舒适优先多模态感知驱动的智能玻璃自适应控制方法

Adaptive Control Method of Smart Glass Driven by Multimodal Sensing with Comfort Priority

摘要: 在全球变暖与能源紧缺背景下，建筑空调系统能耗占比持续攀升，大面积玻璃结构引起的热交换损失进一步加剧了这一问题。本文系统梳理了智能玻璃控制领域的主要技术流派，指出现有方法在热舒适建模精度与多模态环境自适应能力方面的不足，进而提出一种面向绿色建筑的舒适优先、多模态感知驱动的智能玻璃自适应控制方法。系统基于低功耗ESP32搭建物联网终端，集成传感器实时采集室内外环境参数；依托ASHRAE热舒适数据库，采用随机森林建立热舒适参考标准，并利用支持向量机构建从环境状态到控制行为的映射，形成以舒适度偏差为驱动的闭环控制策略。实验结果表明，该方法使室内全年热舒适时长增加了50 h，平均采光系数提高了2.03%，全年综合能耗减少了423.387 kWh；与普通玻璃窗相比，空调冷负荷显著降低，建筑整体能耗降低约20%。该方法能够动态响应外界环境变化，避免机电设备过度补偿所造成的能量浪费，使建筑冷热负荷需求降低，验证了其在智能玻璃结构中的节能潜力与应用价值。

Abstract: Against the backdrop of global warming and energy shortage, the energy consumption proportion of building air-conditioning systems keeps rising, and heat exchange losses induced by large-area glass structures further aggravate this issue. This paper systematically reviews the main technical schools in the field of smart glass control, identifies the deficiencies of existing methods in thermal comfort modeling accuracy and multimodal environmental adaptability, and thereby proposes an adaptive control method of smart glass driven by multimodal sensing with comfort priority for green buildings. The Internet of Things (IoT) terminal is built based on low-power ESP32, with integrated sensors to collect indoor and outdoor environmental parameters in real time. Relying on the ASHRAE thermal comfort database, random forest is adopted to establish the thermal comfort reference standard, and support vector machine is used to construct the mapping from environmental states to control actions, forming a closed-loop control strategy driven by comfort deviation. Experimental results demonstrate that the proposed method increases the annual indoor thermal comfort duration by 50 h, raises the average daylight factor by 2.03%, and cuts the annual total energy consumption by 423.387 kWh. Compared with conventional glass windows, the air-conditioning cooling load is remarkably reduced, and the overall building energy consumption drops by approximately 20%. The proposed method can dynamically respond to external environmental variations and avoid energy waste caused by excessive compensation of electromechanical equipment, thus lowering the building heating and cooling load demands. Its energy-saving potential and application value in smart glass structures are verified.