Research on Coordinated Optimization of Smart Agriculture Microgrid Configuration and Production Energy Efficiency Driven by Multi-objective Evolutionary Algorithm

Abstract

The smart agriculture park is integrating energy issues into the production process. Electricity is no longer an external input but a continuous condition that maintains the environment for crops. Existing research on microgrids mostly focuses on solving costs, carbon emissions, and reliability, while agricultural loads are often compressed into an exogenous demand curve, making it difficult to identify the transmission of supply shortages to yield and quality. This study establishes a "source-storageload-production" coupled model, placing capacity selection, grid connection boundaries, and agricultural loads in the same code; the target system also examines economic efficiency, emission constraints, supply shortage risks, and production energy efficiency. Based on the results of continuous observations in a smart agriculture park in North China Plain in 2024, the carbon constraint collaborative optimization scheme reduced the total annualized cost by 20.89% compared to the grid benchmark; the carbon emission reduction was 72.18%; unit output energy consumption decreased by 27.23%; production energy efficiency increased by 37.45%, and the critical load guarantee rate reached 99.7%. Compared to five benchmark algorithms, PEC-MOEA achieved higher HV and lower IGD, and had a higher proportion of feasible solutions. The study shows that the core of agricultural microgrid planning is not merely about increasing low-carbon installed capacity, but about integrating energy timing, production tasks, and reliability baselines into the same set of constraints.