Optimization of the design and operation of sustainable and resilient buildings of 21st century

申请者具有高度原创性和创新性的研究计划的长期目标是优化21世纪可持续和有弹性的建筑的设计和运营。恢复力是在面对逆境、创伤、悲剧、威胁等情况下很好地适应并迅速从这些情况中恢复的过程。短期目标1侧重于开发一种供暖、通风和空调(HVAC)系统同时多重故障的自动检测和诊断方法(ADDSMF)。在申请人的设想中，安装在商业和机构建筑中的建筑自动化系统(BAS)应该模拟人类的神经系统，并具有一些增强的功能，如数据验证、知识提取和故障检测和诊断(FDD)。这种方法的开发基于BAS的测量质量和有效的自适应数学模型，这些模型不断地从新的测量中学习以预测性能基准，检测暖通空调系统中的同步故障，识别和自我校准故障传感器。该研究计划专注于应用机器学习(ML)技术，这是人工智能(AI)领域的一个子领域，用于开发此类自适应模型。提出了两类不同的模型：黑箱模型和灰箱模型。将开发ADDSMSF模型的原型库。在商业建筑物中使用这种模型将有助于快速检测故障，并降低运行成本和病毒在建筑物中的传播。这些模式应该整合到加拿大公司生产的BAS中，这将在国内和国际上变得更具竞争力。短期目标2侧重于开发暖通空调系统的多目标优化框架，并将其应用于加拿大国家建筑能源法(NECB)参考暖通空调系统的优化选择。申请人提议开发一个大规模多目标优化框架，用于选择暖通空调系统的配置和确定设备容量。该框架将与建筑能源分析程序(如EnergyPlus，Equest)相连接。这是一个寻找最小生命周期能耗、生命周期成本、生命周期(火用)破坏和生命周期环境影响这四个相互冲突的目标函数的问题。不能保护居住者免受预期的室外威胁(物理、化学和生物)的暖通空调系统的设计替代方案被淘汰。建议的优化框架将用于为加拿大主要城市的不同建筑类型选择低能源目标值的创新参考暖通空调系统。这项研究的结果有助于NECB的更新。

The long-term objective of the applicant's highly original and innovative research program is the optimization of design and operation of sustainable and resilient buildings of 21st century. Resilience is the process of adapting well in the face of adversity, trauma, tragedy, threats etc. and recovering quickly from such situations. Short-term objective 1 focuses on the development of a method for Automated Detection and Diagnosis of Simultaneous Multiple Faults (ADDSMF) of heating, ventilation and air conditioning (HVAC) systems. In the applicant's vision, Building Automation Systems (BAS) installed in commercial and institutional buildings should mimic the human nervous system, with some enhanced features such as data validation, knowledge extraction, and fault detection and diagnosis (FDD). The development of such a method is based on quality of measurements from BAS, and efficient adaptive mathematical models that learn continuously from new measurements to predict the performance benchmarks, detect simultaneous faults in HVAC systems, identify and self-calibrate the faulty sensors. The research program focuses on the application of Machine Learning (ML) techniques, a subfield of artificial intelligence (AI) domain, for the development of such adaptive models. Two different classes of models are proposed: Black-box models and Gray-box models. A prototype library of ADDSMSF models will be developed. The use of such models in commercial buildings will contribute to fast detection of faults, and reduction of operation costs and propagation of viruses in buildings. The models should be integrated in BAS produced by Canadian companies, which will become more competitive, nationally and internationally. Short-term objective 2 focuses on the development of multi-objective optimization framework of HVAC systems, with application to optimum selection of reference HVAC systems of National Energy Code of Canada for Buildings (NECB). The applicant proposes the development of a large-scale multi-objective optimization framework for the selection of configuration of HVAC systems and sizing the equipment capacity. The framework will be connected with a building energy analysis program (e.g., EnergyPlus, eQUEST). This is a problem of searching for minimum life cycle energy use, life cycle cost, life cycle exergy destruction, and life cycle environmental impact, four conflicting objective functions. Design alternatives of HVAC systems that cannot protect the occupants against anticipated outdoor threats (physical, chemical and biological) are eliminated. The proposed optimization framework will be used for the selection of innovative reference HVAC systems with low energy target value, for different building types in major cities in Canada. Outcomes of this research contribute to the update of NECB.