Fuzzy-based multi-agent approach for reliability assessment and improvement of power system protection
Assessment of reliability is one of the important topics in a power system which needs more decentralized mechanism to enable an electrical load to continue receiving the power in the event of its disconnection from the main power grid. Despite the huge remarkable breakthrough in software technologi...
| Main Author: | |
|---|---|
| Format: | Thesis |
| Language: | English English English |
| Published: |
2015
|
| Subjects: | |
| Online Access: | http://umpir.ump.edu.my/id/eprint/13174/ http://umpir.ump.edu.my/id/eprint/13174/ http://umpir.ump.edu.my/id/eprint/13174/1/FKEE%20-%20NADHEER%20ABDULRIDHA%20SHALASH%20-%20CD%209632.pdf http://umpir.ump.edu.my/id/eprint/13174/2/FKEE%20-%20NADHEER%20ABDULRIDHA%20SHALASH%20-%20CD%209632%20-%20CHAP%201.pdf http://umpir.ump.edu.my/id/eprint/13174/10/FKEE%20-%20NADHEER%20ABDULRIDHA%20SHALASH%20-%20CD%209632%20-%20CHAP%203.pdf |
| Summary: | Assessment of reliability is one of the important topics in a power system which needs more decentralized mechanism to enable an electrical load to continue receiving the power in the event of its disconnection from the main power grid. Despite the huge remarkable breakthrough in software technologies, most judgements have to be based on human experts in most of the planning and operation in power system. Most of the techniques are used to address the power system reliability which consumes long computation time. Hence, an inaccurate result for system operators is inevitable as most of the variables affecting the reliability changes with time. As a result of this setback, Multi Agent System (MAS) and Fuzzy models could be used as a reliable assessment of such challenges. MAS is a collection of agents that have been applied to several power system problems such as, those in operation, markets, diagnosis and protection. In this study, in order to assess the system reliability, distribution protection system design and coordination, two models of MAS techniques to determine the suitability of the Distributed Generator (DG) location based on power system reliability and new index reliability of the relay operating time are proposed. The first MAS model is designed using three agents. The first agent is a grounding indices, which is responsible for installing DG randomly by grounding indices. The second agent is a reliability evaluation agent that uses a recursive algorithm to predict the suitability generator based on the frequency and duration reliability indices in each state while the third agent is the storage and transfer of data between the other two agents. Meanwhile, the second MAS model has been designed using two agents as follows: the first agent is a fault current agent that is to determine the fault current at all points before and after grounding; the second agent is the time operating of the agent which is used to determine the relay operating time before and after modifying fault current. The simulation results for the first and second models are done using the data obtained from Malaysia distribution network (DISCO-Net) and 69 bus test system that were implemented using Java Agent Development Framework package software. The simulation results show the effectiveness of proposed MAS approaches for selecting the best DG location in a function of improved power system grounding and reliability. Meanwhile, the simulation results for second model shown that the failure rate decreased to approximately 40% for over current and earth fault relays. The fast reliability indices are also obtained in assessing the performance of the protection system from the selection of DISCO-Net. Finally, to improve the evaluation of the reliability, the approaches of using MAS for connection of probability with the reliability fuzzy model is proposed. The probability agent is to determine the capacity in service while a fuzzy model agent is to estimate the operation or failure probability. In addition, another two agents have been developed based on Monte Carlo simulation. The first agent employed fuzzy parameters such as, current with its means and variances and the second agent is the probability of outage capacity for each state. All these agents have been applied in terms of the loss of load probability (LOLP) and loss of load expectation (LOLE), which have been implemented based on a IEEE-57 bus test system and DISCO-Net. The outcomes had shown that the fuzzy parameters of Monte Carlo simulation provided a better limitation for variance techniques in uncertainty load levels. |
|---|