察,我们评估了部分现有 ML 工作因在模型设计和训练中忽视这些电网特有模式而导致的泛化风险。

英文摘要:

Recent years have seen a rich literature of data-driven approaches designed for power grid applications. However, insufficient consideration of domain knowledge can impose a high risk to the practicality of the methods. Specifically, ignoring the grid-specific spatiotemporal patterns (in load, generation, and topology, etc.) can lead to outputting infeasible, unrealizable, or completely meaningless predictions on new inputs. To address this concern, this paper investigates real-world operational data to provide insights into power grid behavioral patterns, including the time-varying topology, load, and generation, as well as the spatial differences (in peak hours, diverse styles) between individual loads and generations. Then based on these observations, we evaluate the generalization risks in some existing ML works causedby ignoring these grid-specific patterns in model design and training.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