A Wind Energy Prediction Scheme Combining Cauchy Variation and Reverse Learning Strategy

Abstract

Modular multilevel converters (MMCs) can be a reliable solution since they have modular structure and high quality output waveform for permanent magnet synchronous generator (PMSG) based wind energy conversion system (WECS). Capacitor voltage balancing in nearest level modulation (NLM) is required to keep the capacitor voltage of each submodule of MMC constant. In this paper, an efficient capacitor voltage balancing scheme under NLM is proposed for PMSG based WECS with MMC topology. Through proposed control scheme, arm voltages are separately controlled and voltage ripple of around 1.5% is obtained. This result provides high quality output waveform at the point of common coupling (PCC). Furthermore, DC-link voltage control is achieved via hysteresis current control based proportional-integral controller. The ripple of DC-link voltage is obtained quite well as nearly 0.25%. In addition, load voltage control is accomplished using dq reference frame-based voltage control scheme for voltage and frequency stabilization at the PCC by regulating the voltage at its reference value. Simulation studies show that all proposed control schemes give satisfactory results for MMC based WECS under variable dynamic operation modes. Finally, experimental verification is performed using laboratory prototype to show the applicability of the proposed capacitor voltage balancing scheme.

Keywords

• carbon emissions
• cauchy mutation
• long short-term memory
• reverse learning
• synchrosqueezed wavelet transforms