Concepts And Algorithms For Power Network Reliability Evaluation Using Self-Organizing Maps And New Insights Into The Structure Of State Space

本研究的目的是探讨和发展电网可靠性分析的概念和方法。重点是系统和节点的可靠性。在本提案中，节点和区域这两个术语可互换使用，以表示单个主要负荷或发电注入点或集合。这样的几点。在之前美国国家科学基金会的支持下进行的这项研究得出了几点见解。拟议的研究是基于这些见解以及调查的新概念。以下是本研究的主要内容。识别了多节点电力网络的几个特性。多节点电力系统的这些性质将被严格证明，并被用于改进现有算法，以及开发更有效的电力系统可靠性分析方法。基于这些性质，提出了边界点方法的基本思想。这一理想将被研究以寻找使其成为计算高效方法的方法。为了开发更有效的多节点可靠性分析方法，提出了三个概念。这些概念是根据我们在以前的调查中获得的见解确定的。我们将详细探讨这些概念，并开发更多技术。通常有一个感兴趣的领域需要调查，人们希望有效地获得该领域的可靠性。提出了一种在不分析整个系统的情况下确定感兴趣节点的可靠性的方法。这一方法将被探索、实施和测试。自组织映射(SOM)提供了压缩数据和提取关系的能力。对SOM概念进行了研究，用于多节点可靠性评估。将自组织映射模型与蒙特卡罗仿真相结合，为电网可靠性分析开辟了一条新的途径。电力系统是容量受限的网络。开发的技术还将被应用于其他容量受限的网络，如电信、气体流动和水流动网络。

The objective of this research is to investigate and develop concepts and methods for reliability analysis of power networks. The emphasis is both on system and nodal reliability. The terms node and area are used interchangeably in this proposal to represent a single major load or generation injection point or a collection. of such points. The research conducted under the previous NSF support has led to several insights. The proposed research is based on these insights as well as new concepts for investigation. Following will be the major thrusts of this investigation.Several properties of the multi-node power networks have been identified. These properties of the multi-node power systems will be rigorously proven and used to improve the existing algorithms as well as develop more efficient methods of reliability analysis of power systems. The basic idea of boundary point method has been proposed on these properties. This ideal will be investigated to find means to make it a computationally efficient method.Three concepts have been proposed for developing more efficient methods of multi-node reliability analysis. These concepts have been identified based on the insights that we have obtained during our previous investigation . These concepts will be explored in detail and more techniques will be developed. Often there is an area of interest for investigation and one would like to obtain reliability of this areaefficiently. A method is proposed for determining the reliability of a node of interest without analyzing the whole system. This method will be explored, implemented and tested. The self-organizing maps (SOM) provide the capability to compress data and extract relationships. The SOM concept will be investigated for multi-node reliability evaluation. It appears that combination of SOM and Monte Carlo Simulation could result in a new way of analyzing the reliability of power networks. Power systems are capacitated networks. The developed techniques will also be investigated for applications to other capacitated networks such as telecommunication, gas-flow and water-flow networks.