当前位置:
X-MOL 学术
›
Energy Build.
›
论文详情
Our official English website, www.x-mol.net, welcomes your
feedback! (Note: you will need to create a separate account there.)
A learning-based model predictive control method for unlocking the potential of building energy flexibility
Energy and Buildings ( IF 6.6 ) Pub Date : 2025-01-12 , DOI: 10.1016/j.enbuild.2025.115299
Jie Zhu, Jide Niu, Sicheng Zhan, Zhe Tian, Adrian Chong, Huilong Wang, Haizhu Zhou
Energy and Buildings ( IF 6.6 ) Pub Date : 2025-01-12 , DOI: 10.1016/j.enbuild.2025.115299
Jie Zhu, Jide Niu, Sicheng Zhan, Zhe Tian, Adrian Chong, Huilong Wang, Haizhu Zhou
Buildings are regarded as promising energy flexibility resources due to their significant energy consumption and better integration into the electricity grid. To fully exploit the potential of building flexibility, optimized operational strategies need to be developed in which cost savings and thermal comfort should be considered. Model predictive control (MPC) is widely acknowledged as one of the effective methods for developing optimal strategies. However, the practical implementation of traditional MPC is often hindered by substantial computational burdens. Specifically, MPC demands intensive software and hardware resources due to its reliance on real-time iterative optimization at each control interval. Moreover, the complexity of control tasks related to energy flexibility, along with the demand for large-scale implementation of MPC within buildings, makes this issue more pronounced. Therefore, this study proposes a learning-based model predictive control (LBMPC) method employing machine learning models to learn and imitate MPC behavior from a dataset containing optimal control trajectories. This method eliminates the need for online optimization, significantly reducing dependency on computational resources. The proposed method consists of three parts: identifying a control-oriented building thermal model to formulate the MPC problem, conducting offline simulations to generate the training dataset, and training classification and regression trees (CART) to learn optimal control actions from dataset. The control performance of the proposed method in a multi-zone office building was evaluated using a high-fidelity co-simulation testbed constructed with Modelica and Spawn of EnergyPlus. The results indicate that, compared to the baseline control strategy, traditional MPC and LBMPC reduced energy costs by 13.89% and 12.27%, respectively, and peak electrical loads by 30.35% and 24.78%, respectively, without compromising thermal comfort. Especially, compared to traditional MPC, LBMPC can significantly reduce the computational cost by as much as 99.6%, with only a small trade-off in performance. Besides, the impacts of input features, meteorological conditions, and model accuracy on the control performance of the proposed method are discussed in detail.
中文翻译:
一种基于学习的模型预测控制方法,用于释放建筑能源灵活性的潜力
建筑物被认为是有前途的能源灵活性资源,因为它们的能源消耗量很大,并且能更好地融入电网。为了充分利用建筑灵活性的潜力,需要制定优化的运营策略,其中应考虑成本节约和热舒适性。模型预测控制 (MPC) 被广泛认为是制定最佳策略的有效方法之一。然而,传统 MPC 的实际实现经常受到大量计算负担的阻碍。具体来说,MPC 需要大量的软件和硬件资源,因为它依赖于每个控制间隔的实时迭代优化。此外,与能源灵活性相关的控制任务的复杂性,以及在建筑物内大规模实施 MPC 的需求,使这个问题更加突出。因此,本研究提出了一种基于学习的模型预测控制 (LBMPC) 方法,该方法采用机器学习模型从包含最佳控制轨迹的数据集中学习和模仿 MPC 行为。这种方法消除了在线优化的需要,大大减少了对计算资源的依赖。所提出的方法包括三部分:确定面向控制的建筑热模型以制定 MPC 问题,进行离线仿真以生成训练数据集,以及训练分类和回归树 (CART) 以从数据集中学习最佳控制动作。使用由 Modelica 和 Spawn of EnergyPlus 构建的高保真协同仿真测试平台评估了所提方法在多区域办公楼中的控制性能。 结果表明,与基线控制策略相比,传统 MPC 和 LBMPC 在不影响热舒适性的情况下,能源成本分别降低了 13.89% 和 12.27%,峰值电力负载分别降低了 30.35% 和 24.78%。特别是,与传统 MPC 相比,LBMPC 可以显著降低高达 99.6% 的计算成本,而性能上的牺牲很小。此外,详细讨论了输入特征、气象条件和模型精度对所提方法控制性能的影响。
更新日期:2025-01-12
中文翻译:

一种基于学习的模型预测控制方法,用于释放建筑能源灵活性的潜力
建筑物被认为是有前途的能源灵活性资源,因为它们的能源消耗量很大,并且能更好地融入电网。为了充分利用建筑灵活性的潜力,需要制定优化的运营策略,其中应考虑成本节约和热舒适性。模型预测控制 (MPC) 被广泛认为是制定最佳策略的有效方法之一。然而,传统 MPC 的实际实现经常受到大量计算负担的阻碍。具体来说,MPC 需要大量的软件和硬件资源,因为它依赖于每个控制间隔的实时迭代优化。此外,与能源灵活性相关的控制任务的复杂性,以及在建筑物内大规模实施 MPC 的需求,使这个问题更加突出。因此,本研究提出了一种基于学习的模型预测控制 (LBMPC) 方法,该方法采用机器学习模型从包含最佳控制轨迹的数据集中学习和模仿 MPC 行为。这种方法消除了在线优化的需要,大大减少了对计算资源的依赖。所提出的方法包括三部分:确定面向控制的建筑热模型以制定 MPC 问题,进行离线仿真以生成训练数据集,以及训练分类和回归树 (CART) 以从数据集中学习最佳控制动作。使用由 Modelica 和 Spawn of EnergyPlus 构建的高保真协同仿真测试平台评估了所提方法在多区域办公楼中的控制性能。 结果表明,与基线控制策略相比,传统 MPC 和 LBMPC 在不影响热舒适性的情况下,能源成本分别降低了 13.89% 和 12.27%,峰值电力负载分别降低了 30.35% 和 24.78%。特别是,与传统 MPC 相比,LBMPC 可以显著降低高达 99.6% 的计算成本,而性能上的牺牲很小。此外,详细讨论了输入特征、气象条件和模型精度对所提方法控制性能的影响。