Determination Of Heat Transfer Coefficients In Heat Exchangers By Genetic Algorithm

Genetic algorithm (GA), developed by John Holland in the 1970s, is gaining widespread attention throughout the years. In particular, GA is being extensively used in the field of optimization. This paper discusses the utilization of GA in estimating the coefficients and exponents in developing the Nusselt Number correlation in two different heat exchanger geometries, with the aim to reduce the error incurred in predicting heat transfer coefficients and heat transfer rates. With such reduction, for instance, the design of heat exchangers can be more compact, incurring less cost in the process.

In this study, the scope is limited to Nusselt Number correlations in the plain tube concentric heat exchanger and brazed plate heat exchanger. However, the method has general applicability for other types of heat exchangers. To achieve the objective, proper modeling of the problem to suit the form required by GA is important. Then, the knowledge of coding is required so that the GA can be implemented.

Based upon data from the industry, comparisons are drawn with the correlation developed by conventional methods. The GA approach is shown to lower the errors and give greater confidence on the predictive correlations for these cases.