Development of a Hybridized Model for Detection of Voltage Collapse in Electrical Power Systems

Abstract

Voltage collapse is an occurrence in power systems that are heavily loaded, faulted and have reactive power shortages. It is a system instability involving large disturbances (including rapid increase in load or power transfer) and usually associated with reactive power deficits. Numerous power system blackouts in the past indicates/shows that enough researches have not been done to solve the problem of voltage instability and the resultant voltage collapse. This research paper therefore develops a hybridized sensitivity based voltage collapse prediction index model for detection of voltage collapse in electrical power systems. Sensitivity based method and the voltage collapse prediction index method were hybridized using the linearized basic power flow equations. Newton-Rapson technique was employed to solve the linearized power flow equation to compute the changes in real and reactive powers with respect to change in voltage magnitudes and voltage phase angles. The stressed version of the standard IEEE 24 bus system was used as an input parameter for the computation.Load shedding technique was then embarked on to find the weak bus for load shedding and the required generation reduction to maintain power balancing while buses having an unstable voltage profile were selected using the developed model. Bus number 3 had the least voltage magnitude of 1.018 p.u before stressing while bus number 16 recorded the highest voltage magnitude of 1.082 p.u even though the voltage magnitudes of these two buses were 0.623 p.u and 0.4611 p.u respectively after being stressed. The least and highest loads after shedding with the sensitivity based method were 0.11925 p.u and 0.6214 p.u respectively while the least and highest loads after shedding with the VCPI method were 0.1368 p.u and 0.6148 p.u respectively. With the hybridized sensitivity based VCPI model, the least and highest loads after shedding were 0.2068 p.u and 0.6314 p.u respectively confirming the efficiency of the HSBVCPI model.The total load demand met by the system with the hybridized model was higher compared to that of the sensitivity based method and the VCPI method. The power generated after the multi-stage load shedding by the sensitivity based method, VCPI method and the hybridized model were 2.12 p.u, 3.05 p.u and 3.65 p.u respectively showing that the power balance equations were satisfied since the total generation was greater than the total demand. The hybridized model improves better the voltage profiles of many of the load buses as compared to the other two methods. Bus numbers 4, 5, 12 and 15 were selected for load shedding with the hybridized model since all the load buses had loads less or equal to their loadability margin, hence, all the buses have satisfied the loadability condition. Even though, the VCPI method perform better than the sensitivity based method in detecting voltage collapse, the hybridized model performs best in term of detection of power system voltage collapse and load shedding implementation.

Keywords: Voltage Collapse, Sensitivity, Voltage Collapse Prediction Index (VCPI), Voltage Instability, Load Shedding, Power Balance, Newton-Raphson technique.