This paper presents DC fault analysis for a dual active bridge DC/DC converter which comprises of two active bridges and an internal medium frequency transformer. This topology provides galvanic isolation, voltage step up/down and bidirectional power transfer. The DC fault study assumes the DC terminal voltage of the converter is at zero. The steady state fault current is limited to a low magnitude which is less than the rated value without any controller action depending on the design of the converter. The DC faults current magnitudes are analyzed with AC equivalent circuit where only the fundamental component of the AC voltages and inductor current are considered. Phase shift and AC voltage magnitude modulation control methods are selected. A detailed dual active bridge DC/DC converter 3 MW, 4/40 kV based system is simulated using MATLAB/Simulink to validate the proposed analytical study. Further, hardware testing is conducted to confirm the DC fault studies with a 500 W 24/100 V prototype.
11 January 2022 in Education and Information Technologies
In classical power systems, frequency measurements are transferred via a specialised communication channel, resulting in time delay. The time delay plays a major role in a power system, which can reduce the dynamic performance of the load–frequency control (LFC) system and can destabilise the system. The research to date has tended to focus on developing a new algorithm to determine the delay margin (DM) rather than looking into a hybrid algorithm which includes a nature-inspired metaheuristic optimisation technique. This paper introduces a novel method for computing the DM based on grey wolf optimisation (GWO), specifically for the constant time delay. In the proposed method, GWO is employed to optimise the minimum error of the spectral radius and to determine the best design variable of the crossing frequency. With the help of the proposed method, the sweeping range is no longer required, which improves the accuracy of the result. To evaluate the proposed method, a two-area network power system is considered as a case study. Furthermore, the effect of the PI controller gains on the DM is taken into account. The proposed method efficacy is demonstrated by comparing it with the most recently published methods. The results demonstrate that the proposed method is remarkably better than the existing methods found in the literature, where the smallest percentage inaccuracy using the simulation-based DM based on GWO is found to be 0.000%.