Search

Pulsewidth-modulated (PWM) inverters are used more and more to operate electrical machines and to interface renewable energy systems With the utility grid. However, there are abundant high-frequency harmonics in the output voltage of a PWM inverter, which increase the iron losses and result in derating of the machine or transformer connected to them. Predicting the iron losses caused by the PWM supply is critical for the design of electrical machines and transformers operated by PWM inverters. These losses are primarily attributed to eddy-current loss caused by the PWM Supply. In this paper, after analyzing the harmonic components of PWM voltage, we derive the effects of different parameters of PWM switching on the eddy-current loss. We compare the iron losses modeled with the proposed analytical methods on a three-phase transformer, a dc motor, and an induction motor with the results of time-stepping finite-element analysis and experiments. We provide detailed equations for the prediction of iron losses. These equations can be directly applied in the design and control of PWM converters and electric motors to improve energy efficiency in electrical machines and transformers operated from PWM converters. [ABSTRACT FROM AUTHOR]

• Regional graph attention is used to efficiently capture neighborhood features. • Value decomposition structure is applied for the training of multi-agents. • Proposed distributed dispatching method learns the more effective strategy. • Proposed algorithm is adaptive to network topology change and time granularity scale. In this article, an adaptive active power rolling dispatch strategy based on distributed deep reinforcement learning is proposed to deal with the uncertainty of high-proportioned renewable energy. For each agent, by using recurrent neural network layers and graph attention layers in its network structure, we aim to improve the generalization ability of the multiple agents in active power flow control. Furthermore, a regional graph attention network algorithm, which can effectively help agents aggregate the regional information of their neighborhood, is proposed to improve the information capture ability of agents. We adopt the structure of 'centralized training, distributed execution' to help agents improve the effectiveness of proposed methods in dynamic environments. The case studies demonstrate that the proposed algorithm can help multi-agents learn effective active power control strategies. Each agent has a strong generalization ability in terms of time granularity and network topology. We expect that such an approach can improve the practicability and adaptability of distributed AI method on power system control issues. [ABSTRACT FROM AUTHOR]