Human-centered design optimization for noise control technologies in built environments

噪声通过影响建筑物居住者的舒适性和性能，对空间质量产生重大影响。噪声控制领域中仅试图降低物理噪声水平的传统方法具有某些固有的局限性。这些方法往往忽视居住者的特征和行为。如果不仔细考虑空间中的乘员模式，在无人居住的条件下量化声学质量的努力可能无法满足用户的声学要求。因此，包含乘员舒适性的优化噪音控制产品设计是一项尚未得到很好确立的重要知识。因此，拟议研究计划的长期目标是通过提供以人为中心的声学设计优化方法来改善建筑空间的声学舒适性。短期目标是：(A)开发占用教室活动噪声水平的自动检测方法；(B)优化微穿孔嵌板消音器，以消除暖通空调音调噪声；(C)优化占用开放式办公室的隔音系统。为了实现研究提案中概述的目标，将进行一系列声学测量、数据分析、数值模拟和主观测试。首先，将收集已占用空间的声学数据，以获得每种空间类型的占用模式。建筑空间的准确背景噪声频谱将可用于预测声学活动。无监督机器学习算法，如高斯混合模型，将被用来对背景噪声水平进行分类。教室的语音清晰度、开放式办公室的隔音系统和暖通空调系统的管道消声器将通过识别关键部件进行优化。最后，基于降噪技术优化的乘员舒适性改善将在虚拟现实实验室中得到验证。本研究的结果将为室内空间噪声控制技术的优化设计提供一种方法。这项拟议研究的成功完成，可望提高管道消声器和隔音系统设计的成本效益。研究结果也可作为可持续建筑的循证设计方法。这些努力的结果将是实现建筑空间的最佳音质，从而改善居住者的舒适性和在建筑环境中的表现。

Noise has a significant effect on the quality of space by influencing building occupant comfort and performance. Traditional approaches in the noise control area that attempt only to reduce physical noise levels have certain inherent limitations. These approaches tend to overlook occupant characteristics and behaviours. Without careful consideration of occupant patterns in spaces, efforts to quantify acoustic quality in unoccupied conditions can fail to meet the acoustic requirements for users. An optimized noise control product design incorporating occupant comfort is thus a significant piece of knowledge that has not yet been well-established. The long-term objective of the proposed research program is, therefore, to improve acoustic comfort in building spaces by providing a human-centered acoustic design optimization methodology. The short-term objectives are: (a) to develop an automatic activity noise level detection method in occupied classrooms, (b) to optimize a microperforated panel-embedded silencer for tonal HVAC noise, and (c) to optimize a sound masking system for occupied open-plan offices. To achieve the objectives outlined in the research proposal, a series of acoustic measurements, data analyses, numerical simulations, and subjective tests will be conducted. First, acoustic data in occupied spaces will be collected to obtain occupancy patterns for each space type. Accurate background noise spectra in building spaces will be made available for predicting acoustic activities. Unsupervised machine learning algorithms such as a Gaussian mixture model will be used to classify the background noise levels. The speech intelligibility for classrooms, sound masking system for open-plan offices, and a duct silencer for HVAC systems will be optimized by identifying crucial components. Finally, improved occupant comfort based on optimization of the noise reduction technologies will be validated in a virtual reality laboratory. The results of this research will provide an optimization methodology for designing noise-control technologies in occupied indoor spaces. The successful completion of the proposed research can be expected to improve the cost-effectiveness of the duct silencer and sound masking system design. The outcomes also can be used as an evidence-based design approach for sustainable buildings. The culmination of these efforts will be the achievement of optimal acoustic quality in building spaces, thereby improving occupant comfort and performance in the built environment.