关键词: 风电拉挤板, 生产排程, 多目标优化, 模糊交货期, 小龙虾算法

Abstract: Wind power is a cost-effective, clean and mature energy source, offering a sustainable alternative to fossil fuels. Wind turbine extrusion plates, critical structural components of turbine blades, are manufactured through a continuous, order-driven and flexible production process to meet diverse market needs. However, the Wind Turbine Extrusion Plate Production Scheduling Problem (PSP-WTEP) is highly complex and falls into the NP-hard category, as it must consider constraints such as mold replacements, sequential flush sets and operational limitations. PSP-WTEP can be divided into three sub-problems: assigning production areas for orders, allocating processing equipment within those areas and sequencing components on machines. These challenges require a multi-objective optimization model to balance equipment load, maximize customer satisfaction, and minimize mold changeover times, all while adhering to fuzzy delivery deadlines. Effectively solving PSP-WTEP is essential for improving production efficiency, reducing costs and promoting innovation in the wind power industry.
To tackle this issue, the study develops the Enhanced Multi-Objective CrayfishOptimization Algorithm (EMOCOA), an improvement over the Crayfish Optimization Algorithm (COA). EMOCOA incorporates advanced strategies such as non-dominated sorting, refractive inverse learning for population initialization, nonlinear convergence factors, sub-swing perturbations and dynamic step adjustments. The algorithm employs a two-layer coding system for components and regions, converting continuous variables into discrete codes using Ranked Order Value (ROV) encoding. It adopts a “backward-repair-optimize” decoding strategy to resolve fuzzy delivery constraints and improve solution feasibility. These enhancements allow EMOCOA to effectively balance global exploration and local exploitation, handle the coupling of PSP-WTEP sub-problems, and deliver diverse and high-quality solutions that meet production goals. The algorithm’s hybrid learning strategies ensure fast convergence and robust performance, even in complex scenarios.
Extensive experiments highlight the superiority of EMOCOA over the Multi-Objective Crayfish Optimization Algorithm (MOCOA) and Multi-Objective Whale Optimization Algorithm (MOWOA). EMOCOA consistently outperforms its counterparts, generating 10 Pareto solutions compared to 7 by MOCOA and 8 by MOWOA, with better results on key evaluation metrics such as IGD and HV. These metrics indicate that EMOCOA provides improved solution diversity and convergence, making it more practical for real-world application. Production managers can use methods such as hierarchical analysis and entropy weight to rank and select Pareto solutions for implementation. By solving various scenarios and demonstrating its adaptability, EMOCOA proves to be a powerful tool for optimizing production scheduling, reducing costs, and increasing efficiency. The algorithm supports intelligent upgrades in wind turbine extrusion plates manufacturing, facilitating the sustainable growth and technological advancement of the wind power industry.

Key words: wind turbine extrusion plate, production scheduling, multi-objective optimization, fuzzy lead times, crayfish optimization algorithm