Implementation of P-graph modules in undergraduate chemical engineering degree programs: experiences in Malaysia and the Philippines

Abstract

The process graph (P-graph) framework was introduced in the early 1990s as a graph theoretic approach to Process Network Synthesis (PNS) problems in chemical engineering. This framework has several advantageous features, such as mathematically rigorous generation of maximal structures (i.e., superstructures), elucidation of combinatorially feasible solution structures, and efficient search of solution space for optimal PNS compared to conventional branch-and-bound algorithms for mixed integer linear programming (MILP) models. In the three decades since its inception, the P-graph framework has proven to be a viable approach to a wide array of PNS and other structurally analogous problems; the methodology has also matured, as evidenced by the development of key software tools (PNS Draw, PNS Studio and P-graph Studio) and its appearance in modern textbooks. Nevertheless, the P-graph framework has yet to achieve broad, mainstream penetration among grassroots chemical engineering users, especially those without a strong mathematical programming background. In this paper, we discuss the implementation of instructional P-graph modules in undergraduate chemical engineering degree programs at the University of Nottingham, Malaysia Campus and De La Salle University, Manila, Philippines. These modules have primarily focused on the design of Green Processes. Results show that the visual nature of the P-graph methodology is an advantage for practical engineering decision-making, and that it complements the learning of mainstream techniques such as mathematical programming.

Introduction

Contemporary engineering decision-making must now account for various sustainability considerations (Steffen et al., 2015) and anthropogenic climate change in particular (IPCC, 2014). Various methodologies in process systems engineering (PSE) and process integration (PI) have been shown to be useful for enhancing the sustainability of industrial systems, using such measures as energy efficiency, fuel substitution, cleaner production, CO₂ capture and waste management (Yong et al., 2016). However, in practice, decision makers need to balance economic and environmental goals and thus take trade-offs into account via sensitivity analysis; in particular, it is useful to be able to determine how small penalties in economic goals can result in drastic reductions in environmental impact (Friedler, 2015). It has thus been recognized that the capability to determine optimal and near-optimal solutions is essential for planning and designing sustainable man-made systems (Voll et al., 2015).

P-graph methodology is a graph theoretic framework for solving process network synthesis (PNS) problems, whose combinatorial nature pose distinct modelling and computational challenges. Its development dates back to the early 1990s, when a series of papers described its basic axioms (Friedler et al., 1992a), the consequent combinatorial algorithms (Friedler et al., 1992b) as well as a rigorous framework for generating a maximal structure (i.e., superstructure) (Friedler et al., 1993). The P-graph framework enables rigorous model-building and efficient generation of optimal (as well as near-optimal) solutions, primarily by utilizing unique information embedded in all PNS problems. P-graph methodology has been applied to a wide array of PNS as well as structurally analogous problems. It can also be used in conjunction with mathematical programming techniques (Bertok et al., 2013a). Its capability to generate multiple structures for any given PNS problem makes it potentially useful for designing systems for reliability (Bertok et al., 2013a, Bertok et al., 2013b) and flexibility (König et al., 2014); likewise, temporal aspects have also been addressed via time-constrained models (Frits and Bertok, 2014), multi-period systems (Heckl et al., 2015a), and multi-period systems with lower limits for part load operation (Tan and Aviso, 2016). Many of the key P-graph developments from the early 1990s until 2013 are described in a key review article by Lam (2013). Some of the additional works on P-graph applications from 2013 to the present are listed in Table 1, including papers in journal issues and conference sessions in honour of the late Prof. L. T. Fan, one of the early proponents of the P-graph framework. Klemeš and Varbanov (2015) recently noted key trends in recent P-graph literature – i.e., diversification of application domains, as well as the emergence of new research groups outside of the EU. Furthermore, P-graph methodology has become sufficiently established for inclusion in modern textbooks (e.g., Peters et al., 2003) and reference books (Klemeš et al., 2011); a recent P-graph article has also appeared in a professional publication for practicing engineers (Cabezas et al., 2015).

This work was motivated by discussions at an expert workshop on P-graphs that took place in Veszprem, Hungary in January 2015. Among the key outcomes of the meeting was the need to understand how to popularize the P-graph framework to achieve broader mainstream penetration of this powerful tool for various PNS problems. Because of our prior experience in the integration of P-graph topics in existing undergraduate curriculum at the University of Nottingham Malaysia Campus and De La Salle University, Manila, Philippines, it was subsequently agreed to document how the topic has been implemented, with moderate success, in these institutions. This academic collaboration could enhance the effectiveness of teaching and learning process via newly developed methods, which is one of the main challenges for Engineering education (Klemeš et al., 2013). A survey was also conducted to assess student preferences with respect to the use of the P-graph framework vis-à-vis the alternative (i.e., commercial equation-based optimization software such as GAMS and LINGO). This paper provides an account of these activities, and is organized as follows. The next section gives a brief description of P-graph methodology, as well as an overview of available scientific literature, instructional materials and software tools. Then, we give accounts of implementation of P-graph topics in undergraduate curriculum in these two institutions, within the context of distinct features of chemical engineering education in Malaysia and the Philippines. Key insights drawn from recurring preferences seen in both sets of surveys are also discussed. Finally, conclusions and prospects for future work are given at the end of the paper.