Strategic optimization of water reuse in wafer fabs via multi-constraint linear programming technique

The risk of water shortage has been posing as a threat to water demanding industries in Taiwan, including the high-tech industries where ultrapure water is needed for the production of microchips. Such risks are especially unpredictable in the age of climate change, where more frequent extreme climate events such as prolonged droughts have sent these industries scrambling for securing water supply at a very high cost. The national policy also mandates strict water recycling standards for these high-tech plants, while the risk of water supply shortage also forces the industry to be water-conscious. However, most plants set their water recycling strategies based on experience or “rules of thumb” practices, without implementing optimization tools that can help making decisions in a more scientific approach. In this study we applied linear programming technique to optimize the water recovery path for a microchip assembly plant. A water balance diagram was formulated and completed to determine the existing water recycling performance, and the data was converted to a water flow network. The water flow network was then derived with a mathematical model to formulate a linear optimization problem. The proposed linear programming model is composed of mass balance constraints, unit specification constraints, capacity constraints as well as water quality constraints (discharge limits). The linear programming method was effectively appplied to improve the efficiency of water reuse. With the installation of the regeneration units, an increase of ∼40.1% in the volume of reused water was predicted. The results from water cost structure also indicated that, at higher water tariff, water reuses through reclaiming and generating spent effluents can alleviate the overall water consumption costs.