In the transient stability analysis,the grid-forming converter considering current limiting will appear as a switching system,and its switching mechanism is still unclear.Aiming at a dual power synchronization loop(PSL)system considering current limiting,where the current limiting measure uses a current saturation algorithm.The switching boundaries of the dual-machine system in terms of phase angle are derived.It is found that the dual-machine system considering current limiting has four working states and four switching boundaries within a phase angle period,and the positions of two switching boundaries are related to the saturation current angle settings of the two converters.Based on this,the transient stability of the studied system is further analyzed,obtaining the stable equilibrium points of the four working states.The stable equilibrium points are configured by changing the current saturation angle set values of the two converters,which can avoid the situation that one converter is locked in the current limiting control mode.The simula-tion verifies the correctness and effectiveness of the proposed theory.
A dual-loop current control is discussed for LCL -type shunt active power filters (APFs), where inverter-side current feedback and point of common coupling (PCC) voltage feedforward both provide active damping (AD) effects for LCL resonance. However, two ADs are equivalently applied to different elements of LCL filter with narrow damping boundaries, which causes less robustness against grid impedance variation. Hence, this paper has presented a comprehensive approach to strengthen the system robustness against grid impedance variation. On the basis of phase compensated resonant unit, a delay-compensation scheme is firstly proposed to extend the damping boundary of inverter-side current feedback up to almost Nyquist frequency ( $f_{s}/2$ ). Theory analysis considering PCC voltage feedforward indicates that the control system can keep robust to grid impedance when the initial LCL resonance frequency falls between $f_{s}/3$ and the extended damping boundary. Moreover, a typical weighted proportional-resonant feedforward scheme is suggested to prevent high-frequency resonant units from inducing system instability under a large grid impedance. We finally validate the effectiveness of the proposed approach through experiments.
Under inverter operating conditions, significant loss is incurred by the insulated gate bipolar transistor (IGBT) (referred to as T2 tube) in the lower part of the half-bridge submodule. The reduction of loss is beneficial for the improvement of equipment operation reliability. At the same time, the suppression of capacitor ripple voltage has the advantage for reducing capacitor demand and enhancing power density. However, attention is not given by existing optimization control strategies to the contradiction between loss distribution optimization and capacitor ripple voltage, making it difficult to balance equipment operation reliability and power density. Therefore, a comprehensive optimization method that combines the reduction of T2 transistor losses and the suppression of capacitor voltage ripple is proposed in this article. Firstly, the inherent contradiction between reducing the on-state loss of the T2 transistor and suppressing capacitor voltage ripple is explained by analyzing the impact path of charge on device loss and capacitor ripple voltage. Then, by introducing a penalty function, a comprehensive objective function is established that takes into account T2 transistor losses and capacitor voltage ripple. Subsequently, using the active bypass strategy as an example, a comprehensive optimization method based on the injection of second harmonic current and third harmonic voltage is proposed by analyzing the impact of the second harmonic current and third harmonic voltage injection on T2 transistor loss and capacitor voltage ripple. Finally, a simulation model is built in MATLAB/Simulink and PLECS for verification. The simulation results suggest that the reliability and power density of the device increase by the comprehensive optimization method, considering both T2 transistor losses and capacitor voltage ripple.
With the proposal of the “dual-carbon” target and the rapid development of distributed generation technology, distribution areas (DAs) are faced with the problems of difficult to expand capacity, difficult to consume distributed generation, unable to effectively use the remaining capacity of each DA, etc. The flexible interconnection system (FIS) has gradually become a solution to improve the power supply level of DAs. To fully utilize the advantages of the flexible power flow control of FIS, a $v_{\mathrm {{dc}}}-P_{\mathrm {{grid}}}$ droop control strategy is proposed, which takes into account load balancing. This strategy balances the AC source output of each DA during normal operation. Subsequently, a simplified circuit model of the system is developed, and the stability of the system is analyzed under large disturbances when a constant power load is connected to the DC bus, using the mixed potential function theory. The stability criteria of the system is obtained, and verified by the real-time hardware-in-loop (HIL) tests.
In deep learning, the load data with non-temporal factors are difficult to process by sequence models. This problem results in insufficient precision of the prediction. Therefore, a short-term load forecasting method based on convolutional neural network (CNN), self-attention encoder-decoder network (SAEDN) and residual-refinement (Res) is proposed. In this method, feature extraction module is composed of a two-dimensional convolutional neural network, which is used to mine the local correlation between data and obtain high-dimensional data features. The initial load fore-casting module consists of a self-attention encoder-decoder network and a feedforward neural network (FFN). The module utilizes self-attention mechanisms to encode high-dimensional features. This operation can obtain the global correlation between data. Therefore, the model is able to retain important information based on the coupling relationship between the data in data mixed with non-time series factors. Then, self-attention decoding is per-formed and the feedforward neural network is used to regression initial load. This paper introduces the residual mechanism to build the load optimization module. The module generates residual load values to optimize the initial load. The simulation results show that the proposed load forecasting method has advantages in terms of prediction accuracy and prediction stability.
