Isoterma de adsorção de Langmuir com estruturas de erros autorregressivos regulares e irregulares

Abstract

Adsorption is a the solid-fluid type mass transfer operation in which there is chemical or physical accumulation of a substance or material chemical mechanisms at the interface between the solid surface call adsorbent and the solution called adsorbate, widely used for removing pollutants in industrial effluents. In many studies, in which the adsorption is modeled by the Langmuir isotherm, errors are correlated and data collection is not always done regularly (equidistant). The autocorrelation of errors and irregularities in the measurements of the independent variable can affect the estimation of the model parameters. One way to model dependent errors in a regression model is to use an autoregressive process which in turn implies that observations are carried out at equidistant intervals. However, the independent variable setting is often performed at irregular intervals causing a reduction of information obtained from the data set. A possible improvement of the quality of adjustment of these models, considering the irregularity is the use of irregular autoregressive process. This study aimed to compare the impact on the estimation of the Langmuir adsorption isotherm parameters with different structures autoregressive errors, regular and irregular, given the positive autocorrelation. It evaluated also the estimation for different sample sizes, different autocorrelation values of error and different positions of the observations not equidistant. It was found that there is the need to respect the model assumptions and therefore the ignoring autocorrelation produces bias in the parameter estimates and irregular autoregressive model is more precise and accurate in most of the scenarios analyzed.