Analysis of Voltage Stability Indices through RTDSCopyright: RWTH Aachen
A widely accepted definition of voltage stability is the ability of the power system to maintain steady acceptable voltages at all buses in the system under normal operating conditions and after being subjected to a disturbance. The quantitative assessment of how close the system is from a voltage instability is critical for timely intervention. The information on the proximity can determined and formulated through Voltage Stability Indices. At ACS we have investigated the performance of different line voltage stability indices previously presented in literature, including Lmn Index, Fast Voltage Stability Index (FVSI), Voltage Collapse Point Indicators (VCPI), and LQP Index, as well as the traditional Jacobian index based on the minimum eigenvalue of the Jacobian matrix. These indices were tested on a small 5-bus system and on the IEEE 39-bus system. Those feature the characteristics of classical power systems. The simulation tool used was RTDS and the indices where computed using the control blocks components in order to monitor the values in real time. This method was chosen to have the indices values available for further future work on control algorithm development. All the indices were found consistent with their theoretical background and the performance comparison was based on three characteristics: accuracy, robustness to uncertainty, and adequacy for control purposes. We also analyzed how dependable the calculated value is. To assess this characteristic we used the last index value before voltage collapse occurs, the increase of value at the voltage collapse point and linearity in the stable region and the first index value after voltage collapse. This resulted in a ranking where the best index was VCPI (p), followed by LQP, Lmn and VCPI (2)(l). Another aspect that was analyzed, is how the index value while approaching the instability point is dependent upon the nature of the apparent power increase, whether the increase was from both active and reactive, or if it was just due to the increase of reactive power. This should be understood as how much does the index tells about the stability of the system by itself, without viewing any other measurements of the system like reactive power increase at the loads. From the results and based on these comparison characteristics, VCPI (p) was found to have the best performance from the indices studied.
We plan to move the index calculation away from the control toolbox into own components created with the CBuilder; as this would enable us to include mathematical more complex indices as the Jacobian index in RTDS directly. Another step for us is to implement the PMU based indices, this effort is done in the work of the distributed intelligence platform which comprises the PMU emulator. We also plan to replace the generators of the test systems with renewable energy sources. Another major step consists in closing the loop actually using the indices for control and hence evaluating their performance.