Scholarlink

Abstract <p>We address the current problem of ensuring the protection of energy system elements from the effects of large disturbances. Ranking the system elements in terms of their criticality allows us to solve this problem more effectively. Based on the ranking, we will be able to select and protect the most critical elements. The criticality of an element is defined as the degree of impact of its failure on the performance of the system as a whole. In this context, we propose a comprehensive approach to assessing the element criticality using different methods. In particular, we use the brute force for element group failures, the global sensitivity analysis, and the study of topological and flow parameters of element criticality. The usage of global sensitivity analysis is a novelty of our approach compared to known similar approaches. We have also developed service-oriented applications to implement the above methods. We rationally allocate resources to run services and provide a trade-off between computation time, processor load, and load balancing across resource nodes. This is achieved by taking into account the method execution overheads, resource characteristics, and administrative quotas for the resource usage. The benefits of our approach have been demonstrated in assessing the criticality of elements (transport facilities) of a small gas supply system. In contrast to the known approaches of similar purpose, we can simultaneously classify three groups of elements: vulnerable elements, dependent elements, and bottlenecks. Based on the correlation analysis of the criticality of the classified elements, we have showed the sufficiency of considering the failures of all possible pairs of elements to detect dependent elements. We also assess the methods in terms of their computational accuracy and execution overhead. We achieved an acceptable solution time of just over 6 hours and an average CPU utilization on resource nodes of 74% to 96%. This is a good result considering the heterogeneity of the resources and the different methods used.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Comprehensive Assessment of Criticality of Energy System Elements Based on Service-Oriented Approach

  • 期刊论文

  • 发表日期：

    2025年12月1日

查看全文
  • A. Edelev,
  • O. Basharina,
  • M. Voskoboinikov,
  • A. Feoktistov,
  • A. Tchernykh

摘要

Abstract

We address the current problem of ensuring the protection of energy system elements from the effects of large disturbances. Ranking the system elements in terms of their criticality allows us to solve this problem more effectively. Based on the ranking, we will be able to select and protect the most critical elements. The criticality of an element is defined as the degree of impact of its failure on the performance of the system as a whole. In this context, we propose a comprehensive approach to assessing the element criticality using different methods. In particular, we use the brute force for element group failures, the global sensitivity analysis, and the study of topological and flow parameters of element criticality. The usage of global sensitivity analysis is a novelty of our approach compared to known similar approaches. We have also developed service-oriented applications to implement the above methods. We rationally allocate resources to run services and provide a trade-off between computation time, processor load, and load balancing across resource nodes. This is achieved by taking into account the method execution overheads, resource characteristics, and administrative quotas for the resource usage. The benefits of our approach have been demonstrated in assessing the criticality of elements (transport facilities) of a small gas supply system. In contrast to the known approaches of similar purpose, we can simultaneously classify three groups of elements: vulnerable elements, dependent elements, and bottlenecks. Based on the correlation analysis of the criticality of the classified elements, we have showed the sufficiency of considering the failures of all possible pairs of elements to detect dependent elements. We also assess the methods in terms of their computational accuracy and execution overhead. We achieved an acceptable solution time of just over 6 hours and an average CPU utilization on resource nodes of 74% to 96%. This is a good result considering the heterogeneity of the resources and the different methods used.