Pareto domain: an invaluable source of process information

Area of Science:

• Chemical Engineering
• Process Optimization
• Computational Chemistry

Background:

• Chemical processes face competitive markets and strict environmental regulations, necessitating optimal operation.
• Balancing multiple, often conflicting, objectives is crucial for process efficiency and decision-making.
• Multi-objective optimization (MOO) offers a framework to navigate these trade-offs.

Purpose of the Study:

• To review methods for solving MOO problems and selecting optimal solutions.
• To demonstrate the value of MOO in chemical process design and operation.
• To illustrate how MOO yields richer process insights than single-objective approaches.

Main Methods:

• Defining the Pareto domain of non-dominated solutions.
• Ranking Pareto-optimal solutions using expert preferences, visualization, or algorithms.
• Solving four case studies: PI controller design, SO2 to SO3 reactor, distillation column, and acrolein reactor.

Main Results:

• MOO effectively identifies a range of optimal trade-offs between conflicting objectives.
• The Pareto domain provides comprehensive information about process behavior.
• Case studies highlight the benefits of MOO for understanding variable-objective relationships.

Conclusions:

• MOO is a powerful tool for chemical process optimization, offering superior insights over traditional methods.
• Generating and utilizing the Pareto domain enhances understanding of complex process dynamics.
• MOO facilitates informed decision-making by revealing the spectrum of achievable performance outcomes.