Optimized control of parallel heat pump units based on the satin bower bird algorithm in a distributed architecture

Compared to traditional shallow borehole heat exchanger (BHE), medium-deep borehole heat exchanger (MDBHE) exhibits significant advantages in heat extraction capability. Utilizing MDBHE as the heat source for heating systems can effectively achieve primary energy savings. However, in building heating applications, MDBHE is generally combined with heat pump units, which typically operate in parallel. The optimal load distribution of parallel heat pump units is a critical issue in MDBHE heating systems. Reasonable control to meet load demands is of great importance for the energy-efficient operation of the system. To address these issues, this study first establishes and compares optimization models of centralized and distributed systems targeting parallel heat pump units. The Distributed Satin Bowerbird Optimization (D-SBO) algorithm is proposed to address the load optimization distribution problem in parallel heat pump units within MDBHE. Experimental results confirm the robustness of the D-SBO algorithm, achieving up to a 23.1% improvement in COP. In case 1, the standard deviations of D-SBO range from 0.05 to 0.65 under a load demand of 40% to 90%. In case 2, the standard deviations range from 0.09 to 2.67 with a load demand of 70% to 90%. While D-SBO yields results comparable to DCSA, it demonstrates superior stability. Additionally, D-SBO provides significant energy savings, ranging from 3.90 kW to 140.38 kW in case 1 compared to GA, and from 1.00 kW to 165.00 kW in case 2 compared to GA and PSO.