A comparative study on energy performance assessment for HVAC systems in high-tech fabs

Abstract

High-technology fabrication plants (fabs) such as semiconductors generally consume enormous energy, where HVAC system accounts for the second largest part at roughly 40–50%. Therefore, the energy performance for HVAC system in semiconductor industry is undoubtedly essential. In this study, several EPMs (Energy Performance Metrics), namely SCL, ${\mathrm{\omega }}_{scl}$, EI_AFR, was proposed to estimate and assess the energy using for HVAC system in a fab through the analysis of theoretical model and influential factors. The energy performance comparison cases were conducted in both sub-system level and system level through comparative analysis on the operation data, simultaneously the presented EPMs were verified and applied in two air systems and two fabs studied. Most importantly, the results show the advantages of SCL in estimating overall cooling load in high-tech fabs, which can be used in the future for designing HVAC system in fabs that with higher energy efficiency and ${\mathrm{\omega }}_{scl}$ that can be applied to evaluating the stability and validation in maintaining indoor environmental metrics including indoor temperature, humidity, particles and pressure difference. Moreover, it is generally proved that the annual EPMs like EI_AFR (kWh/Air Flow Rate) turn to provide more consistent comparison than the traditional metrics like EI kWh/Floor Area) and SEUP (kWh/Unit of Production) in assessing overall energy performance of HVAC system in high-tech fabs.

Introduction

High-technology fabrication plants (fabs) such as semiconductors generally consume enormous energy and the energy performance for semiconductor industry is undoubtedly essential. Previous studies have shown that Heating, Ventilation, and Air conditioning (HVAC) system accounts for the second largest energy consumption in high-fabs, at about 40%–50% and energy density of HVAC system for cleanroom in high-tech fabs is generally 10 times that for thermal comfort [1]. Therefore, many studies regarding energy conservation for HVAC system in cleanroom of semiconductor fabs have been conducted. For example, Brown [2] discussed an energy-saving opportunity for the make-up air unit (MAU) of a semiconductor fab. Hu and Tsao [3] compared energy efficiency performance of five different HVAC systems for cleanrooms and pointed out that the MAU + FFU systems exhibited the highest energy efficiency. Tsao and Hu [4,5] et al. investigated the difference in energy efficiency performance of MAU with different pre-cooling and preheating/humidification schemes. K. Kircher et al. [6] made an assessment of three energy-saving opportunities by modeling and simulation and calculate both the energy reduction and payback time to recommend the best strategy for energy efficiency.

Meanwhile, there is a series of energy benchmark works having been done in the past 20 years. Xu [7,8] reported energy consumption and particle control of facility systems and characterized fab energy use in terms of energy use or power demand. In 1997, ISMI sponsored and participated in the international benchmarking study of fourteen 150 mm and 200 mm fabs around the world collecting and sharing energy consumption data for fab process areas and facility operations equipment [10]. Hu et al., in 1999 [9] studied energy benchmark for nine fabs producing 150 mm and 200 mm wafers in Taiwan, China. In 2008, Hu [11] and his team established the energy benchmark of a typical 8-in. DRAM semiconductor fab through field measurement data for chilled water system, PCW system, nitrogen system, vacuum system and UPW system, respectively, which can assess the efficiency of different energy-saving schemes and as a good reference for factory authorities. Then, they [12] characterized the electric energy consumption and production of 300 mm DRAM fabs by using various performance metrics, including PEI, EUI and UOP in 2010. Chang and Hu et al. [13] identified the specific energy consumption of all major energy consuming segments of the Dynamic Random Access Memory module supply chain, including 12” Si-wafer (ingot), wafer fabrication, assembly, testing, and printed circuit board in 2012. By 2017, Hu [14] lead the team developed a new calculator to provide energy conversion factors for each sub-system in the fab, which can be used as a straightforward tool for energy-saving and future design. Published studies for fabs refer to energy consumption normalized by either per unit wafer area or per unit fab area [[7], [8], [9], [10], [11], [12], [13], [14]]. Energy consumption levels in 200 mm and 300 mm semiconductor fabs in Asia, North America, and Europe were studied to gather baseline data on different facility systems to develop energy efficiency metrics, as the effectiveness of the present EPMs described as energy consumption either per unit wafer area or per unit fab area varies a lot in fabs with different energy consumption levels [15].

However, the published data on energy demand for semiconductor fabs is quite limited, largely due to the facts that energy consumption data are considered as confidential information for manufacturers. As a result, few studies focus on the energy benchmark and energy performance for the facility system, especially for HVAC system. The objective of this study, based on measured data or appropriate assumptions when operating data is not available, is to develop better EPMs (Energy Performance Metrics) to estimate and assess the energy uses for HAVC systems. This paper briefly introduces the specification of the fabs and air systems studied in the first place. Through several theoretical models and influential factor analysis of cooling load in high-fabs, the new EPMs is presented. The new EPMs is verified and applied in the following comparative cases, on sub-system level (air systems) and system level (two fabs) separately. Finally, conclusions are drawn.