Font Size: a A A

Modeling and Controlling a Hybrid Multi-Agent based Microgrid in Presence of Different Physical and Cyber Components

Posted on:2016-11-17Degree:Ph.DType:Dissertation
University:West Virginia UniversityCandidate:Keshtkar, HessamFull Text:PDF
GTID:1472390017980931Subject:Electrical engineering
This dissertation starts with modeling of two different and important parts of the distribution power systems, i.e. distribution line and photovoltaic (PV) systems. Firstly, it studies different approximation methods and develops a new approach for simplification of Carson's equations to model distribution lines for unbalanced power flow and short circuit analysis. The results of applying the proposed method on a three-phase unbalanced distribution system are compared with different existing methods as well as actual impedance values obtained from numerical integration method. Then steady state modeling and optimal placing of multiple PV system are investigated in order to reduce the total loss in the system. The results show the effectiveness of the proposed method in minimizing the total loss in a distribution power system.;The dissertation starts the discussion about microgrid modeling and control by implementing a novel frequency control approach in a microgrid. This study has been carried out step by step by modeling different part of the power system and proposing different algorithms. Firstly, the application of Renewable Energy Sources (RES) accompanied with Energy Storage Systems (ESS) in a hybrid system is studied in the presence of Distributed Generation (DG) resources in Load Frequency Control (LFC) problem of microgrid power system with significant penetration of wind speed disturbances. The next step is to investigate the effect of PHEVs in modelling and controlling the microgid. Therefore, system with different penetrations of PHEVs and different stochastic behaviors of PHEVs is modeled. Different kinds of control approaches, including PI control as conventional method and proposed optimal LQR and dynamic programming methods, have been utilized and the results have been compared with each other. Then, Multi Agent System (MAS) is utilized as a control solution which contributes the cyber aspects of microgrid system. The modeled microgrid along with dynamic models of different components is implemented in a centralized multi-agent based structure. The robustness of the proposed controller has been tested against different frequency changes including cyber attack implications with different timing and severity. New attack detection through learning method is also proposed and tested. The results show improvement in frequency response of the microgrid system using the proposed control method and defense strategy against cyber attacks.;Finally, a new multi-agent based control method along with an advanced secondary voltage and frequency control using Particle Swarm Optimization (PSO) and Adaptive Dynamic Programming (ADP) is proposed and tested in the modeled microgrid considering nonlinear heterogeneous dynamic models of DGs. The results are shown and compared with conventional control approaches and different multi-agent structures. It is observed that the results are improved by using the new multi-agent structure and secondary control method.;In summary, contributions of this dissertation center in three main topics. Firstly, new accurate methods for modeling the distribution line impedance and PV system is developed. Then advanced control and defense strategy method for frequency regulation against cyber intrusions and load changes in a microgrid is proposed. Finally, a new hierarchical multi-agent based control algorithm is designed for secondary voltage and frequency control of the microgrid. (Abstract shortened by ProQuest.).
Keywords/Search Tags:Different, Microgrid, Multi-agent, Modeling, System, Frequency control, Cyber, Distribution
Related items