Uncertainty-based robust optimal design and control of cleanroom air-conditioning systems

Abstract

The total floor area and energy consumption of buildings with spaces requiring strict temperature and humidity control, such as pharmaceutical cleanrooms, hospitals, semiconductor/microchip factories (denoted as "cleanrooms" for brevity), have been growing rapidly worldwide. The energy intensity of cleanroom air-conditioning systems is usually 10-100 times greater than the average energy intensity of office buildings, due to the complexity of the systems and their operational needs for strict temperature and humidity control. However, the energy conservation issue in cleanrooms has not attracted sufficient attention. This PhD study attempts to comprehensively explore the ventilation strategy including the energy-efficient design and control of cleanroom air-conditioning systems by addressing the following questions which are not well answered in existing studies: • What is the most suitable ventilation strategy for cleanroom air-conditioning systems that can be adaptive to changes in different working conditions? • How to design cleanroom air-conditioning systems that can maximize their potentials in energy saving and cost reduction under uncertainties? • How to control cleanroom air-conditioning systems in operation that can enable systems to operate at high energy efficiency and reliability under measurement uncertainties? To ensure the high performance of air-conditioning systems as its design expectation, the ventilation strategy adopted should be adaptive to changes in internal sensible/latent load and ambient conditions. An "adaptive full-range decoupled ventilation (ADV) strategy" is proposed to minimize the system energy consumption by incorporating the advantages of existing ventilation strategies and adopting a novel "adaptive economizer". The main advantage of the ADV strategy is that it can select the optimal operation mode (i.e. with the minimum estimated energy use) from available/multiple operation modes. It avoids sub-cooling and reheating as far as beneficial via the best use of MAU and economizer for cooling and dehumidification. The energy and economic performance of the proposed ADV strategy are further evaluated under different climatic conditions. The results show that adopting the ADV strategy can offer significant and promising energy savings. The payback periods are attractive for both existing system retrofit and new system design in most climates. The design for cleanroom air-conditioning systems is a complicated task due to the coupling operation and counteraction/interaction among their components. An uncertainty-based robust optimal design method for cleanroom air-conditioning systems is developed for implementing the ADV strategy considering uncertainties. To address the issue of asynchronous loads in different zones/spaces with reduced computation demand, a probabilistic diversity factor method is developed to quantify the effects of uncertainties of space load diversities in multiple zones/spaces. The robust optimal design method is validated based on the actual air-conditioning systems in an existing pharmaceutical building. The test results show that the air-conditioning system, which is designed for the ADV strategy using the proposed design method, offers superior economic performance and satisfaction of service. Besides the robust optimal design, an online supervisory control strategy is also essentially needed for practical implementations of the ADV strategy. A risk-based online robust optimal control strategy is developed for multi-zone air-conditioning systems considering component performance degradation and measurement uncertainties. The core element of this strategy is the robust decision-making scheme, which is developed for selecting the optimal control mode of the air-conditioning systems online by compromising between the potential risks and benefits. This online control strategy is validated on a dynamic system simulation platform constructed on the basis of existing air-conditioning systems at full scale. The test results show the proposed control strategy can successfully determine the best operation mode allowing for component performance degradation and measurement uncertainties, and ensure air-conditioning systems to operate at high reliability and energy efficiency. To conclude, the developed adaptive ventilation strategy, the system robust optimal design and robust online optimal control methods can provide significant improvements to current design and control practice of cleanroom air-conditioning systems.