An approach for interpreting thermogravimetric profiles using artificial intelligence

Expert systems are proposed in order to analyze thermogravimetric profiles. The methodology showed to be an efficient tool as a complementary procedure for experimentalists in the area of themogravimetric analysis. Several simple and complex molecules were used to validate the developed methodology. The advantage of the actual approach can be attributed to its efficient error minimization for each weigh loss and global process. In addition, the proposed expert system presents a low computational demanding. The validation analysis using compounds such as salicylato (amine) Co (III) complexes, norfloxacin complexes of manganese (II) and cobalt (II), and calcium borate produced an average confidence interval of about 2%.