6-phase DFIG for wind energy conversion system: A hybrid approach. (October 2022)

• Record Type: Journal Article
• Authors: Chellaswamy, C.; Geetha, T.S.; Thiruvalar Selvan, P.; Arunkumar, A.
• Is Part Of: Sustainable energy technologies and assessments. Volume 53:Part B(2022)
• Publication Date: 2022-10
• Subjects: Quantum process -- 6-phase DFIG -- Wind energy conversion system -- Deep reinforcement learning algorithm -- Hybrid approach; QPDRL quantum process and deep reinforcement learning -- DFIG doubly-fed induction generator -- PID proportional integral derivative -- THD total harmonic distorsion -- WECS wind energy conversion systems -- AIA artificial intelligence algorithms -- TAM traditional analytical methods -- CAA conventional analytic algorithms -- FO-PID fractional-order PID -- RL reinforcement learning -- DRL deep reinforcement learning -- MES multi-energy scheme -- GGWO grouped grey wolf optimization -- ANN artificial neural network -- DBNs deep belief networks -- RLA reinforcement learning algorithm -- DFIG-WTs DFIG-based wind turbines -- PSO particle swarm optimization -- GA genetic algorithm -- FOPI fractional order proportional integral -- PI proportional integral; Renewable energy sources -- Periodicals; Energy development -- Technological innovations -- Periodicals; Electric power production -- Periodicals; Energy storage -- Periodicals; 333.79
• DOI: 10.1016/j.seta.2022.102497
• Languages: English
• ISSNs: 2213-1388

Abstract

Highlights: A hybrid technique is proposed for a six-phase DFIG-based wind energy system. Deep reinforcement learning and quantum processes are combined together to form control strategies. It predicts the next systemic state and avoids local optimal solutions. The performance of the proposed method is tested using three different statistical indices. Good control performance on machine side and grid side converters is effectively obtained. Abstract: This research presents a novel wind power system based on a six-phase doubly-fed induction generator (DFIG). Optimization approaches are required to improve the efficiency of the traditional controllers. This study introduces a blended method for DFIG-based wind power transformation systems that combines quantum process and deep reinforcement learning (QPDRL) to improve control efficiency. It will be driven by using online control algorithms to eliminate the optimizing step and upgrade online control strategies. The proposed QPDRL can prevent local optimum solutions, forecast the future essential phase, and update DFIG-based wind power plants' regulation methods online. For two distinct scenarios, the QPDRL was contrasted with the proportional integral derivative (PID) controller, fractional-order PID, and reinforcement learning (for changeable air velocity, there are two types of arbitrary and step amplitudes). Matlab software was used to experiment. As air velocity variations exist, the findings revealed a 62% reduction in the DC link voltage ripples and a 99% reduction in speed overshoot with wind velocities overrun. Finally, comparing PID controls revealed a 42.15 percent reduction in grid current THD and an 11.38 percent reduction in the generator current.