Shell-and-tube heat exchangers are widely used across various industries due to their high efficiency in energy recovery, drying, and cooling processes. However, optimizing these exchangers presents significant challenges due to the discrete nature of the design variables and the complexity of the governing equations, which are discontinuous and non-differentiable. To address these challenges, metaheuristic algorithms such as Genetic Algorithm, Particle Swarm Optimization, and Differential Evolution are commonly applied. This study introduces two enhanced optimization algorithms: Opposition-Based Differential Evolution and Comprehensive Opposition-Based Learning. These methods incorporate the concept of opposition both in generating initial solutions and in iterative optimization processes, enabling the efficient identification of optimal solutions.

The optimization process focuses on key design variables, including tube diameter, shell diameter, and baffle spacing, which directly determine the complete geometry of the heat exchanger. The performance of these algorithms was evaluated across two case studies from prior research. Results showed an 18.88% improvement in cost reduction for the first case study and a 16.37% improvement for the second compared to conventional methods. Additionally, these algorithms outperformed Aspen EDR, a commercial heat exchanger design software, in identifying cost-effective and geometrically feasible solutions. Since Aspen EDR has limited modules for heat exchanger design and fails to provide solutions for complex geometries, such as helical or coiled tube configurations, the proposed algorithms offer a reliable and efficient alternative in such scenarios.
This research demonstrates the potential of these enhanced optimization algorithms to overcome traditional design limitations, providing a more versatile and effective approach