الفهرس 
INTRODUCTIONOnly 14 pages are availabe for public view
 |  | from | 167 |  |  |
| | | from | 167 | | |
Abstract
Optimal coordination of directional overcurrent relays (DOCRs) has been studied in many researches based on the suppose that the fault current passes through these relays has a fixed value. However, during short circuit (SC) conditions in real distribution and industrial networks, the SC current doesn’t have a fixed value. That is due to the exist of synchronous generators and the transient behavior of these synchronous machine and due to the contribution of induction motors (IMs) during SC conditions; hence the DOCRs see a time variable fault current waves. Therefore, assuming a fixed SC current during the calculation of relays settings may lead to miscoordination states in real networks. So that, to avoid possible miscoordination between the DOCRs, the transient states of fault current must be considered during coordination study of the DOCRs relays.
The main contribution in this thesis is the introduction of two algorithms; proposed one-level user-defined characteristics method and proposed non-standard two-level user-defined characteristics method; as a solution for the DOCRs coordination problem in the presence of transient fault current contributed from both DGs and utility. These two proposed characteristics (one-level and non-standard two-level) of DOCRs are based on the user-defined dynamic model of overcurrent relays. It should be focused that the name of one-level characteristics and two-level characteristics has an essential relation to the shape of the relay characteristics curve. As, the one-level characteristics means that the curve has no jump points, i.e., a smooth curve from start to end. In contrast, the two-level characteristic means that the curve contains a jump point that moves it from level to level, i.e., it has a discontinuity point. Moreover, the user-defined terminology means that the user could define time over current characteristics which differ from the standard time over current characteristics. While, the characteristics curve of a DOCR in the two-level based method is described by the non-standard terminology on contrary to the characteristics curve of a DOCR in the one-level based method, as the shape of the DOCR curve in the two-level based method no longer has the common known standard curves of the OC relays (inverse, very-inverse or extremely-inverse). In other words, the operating points of a DOCR in the two-level based method will not pass through a unique smooth level, but it will follow a curve per the separation point and another curve after the separation point. In the presented user-defined dynamic models of the OC relay, α and β; the characteristics coefficients of a relay; are optimally selected in addition to the two traditional settings time multiplying settings (TMS) and pickup current (Ip). It worth to mention that, through all the presented techniques in this thesis, the OC relay four parameters are considered to have continuous values not restricted to certain discrete values, un like the common IEC standard values for α and β coefficients of the inverse, very-inverse or extremely-inverse characteristic curves.
In this thesis, the nonlinear operating time equation of the OC relay is re-formulated as a linear optimization problem for simplicity and for reducing the computational time.
The coordination problem in this thesis is solved by three proposed optimization techniques; two hybrid meta-heuristic techniques: gravitational search algorithm- sequential quadratic programming (GSASQP) and particle swarm optimization- gravitational search algorithm (PSOGSA) and one mathematical method: find minimum constraints (FMINCON). Each one of the presented solution techniques is applied to three different test systems- IEEE 3, 9 and 14 bus systems.
Regarding to the optimal coordination of DOCRs considering the one-level characteristics of the DOCRs, each one of the three systems will be studied in four different models - to show the efficiency of the proposed one –level user-defined dynamic model technique. The four models for each system are; model 1: optimal coordination based on the conventional one-level characteristics of DOCRs while considering the steady-state fault current value, model 2: optimal coordination based on the user-defined one-level characteristics of DOCRs while considering the steady-state fault current value, model 3: optimal coordination based on the conventional one-level dynamic model characteristics of DOCRs while considering the transient states of fault current and model 4: the proposed method; optimal coordination based on the one-level user-defined dynamic model characteristics of DOCRs while considering the transient states of fault current; which optimally optimize the four parameters of relays while considering the transient fault current. The results of optimal coordination of DOCRs considering the one-level characteristics of DOCRs show a superiority of the meta-heuristic techniques over the mathematical method also show that the presented one-level user-defined dynamic model of OC relays technique is the most efficient method, among the other one-level coordination based methods, for DOCRs coordination in presence of transient fault.
For further enhancement of the coordination results obtained based on the presented one-level user-defined dynamic model of OC relays, the proposed non-standard two-level user-defined dynamic model characteristics is suggested. The optimization results based on the of the proposed non-standard two-level user-defined dynamic model method show a great reduction in the relays operating times and in the objective function compared to the best method in the one-level optimization based models (the proposed one-level user-defined dynamic model method). Also, the DOCRs coordination results based on the proposed non-standard two-level user-defined dynamic model confirm the superiority of the meta-heuristic techniques over the mathematical method.
Here, MATLAB Simulink program is used to obtain the transient SC current waves of the test systems under study while MATLAB M-file is used to apply the proposed technique and get optimization results.