Inferential sensors and inferential control schemes for use in building automation

本研究旨在探讨如何通过使用推理传感器和推理控制方案来改善对建筑环境的控制。通过这项研究，将产生有关建筑自动化，环境监测和建筑环境评估的新知识。短期目标是通过开发和使用推理传感器和推理控制技术来加强目前的建筑物控制算法和方案，并开发一种广泛接受的方法和实验设施，用于测试建筑物自动化设备和系统的性能。长期目标是研究推理传感和推理控制技术的数学和物理基础，并了解它们对经典控制理论的影响。 在建筑环境中，由于经济和/或技术原因，对于建筑物系统的最佳操作至关重要的许多变量即使不是不可能也难以直接测量。推理传感器基于推理建模，该推理建模将一个关键变量与可以容易地测量的多个相关变量相关联。在推理建模中，将采用多种方法，包括物理模型、随机和统计方法以及神经模糊系统。推理传感技术将与传统的控制方案相结合，设计推理控制方案。两种方法将被应用：基于模型的预测控制和级联反馈推理控制。推理控制方案将极大地简化楼宇自动化系统的设计和实现，研究目标将通过各种研究活动来实现：文献研究和调查，推理传感器和控制算法的理论研究，仿真研究和实验室实验研究。这些技术和新知识将有助于提高建筑物的能源效率和室内环境质量。

This research aims to investigate how to improve the control of the built environment through the use of inferential sensors and inferential control schemes. Through this research, new knowledge on building automation, environmental monitoring, and assessment of the built environment will be generated. The short-term objectives are to enhance current building control algorithms and schemes through the development and use of inferential sensor and inferential control technology and to develop a widely accepted method and experimental facility for testing the performance of building automation devices and systems. The long-term objectives are to study mathematic and physical fundamentals of inferential sensing and inferential control technology and to understand their impacts on the classic control theory.  In the built environment, many variables that are critical for the optimal operation of the building systems are difficult, if not impossible, to be directly measured due to economical and/or technical reasons. An inferential sensor is based on inferential modeling that associate one critical variable with a number of relevant variables that can be easily measured. A variety of approaches will be applied in the inferential modeling, including physical model, stochastic and statistical method, and neuro-fuzzy systems. The inferential sensing technologies will be integrated with conventional control schemes to design inferential control schemes. Two approaches will be applied: model-based predictive control and cascaded feedback inferential control. The inferential control schemes will significantly simplify the design and implementation of building automation systems.The research objectives will be achieved through a variety of research activities: literature study and survey, theoretical study on inferential sensors and control algorithms, simulation study, and experimental study in laboratory. The technologies and new knowledge will help improve energy efficiency and indoor environment quality of buildings.