Development and evaluation of an evolutionary tool for obtaining the crystallization rate constant of semicrystalline polymers (HDPE case) as a function of temperature

This paper presents a methodology based upon genetic algorithms, which is intended to find from a set of data, the parameters of an approximating function consisting of radial basis functions. The aim is to obtain a smooth, continuous and integrable approximating function, capable of satisfactorily expressing the relationship between temperature and the kinetic crystallization constant K(T) used in the different models of non-isothermal crystallization. The proposed methodology was tested over HDPE crystallization data at different cooling velocities (5 and 20 °C/min). Additionally, a different evolutionary methodology, previously developed [Master degree thesis, Postgrado de Investigación de Operaciones, Facultad de Ingenierı́a, Universidad Central de Venezuela, 1998], was also tested over the same data. The results show that both of the methodologies tested are capable of finding good approximations to the kinetic crystallization constant K(T).