(183a) Synthesis of Coffee Wastewater Treatment Network with Single Input and Multiple Output Streams Using the P-Graph Framework | AIChE

(183a) Synthesis of Coffee Wastewater Treatment Network with Single Input and Multiple Output Streams Using the P-Graph Framework


Aboagye, E. - Presenter, Rowan University
Pimentel, J., Grupo de Procesos Quimicos y Bioquimicos, Universidad Nacional de Colombia, Bogota, Colombia
Orosz, A., University of Pannonia
Cabezas, H., University of Miskolc
Friedler, F., Pazmany Peter Catholic University
Yenkie, K., Rowan University
Kearns, Q., Rowan University

Globally, the food and beverage industries are considered the third largest water consumer (Valta et al., 2016). Soluble coffee production, one of the main products from the beverage industries, generates a significant amount of wastewater during the extraction and freeze- or spray drying processes (Clarke and Macrae, 1987). Rather than treating the entire waste stream for disposal, recovering and reusing part of the wastewater for cooling processes presents a better alternative at reducing the water footprint of the beverage industry. Process Network Synthesis (PNS) refers to the task of algorithmically constructing a process from a set of building blocks; it has been extensively employed for the synthesis of water and wastewater treatment networks, and optimizing the economic, environmental, or social dimension of the projects (Galan and Grossmann, 1998; Grossmann et al., 2014). Traditional methods for wastewater treatment network synthesis have limitations due to simplified technology models and deficiencies in structural enumeration. On the other hand, the more complex models for the treatment technologies lead to a larger and more complicated mixed integer non-linear optimization model, which cannot be properly solved with the contemporary MINPL solvers. Moreover, the formulation of the structural decisions of the problem in terms of binary variables, may derive in structural-related optimal solutions. Rigorous and exhaustive enumeration of all the possible treatment networks coupled with improved models can help overcome these limitations.


In this work, we implement the P-graph framework for designing wastewater treatment networks (Aboagye et al., 2021; Yenkie et al., 2021). The P-graph framework is a graph-theoretic approach for the design of processes, and has been used extensively to evaluate problems of combinatorial nature, such as supply chains, planning of evacuation routes, and industrial PNS problems (Friedler et al., 1998). The axioms of the P-graph framework lead to the identification of all plausible structures for the treatment networks (Friedler et al., 1992). Furthermore, the axioms provide information relating to the structural feasibility of all synthesized networks (Friedler et al., 2022). The combinatorial algorithms of the P-graph framework, which are based on the axioms, help to exclude any inconsistencies in the structural model, thus avoiding incomplete structures that misdirect the optimization. The exclusion of incomplete structures presents a great advantage in our superstructure-based optimization, as unnecessary complexities are avoided leading to simpler mathematical models to be optimized. Due to the structural enumeration, the method is also able to avoid the structure-related local optimums of the problems, which are a major issue for conventional MINLP-based techniques. Figure 1 is the synthesized superstructure for this work.

Thus, in this work we first identify all the potentially feasible solution structures using the combinatorial axioms of P-graph framework, and then optimize them individually. The first step is generating a superstructure known as the maximal structure using the Maximal Structure Generation (MSG) algorithm in P-Graph Studio software. This superstructure contains all the treatment technologies, flows, connections, bypasses, and mixers that are considered during the synthesis procedure. We consider a three-stage treatment process comprising sedimentation and membranes technologies at the primary stage, activated sludge, rotating biological contactors, and membrane bioreactors at the secondary stage, while the tertiary stage consists of advanced oxidation processes and ion exchange operations, as shown in Figure 1. The next step is to use the Solution-Structure Generation (SSG) algorithm to identify and enumerate each solution from the maximal structure, thus producing a set of feasible networks from the structural point of view. These combinatorially feasible networks may still be infeasible due to process models and constraints such as mass and energy balance, capacity constraints, thermodynamics and kinetics considerations, and sustainability restraints. Due to this possible infeasibility, we evaluate each combinatorially feasible structure using nonlinear programming (NLP). Since every feasible structure must be also combinatorially feasible, this method evaluates all plausible structures including the globally optimal one. To achieve this, we model each of the feasible structures and their associated operating technologies using MATLAB. Then, we use the NLP solvers found in AMPL to solve the model generated by MATLAB through AMPL-API interface. The main objectives for this problem are to minimize the treatment cost as well as to reduce two sustainability metrics, namely the Sustainable Process Index (SPI) and Emergy (Narodoslawsky and Krotscheck, 1995, 2004; Odum, 1988; Ulgiati et al., 1994). The SPI metric helps to quantify the ecological burden of the treatment process in land area while Emergy accounts for the available energy used in solar energy equivalents. Therefore, to set an objective function, we constrain the SPI and Emergy metrics to the problem and optimize the cost function.

