Multi-objective optimal design of semiconductor cleanroom air-conditioning systems considering load uncertainty and equipment degradation

Semiconductor clean fabrications are one of the most dynamically updated building types which require extremely stringent indoor environment controls and exhibit energy intensities 50–100 times greater than office/commercial buildings. The design challenges and complexity of their air-conditioning systems are thus much higher, and improper sizing would lead to more severe space environment deviations and significant cost waste. However, conventional design methods for semiconductor cleanrooms, deriving from those of office/commercial buildings, neglect the unique features of their loads and operation principles. This study therefore proposes a novel optimal design method for semiconductor cleanrooms considering load uncertainty and performance degradation of air-conditioning equipment. The proposed design method optimizes the cleanroom environmental, energy and economic performance simultaneously, through a multi-objective algorithm. To validate the proposed method, three actual semiconductor fabrication buildings are adopted as test cases. Results demonstrate that the proposed method can obtain robust and optimal air-conditioning systems and a remarkable reduction of 92 % in unmet hours is achieved with a mere increase of 5–7 % in life cycle cost, compared with a commonly used design method. The results and experience also show the proposed method can be a promising tool for engineers to make informed decisions aligned with their priorities.