关键词: 空压机负荷预测, 改进的哈里斯鹰优化算法, 深度回声状态网络, 超参数, 智能调度

Abstract: The air compressor system consists of multiple units, and the optimization of unit combination is a nonlinear and large-scale task of multiple objectives and constraints. To address such problems as high energy consumption and serious waste of resources in air compressor scheduling, this paper studies the quantity scheduling problem of air compressors based on the characteristics of air compressor combination. A multi-strategy improved Harris Hawk Optimization Algorithm (MHHO) combined with Deep Echo State Network (DESN) is proposed to predict the load of the air compressor. After obtaining the load required for 24 hours a day, the MHHO algorithm is employed for unit combination scheduling and gas consumption allocation. Our experimental results show the prediction model achieves a higher prediction accuracy, and thus is highly applicable for air compressor load prediction. Intelligent scheduling improves the unit operation efficiency and reduces the system’s energy consumption.