Optimization for Sustainable Process Design

One important way to support the global goal of net-zero emissions is the development of processes that reduce environmental impacts, including CO₂ emissions. This is the domain of process systems engineering (PSE). However, achieving this goal requires the development of PSE tools that focus not only on the optimization of technical and economic performance but the incorporation of environmental performance in process synthesis. Such optimization and decision support tools must be sufficiently robust that can support the exploration and analysis of various process alternatives under uncertainty and produce an optimal balance between profit maximization, energy efficiency, and environmental performance. This paper focuses on developing PSE tools that allow for the synthesis and optimization of processes that meet environmental design goals along with technical and economic performance.

We formulated a multiobjective optimization in MATLAB^® in which the objective functions are minimizing total cost and carbon footprint subject to certain constraints while the process alternatives are modelled as decision (binary) variables. Aspen Plus^® Software is utilized to model these alternative processes with accurate mass and energy balances. The results from the simulation were regressed to formulate mathematical models for the optimization.

A genetic algorithm solver is utilized to generate a Pareto front which shows the trade-off between carbon footprint and total cost, at the same time, giving the optimum reactor configuration at each local minimum. The multiobjective optimization using the stochastic base approach is compared with the gradient-based approach.

These tools will be presented in the specific context of syngas production from municipal solid waste (MSW) via steam gasification. Four different reactor configurations namely, plasma gasifier, entrained flow reactor, fluidized bed reactor, and fixed bed reactor to produced syngas were modelled and optimal configurations at best operating conditions were obtained with the constraint that the molar ratio of hydrogen to carbon monoxide (H2:CO) is suitable for the Fischer Tropsch process.

The results provide a set of optimal process alternatives and insight for decision-makers and companies to decide depending on their priorities. Similarly, the outcome of this work will demonstrate the application of superstructure optimization in process design and its effectiveness in decision-making.