Wind power ramp event is a small probability and large hazard event, for effectively avoiding its impact on the grid, an offshore wind prediction method based optimal combination model proposed. First, basis of output data existing turbines, preprocessing series carried out by using Variational Modal Decomposition optimized novel Seagull Optimization Algorithm, which decomposed into several Intrinsic Mode Function components one residual component to extract salient features signal reduce...

Abstract STATCOM can be used as both dynamic reactive power compensation device and active filter to improve the quality of a grid. This paper describes function implemented into an existing control. The explains firstly filtering control system STATCOM, starting from introduction principle, harmonic current detection tracking After that, parameter design, open-loop characterization simulation verification proposed scheme are carried out. PSCAD results show that in this effectively reduce...

Abstract Under the “dual carbon” strategy, frequent wide-band grid oscillations caused by grid-following converters as interface for renewable energy integration have become common. Existing studies suggest that addition of grid-forming can address this issue, but few explain reasons from a stability mechanism perspective. Additionally, most overlook interaction between and sending-end converter stations. This paper first conducts sequential impedance modeling. Then, based on network...

In order to realize the collection and long-distance transmission of large-scale renewable energy in "Three North" southwest regions China, two schemes through series hybrid DC system are proposed, including island grid connection transmission. The adopts structure line commutated converter (LCC) modular multilevel (MMC), which can make full use advantages LCC MMC. paper presents operation control strategy under modes, designs startup different modes at sending end. Finally, a simulation...

A method of offshore wind power prediction based on VMD and LSTM is proposed to improve the accuracy prediction. For reducing volatility data extract salient features signal, a variational modal decomposition algorithm optimized by seagull used decompose series into several eigenmodal components residual component that may contain uncertainties affecting prediction; decomposed are built model long short-term memory network, extreme learning machine. The final obtained superimposing...

To solve the problems of complex control structure, too many parameters to be tuned and complicated modulation module design in traditional dual closed-loop strategy neutral point clamped (NPC) three-level PV inverter. A is proposed for NPC inverters based on multi-step-ahead finite set model predictive (FCS-MPC). Based derivation mathematical inverter, current inverter αβ frame realized basic principle FCS-MPC, which effectively reduces Additionally, inherent advantages mid-point potential...

Two sets of 150Mvar cascaded multilevel STATCOM equipment are proposed for a substation the South Grid transmission system to enhance dynamic reactive power support capability grid. In this paper, design control is discussed, including steady-state control, voltage transient and blocking coordinated control. The switching conditions these modes additionally determined. To verify that in paper can realize mode while having sufficient fault ride-through recovery supporting capabilities, PSCAD...

In order to improve the fault accuracy and efficiency of fan bearings, a bearing intelligent diagnosis method based on combination particle filter sparse encoder is proposed. The vibration signal preprocessed with filter, then features are extracted separately construct feature matrix; in prevent network structure from overfitting, input into autoencoder added Dropout for classification. Finally, soft-max classifier used classification recognition, compared traditional method. Experimental...

The wind power converter is very prone to failure in the permanent magnet direct-drive system. For healthy operation of system, a fault diagnosis method based on improved wavelet and random forest proposed. In Simulink platform, model system built simulate normal states obtain signals. lifting scheme used construct wavelet, genetic algorithm optimize optimal wavelet. Then relative energy, waveform index, margin index peak are calculated by detail coefficients decomposition. feature vector...