Case Study

We apply this framework to a soluble coffee processing plant of Nestlé USA, located in Freehold, New Jersey, that generates 1,324,894 L/d of wastewater (Wisniewski et al., 2020). The wastewater stream comprises Chemical Oxygen Demand (COD) and Total Suspended Solids (TSS) contaminants with high conductivity. The desire is to have two effluent streams (Outlet#1 and Outlet#2) with different purity levels where one stream is recycled to the process to reduce freshwater consumption while the remaining is discharged into a water body. The purity specification of the effluent stream to be released into the water body is constrained by the acceptable limits set by the USEPA for wastewater discharge. Table 1 shows the inlet and outlet contaminant specifications.

Summary and Conclusion

We have already developed an Excel-based tool for evaluating wastewater treatment networks [http://p-graph.org/wastewater-treatment/]. Thus, this is an extension of the previous work where we integrate nonlinear models into P-graph. Furthermore, this work incorporates a simultaneous assessment of economics and sustainability into the synthesis process. Thus, presenting decision-makers with multiple options for consideration for further detailed design.


Aboagye, E.A., Pimentel, J., Orosz, Á., Cabezas, H., Friedler, F., Yenkie, K.M., 2021. Efficient Design and Sustainability Assessment of Wastewater Treatment Networks using the P-graph Approach: A Tannery Waste Case Study. Chemical Engineering Transactions 88, 6.

Clarke, R.J., Macrae, R. (Eds.), 1987. Coffee. Springer Netherlands, Dordrecht. https://doi.org/10.1007/978-94-009-3417-7

Friedler, F., Fan, L.T., Imreh, B., 1998. Process network synthesis: Problem definition. Networks 31, 119–124.

Friedler, F., Orosz, Á., Pimentel Losada, J., 2022. P-graphs for Process Systems Engineering: Mathematical Models and Algorithms. Springer International Publishing, Cham. https://doi.org/10.1007/978-3-030-92216-0

Friedler, F., Tarján, K., Huang, Y.W., Fan, L.T., 1992. Graph-theoretic approach to process synthesis: axioms and theorems. Chemical Engineering Science 47, 1973–1988. https://doi.org/10.1016/0009-2509(92)80315-4

Galan, B., Grossmann, I.E., 1998. Optimal Design of Distributed Wastewater Treatment Networks. Ind. Eng. Chem. Res. 37, 4036–4048. https://doi.org/10.1021/ie980133h

Grossmann, I.E., Martín, M., Yang, L., 2014. Review of optimization models for integrated process water networks and their application to biofuel processes. Current Opinion in Chemical Engineering, Energy and environmental engineering / Reaction engineering 5, 101–109. https://doi.org/10.1016/j.coche.2014.07.003

Narodoslawsky, M., Krotscheck, C., 1995. The sustainable process index (SPI): evaluating processes according to environmental compatibility. Journal of Hazardous Materials, Selected papers presented at the Conference on Hazardous Waste Remediation 41, 383–397. https://doi.org/10.1016/0304-3894(94)00114-V

Narodoslawsky, M., Krotscheck, Ch., 2004. What can we learn from ecological valuation of processes with the sustainable process index (SPI) — the case study of energy production systems. Journal of Cleaner Production, Advances in cleaner production technologies 12, 111–115. https://doi.org/10.1016/S0959-6526(02)00184-1

Odum, H.T., 1988. Self-Organization, Transformity, and Information. Science 242, 1132–1139. https://doi.org/10.1126/science.242.4882.1132

Ulgiati, S., Odum, H.T., Bastianoni, S., 1994. Emergy use, environmental loading and sustainability an emergy analysis of Italy. Ecological Modelling 73, 215–268. https://doi.org/10.1016/0304-3800(94)90064-7

Valta, K., Moustakas, K., Sotiropoulos, A., Malamis, D., Haralambous, K.J., 2016. Adaptation measures for the food and beverage industry to the impact of climate change on water availability. Desalination and Water Treatment 57, 2336–2343. https://doi.org/10.1080/19443994.2015.1049407

Wisniewski, C.M., Laurio, M.V.O., Slater, C.S., Savelski, M.J., 2020. Environmental and process assessment of a vibratory nanofiltration system for the recovery of water from soluble coffee manufacturing waste. Environ Prog Sustainable Energy 39. https://doi.org/10.1002/ep.13425

Yenkie, K.M., Pimentel, J., Orosz, Á., Cabezas, H., Friedler, F., 2021. The P-graph approach for systematic synthesis of wastewater treatment networks. AIChE Journal 67, e17253. https://doi.org/10.1002/aic.17253