Loading...

Design of an Optimal Multi Agent System for Decentralized Implementation of the Active Distribution Management System's Functions Using State Estimation

Gholami, Mohammad | 2021

250 Viewed
  1. Type of Document: Ph.D. Dissertation
  2. Language: Farsi
  3. Document No: 54528 (05)
  4. University: Sharif University of Technology
  5. Department: Electrical Engineering
  6. Advisor(s): Abbaspour Tehrani Fard, Ali; Moeini Aghtaie, Moein
  7. Abstract:
  8. In recent years, due to the increase in distributed generation, the use of renewable energy sources and the emergence of restructuring and energy markets, traditional distribution networks transform to complex and active distribution networks. These issues, along with requirements such as increasing power quality and reliability, have increased the need to provide active distribution management systems and the use of effective network monitoring, control and protection systems. Due to the complexity and expanse of distribution networks, as well as the development of the use of metering elements and remotely controllable elements, one of the important issues in these systems is the introduction of new monitoring and control strategies for the active distribution management system. In this regard, due to the fact that centralized approaches have weaknesses from perspectives such as time response and reliability, the tendency to use decentralized approaches has increased. Various schemes have been proposed for the decentralized implementation of the operational functions of the distribution management system (monitoring or control), such as state estimation, fault management and voltage control. However, the network division considered for the decentralized execution of these functions is different and predetermined in terms of the number and size of zones, and a study on comprehensive architectural evaluation (hierarchical or distributed) and determining the unit and optimal decomposition of the active distribution network for equipping with local and intermediate processing units (between local processing units and the control center) has not been addressed. To meet this challenge, due to the mutual effects of the operational functions of the distribution management system and its implementation method (network architecture and network zoning) and also the impossibility of using all functions, the state estimation function has been used as the most important function of this system in the dissertation. For this purpose, as the first step, the centralized state estimation algorithm in the active distribution network is modelled and to satisfy the minimum standard accuracy for the network voltage profile, a method for optimal location of the PMU in the network is presented. Next, using a hierarchical distributed architecture, the state estimation algorithm for decentralized execution is developed and its performance is compared with distributed execution methods. Finally, by providing an algorithm to consider all possible scenarios for distribution network divisions, a method for determining the optimal number and size of network zones in the proposed architecture is presented using state estimation operational indices, i.e. accuracy of results, reliability of achieving to the results, and process execution time. Finally, a discussion about the ability to implement some important operational functions of the distribution management system in the proposed architecture is presented. Also, to avoid the expansion of the topics covered in the dissertation and because recent fault location methods in the network are based on state estimation, a new method for modifying the state estimation process in the fault conditions and an algorithm for decentralized implementation of fault location process based on the suggested hierarchically distributed architecture are proposed. All methods and models presented in this dissertation have been studied numerically and the results of different cases have been reported to validate the presented methods
  9. Keywords:
  10. Multiagent System ; Decentralized Strategy ; State Estimation ; Voltage Control ; Fault Management ; Fault Location ; Active Distribution System ; Distributed Management ; Hierarchical Distributed Method

 Digital Object List

 Bookmark

No TOC