多气候区辐射 + 新风空调系统预测控制策略研究

Research on Predictive Control Strategy for Radiation + Fresh Air Conditioning System in Multi-Climate Zones

本研究使用变分模态分解算法优化神经网络预测模型(VMD-BP),将其与多目标粒子群算法(MOPSO)结合实现对辐射 + 新风空调系统控制参数的准确寻优。分别对比基于预测的多目标优化(F-MOPSO)控制策略在沈阳(严寒地区)、北京(寒冷地区)杭州(夏热冬冷地区)、广州(夏热冬暖地区)的节能潜力与室内空气温度的控制精度。通过仿真试验可知,VMD-BP预测模型相较于BP预测模型预测精度更高,F-MOPSO控制策略可实现更加精确的室温控制,在沈阳、北京、杭州和广州均可实现节能,且在广州节能量最高,可为辐射 + 新风空调系统在不同地区的应用提供指导。

In this study, a neural network prediction model (VMD-BP) is optimized using a variational modal decomposition algorithm combined with a multi-objective particle swarm algorithm (MOPSO) to achieve accurate optimization of the control parameters of a radiant + fresh air conditioning system. Compare energy saving potential and control accuracy of indoor air temperature of predic-tion-based multi-objective optimization (F-MOPSO) control strategy in Shenyang (severe cold zone), Beijing (cold zone), Hangzhou (hot summer and cold winter zone) and Guangzhou (hotsummer and warmwinter zone), respectively. Through simulation tests, it can be seen that the VMD-BP prediction model has higher prediction accuracy compared with the BP prediction model, and the F-MOPSO control strategy can achieve more accurate room temperature control, and energy saving can be realized in Shenyang, Beijing, Hangzhou and Guangzhou, and the energy saving is the highest in Guangzhou, so that it can provide a guide for the application of the radiant + fresh air air-conditioning system in different regions.