Peak Arm Current Reduction of Modular Multilevel Converter Based on Particle Swarm Optimizer for Circulating Current Injections

Modular multilevel converters (MMCs) are widely used, and ongoing research aims to enhance and optimize their performance. One key improvement is reducing the peak arm current, which helps boost the current capability of the submodule switches and increase power handling capacity. One approach to achieve this target is to optimally shape the circulating current with a specified amount of AC components, which can be injected into the arm current. Circulating current arises from the mismatch between submodule capacitor voltages and the DC terminals. This paper uses a particle swarm optimisation (PSO) algorithm to determine the optimal values for the injected components. The magnitude and phase of the injected components are treated as variables in the optimisation process, with the goal of minimising the peak arm current as the cost function. The paper also investigates how these optimal injections affect peak-to-peak energy deviations and capacitor voltage ripples. Moreover, the response of the proposed method is investigated under multiple dynamic conditions, including a varying power factor, dynamically changing power references, and grid voltage sags, as the MMC interacts with the grid by exchanging real and reactive power. Results show that the proposed method can reduce the peak arm current by up to 27.42% for specific operating conditions. Additionally, the proposed approach performs better in reducing the peak arm current when the MMC supplies purely reactive current. In contrast, other methods may tend to be unproductive in these areas. An evaluation of losses, based on datasheet information of submodule switches, shows that the effect of the PSO-based optimal injections is marginal to the extent that no significant change is reported in converters’ efficiency compared to the case of zero injections.