Traditional design approaches, such as superstructure optimization, strongly depend on user-defined technologies, stream connections, and reasonable initial guesses for the unknown variables. This results in not only missing out on possible structures that can perform better than the selected one but also not considering essential aspects for practical implementation [1]. Regarding this, the enhanced P-graph framework, integrated with insights from machine learning and realistic technology models, is presented as a novel approach to WWTN synthesis. It offers a unique advantage providing the n-best structures considering its specific connectivity rules for input, intermediate, and terminal nodes [2]. In addition, the novel two-layer process synthesis algorithm is developed which incorporates combinatorial, linear, and nonlinear solvers to integrate the P-graph with realistic nonlinear model equations. This algorithm is composed of the outer layer that contains combinatorial P-graph algorithms, such as Maximal Structure Generation (MSG), and Solution Structure Generation (SSG); and the inner layer, which involves the optimization of individual structures with the different solvers [3] from which the solution structures are ranked based on chosen metrics, such as cost and reliability. However, the data availability on WWTNs' reliability is limited, and this information is often estimated based on expert field assessment. Considering this, applying Machine Learning (ML) methods for regression will allow for the accurate prediction of reliability based on different features, for which data are normally available in the WWTNs [4]. This will support better process design, enable proactive maintenance, and improve overall management. The pipeline network, pumping stations, and the WWTP are modeled with the P-graph framework to achieve this. Detailed and accurate models are developed for treatment technologies. ML methods such as eXtreme gradient boosting (XGBoost) and Neural Networks (NNs) are implemented to calculate the pipeline network's reliability using real data from a municipal WWTP. Finally, the n-best designs are presented with the possible failures in the WWTN and proposed solutions. This allows companies to anticipate and prevent process disruptions, which leads to maintaining safe and reliable operations.
References
[1] K. M. Yenkie, “Integrating the three E’s in wastewater treatment: efficient design, economic viability, and environmental sustainability,” Curr. Opin. Chem. Eng., vol. 26, pp. 131–138, Dec. 2019, doi: 10.1016/j.coche.2019.09.002.
[2] F. Friedler, Á. Orosz, and J. Pimentel Losada, P-graphs for Process Systems Engineering: Mathematical Models and Algorithms. Cham: Springer International Publishing, 2022. doi: 10.1007/978-3-030-92216-0.
[3] J. Pimentel et al., “Enabling technology models with nonlinearities in the synthesis of wastewater treatment networks based on the P-graph framework,” Comput. Chem. Eng., vol. 167, p. 108034, Nov. 2022, doi: 10.1016/j.compchemeng.2022.108034.
[4] Á. Orosz, J. Pimentel, A. Argoti, and F. Friedler, “General formulation of resilience for designing process networks,” Comput. Chem. Eng., vol. 165, p. 107932, Sep. 2022, doi: 10.1016/j.compchemeng.2022.107932.