Distributed model predictive load frequency control of multi-area interconnected power system considering effect of photovoltaic sources and economical constraints

Mohajeri Avval, Armin and Akbarimajd, Adel and Shayeghi, Hosein and Sobhani, Behrooz (2017) Distributed model predictive load frequency control of multi-area interconnected power system considering effect of photovoltaic sources and economical constraints. Masters thesis, University of Mohaghegh Ardabili.

Preview

Text (کنترل پیش‌بین توزیع شده بار فرکانسی سیستم چند ناحیه‌ای با در نظر گرفتن اثر منابع فتوولتائیک و قیود اقتصادی) Armin Mohajeri Avval.pdf Download (715kB) | Preview

Abstract

Load frequency control is the mechanism for balancing between production and consumption as well as reduction of both frequency fluctuations and the interaction power among the areas to zero. With the increasing complexity of the power system, restrictions on the production and use of renewable resources such as photovoltaic sources, the need for load frequency controllers is more increased. In the last few years, the use of the predictive controllers has been increased in renewable power plants due to the characteristic of analysis of the current behavior of the system and its predictive behavior in future. Distributed load frequency controller has the ability to access local information and its adjacent areas, which is one of the advantages of the controller. Hence distributed predictive controller is a good choice in renewable power plants because of ability of adaption to system changes, ant its logical and multi-purpose solution. The purpose of this research is to provide a distributed multi- agent predictive controller strategy to reduce frequency fluctuations by the participation of all areas of the power system. To this end, intelligent agents are used for decision making and determination of control policies in the presence of photovoltaic sources. Distributed controller in each area runs an agent that pursues the aim of stability of power system based on the immediate behavior of each area and the relationship between the agents. In this regard, after the linearization of the discrete nonlinear dynamical model, the predicted behavior time is substituted in the multi-objective optimizer equations. Using a series technique based on the Lagrangian method, the control signal determines the power with the minimum computation time, to allow the continuation of the load supply process. The performance of the proposed method is investigated by simulated in discrete time and the results are analyzed.

Actions (login required)

View Item