Application of a multicriterial optimization to the resolution of X-ray diffraction curves of semicrystalline polymers

摘要

The analysis of wide angle X-ray diffraction (WAXD) curves of semicrystalline polymers is connected with their decomposition into crystalline peaks and amorphous components. To this aim a theoretical curve is constructed which is a best fitted, mathematical model of the experimental one. All parameters of the theoretical curve are found using an optimization procedure. As it has been already proved, a reliable decomposition can be performed only by means of a procedure which effectively performs a multicriterial optimization. It consists in minimization of the sum of squared deviations between the theoretical and experimental curves and simultaneous maximization of the area of the amorphous component. So, the objective function in the optimization procedure is constructed of two criterial functions which represent the two requirements. The proportions between the criterial functions and their significance at different stages of the procedure must be determined by suitable weights. A proper choice of the weights is an important part of the procedure. In this paper a new solution of this problem is presented: the weights are changed dynamically in subsequent steps of the optimization procedure. A few different algorithms of the weights determination are presented and evaluated by means of several statistical method. The optimization procedures equipped with these algorithms are tested using WAXD patterns of popular polymers: Cellulose I, Cellulose II and PET. It is shown that the optimization procedures equipped with the dynamic algorithms of weights determination are much more effective than the procedures using some constant, arbitrarily chosen weights.