Date of Award
Doctor of Philosophy (PhD)
Electrical and Computer Engineering (Holcomb Dept. of)
Ganesh Kumar Venayagamoorthy
Johan H. Enslin
Nowadays, power grid operators are experiencing challenges and pressures to balance the interconnected grid frequency with rapidly increasing photovoltaic (PV) power penetration levels. PV sources are variable and intermittent. To mitigate the effect of this intermittency, power system frequency is regulated towards its security limits. Under aforementioned stressed regimes, frequency oscillations are inevitable, especially during disturbances and may lead to costly consequences as brownout or blackout. Hence, the power system operations need to be improved to make the appropriate decision in time. Specifically, concurrent or beforehand power system precise frequencies simplified straightforward-to-comprehend power system visualizations and cooperated well-performed automatic generation controls (AGC) for multiple areas are needed for operation centers to enhance.
The first study in this dissertation focuses on developing frequency prediction general structures for PV and phasor measurement units integrated electric grids to improve the situational awareness (SA) of the power system operation center in making normal and emergency decisions ahead of time. Thus, in this dissertation, a frequency situational intelligence (FSI) methodology capable of multi-bus type and multi-timescale prediction is presented based on the cellular computational network (CCN) structure with a multi-layer proception (MLP) and a generalized neuron (GN) algorithms. The results present that both CCMLPN and CCGNN can provide precise multi-timescale frequency predictions. Moreover, the CCGNN has a superior performance than the CCMLPN.
The second study of this dissertation is to improve the SA of the operation centers by developing the online visualization tool based on the synchronous generator vulnerability index (GVI) and the corresponding power system vulnerability index (SVI) considering dynamic PV penetration. The GVI and SVI are developed by the coherency grouping results of synchronous generator using K-Harmonic Means Clustering (KHMC) algorithm. Furthermore, the CCGNN based FSI method has been implemented for the online coherency grouping procedure to achieve a faster-than-real-time grouping performance.
Last but not the least, the multi-area AGCs under different PV integrated power system operating conditions are investigated on the multi-area multi-source interconnected testbed, especially with severe load disturbances. Furthermore, an onward asynchronous tuning method and a two-step (synchronous) tuning method utilizing particle swarm optimization algorithm are developed to refine the multi-area AGCs, which provide more opportunities for power system balancing authorities to interconnect freely and to utilize more PV power.
In summary, a number of methods for improving the interconnected power system situational intelligence for a high level of PV power penetration have been presented in this dissertation.
Wei, Yawei, "Situational Intelligence for Improving Power System Operations Under High Penetration of Photovoltaics" (2018). All Dissertations. 2564.