Evaluation of the capacity of Apron feeders used in crushing–screening plants by response surface methodology and artificial intelligence methods

OpenAlex İngilizce

In this study, the capacity (Q) of Apron feeders is investigated through response surface methodology (RSM) and some artificial intelligence methods. In this regard, a comprehensive field survey is performed to compile quantitative data on the common working conditions of Apron feeders used in the Turkish Mining Industry (TMI). Based on the collected data, RSM analyses are performed to reveal the factors affecting the Q of Apron feeders. Accordingly, hopper width (B), the height of the material layer conveyed (D), conveyor speed (V), and fill factor (φ) are determined to be the most critical factors for the Q. Several interaction and contour plots are presented to observe the variations in the Q values. Moreover, several predictive models are also introduced to estimate the Q of apron feeders based on artificial intelligence methods such as multivariate adaptive regression spline (MARS), adaptive neuro-fuzzy inference system (ANFIS), and artificial neural networks (ANN). The performance of the established predictive models is assessed based on scatter plots, and it is found that the predictive model based on RSM methodology provides relatively better results than the ones found on soft computing-based predictive models. The presented predictive models can be reliably used to estimate the Q of Apron feeders with high capacity. However, crushing–screening plant designers should be careful when using established predictive models for assessing low-capacity Apron feeders. Based on the findings obtained, the present study demonstrates the applicability of RSM methodology and several artificial intelligence methods for evaluating the Q of Apron feeders.