2026 International Conference on Sensing, Measurement & Data Analytics in the era of Artificial Intelligence (ICSMD 2026)

To address dispersed wind-farm operation and maintenance data, poor traceability of spare-part demand, and the conflict between inventory cost and supply assurance, this paper proposes a knowledge-driven multi-objective inventory optimization method. A knowledge graph organizes systems, faults/warnings, corrective actions, spare parts, and suppliers, while spare-part identifiers link replacement, inventory, and procurement records. Purchase quantities are formulated as integer decision variables in a bi-objective inventory-cost–stockout-risk model and solved by NSGA-II with integer and warehouse-capacity repair. Using real data from an onshore wind farm, NSGA-II is compared with MOEA/D and MOPSO under identical evaluation budgets and constraints. NSGA-II maintains comparable solution-set quality while substantially reducing runtime, and stably produces cost-priority, balanced, and assurance-priority plans. Field validation shows that the balanced plan improves periodic-demand coverage and lowers procurement cost while satisfying demand and safety-stock requirements.

wind farm,spare parts,knowledge graph,multi-objective inventory optimization,NSGA-II