A finite-time projection neural network to solve the joint optimal dispatching problem of CHP and wind power

• Published in: Neural computing & applications Vol. 34; no. 10; pp. 7405 - 7417
• Main Authors: Wei, Boyu, He, Xing
• Format: Journal Article
• Language: English
• Published: London Springer London 01.05.2022; Springer Nature B.V
• Subjects: Algorithms; Artificial Intelligence; Carbon; Cogeneration; Computational Biology/Bioinformatics; Computational geometry; Computational Science and Engineering; Computer Science; Constraint modelling; Convergence; Convexity; Data Mining and Knowledge Discovery; Heat storage; Image Processing and Computer Vision; Liapunov functions; Neural networks; Operating costs; Optimization; Original Article; Parameters; Probability and Statistics in Computer Science; Projection; Scheduling; Transmission loss; Wind power; Wind power generation; Finite-time projection neural network; Non-convex optimization; Microgrid; Combined heat and power with wind power; Convergence speed
• ISSN: 0941-0643; 1433-3058
• DOI: 10.1007/s00521-021-06867-x

This paper constructs an optimal scheduling model for combined heat and power generation units with heat storage and wind power generation considering carbon transaction costs and optimizes the output of each unit to reduce wind curtailment rate, carbon emissions, and total operating costs. In the case of considering transmission loss, the optimal scheduling model subject to actual operation constraints is expressed as a non-convex optimization problem. Based on a sufficient condition, the equivalent problem transformed from the original optimization problem is expressed as a convex problem, and a finite-time reduced-dimensional projection neural network with time-varying parameters is proposed to solve the problem. The proposed neural network can reach convergence in a limited time. The Lyapunov function is used to prove the convergence of the algorithm. Finally, the effectiveness of the designed neural network is verified by numerical simulation. Compared with reduced-dimensional projection neural network and finite-time fixed-parameter reduced-dimensional projection neural network, our neural network has a faster convergence speed.