Dynamic water footprint assessment (WFA) of manufacturing industry by using cyber-physical systems (CPS)

This thesis reports research that helps manufacturing industry reduce its impact on local water resources and maintain a sustainable Water Footprint (WF) through the use of real-time sensing technologies and Cyber-Physical Systems (CPS). The key to improvement is to enable industry to be aware of its production activities’ impacts on the local water resources, which implies the necessity of an evaluation methodology. The methodology, Water Footprint Assessment (WFA), developed by the Water Footprint Network (WFN), was selected for quantifying the impact of industrial water use in this research. Several gaps in published WFA studies were identified: 1) studies rarely focus on operational WF and grey WF; 2) cumbersome data collection and lack of dynamic monitoring and evaluation cause a considerable lag between WFA and response; 3) there is a lack of a communication network supporting dynamic WFA. As a solution for these gaps, the development of automated dynamic WFA was proposed and implemented as a CPS. Few reports were found in the literature on applying CPS to monitor industrial wastewater quantity & quality. Consequently, a CPS architecture was designed in this research, using smart sensors to collect the required water and wastewater data.

To support the CPS for wastewater quality data collection, a technique for real-time inline measurement of Chemical Oxygen Demand (COD) was developed. COD is the typical critical pollutant in most manufacturing industries, especially food and beverage. The tool development was based on wastewater samples from a selected beverage plant and employed UV-VIS spectroscopy and Partial Least Squares (PLS) regression with variable selection tools, including Interval PLS (iPLS), Synergy Interval PLS (siPLS) and Moving Window PLS (mwPLS). Finally, the mwPLS model with the most optimised performances was adopted. A subsequent method-comparison study based on the Bland-Altman Plot (B-A Plot) was conducted to establish the capability of the developed tool when used for the dynamic calculation of WF. The results implied that this technique is suitable for both short-term (minutes to hours) and long-term (weeks to months) WF management. Two scenario studies in the WITNESS simulation software, based on one production line of the selected beverage plant, indicated significant benefits are possible for short-term and long-term WF management with the proposed CPS system.