Determination of cell voltage and current efficiency in a chlor-alkali membrane cell based on machine learning approach

Abstract

Due to importance of cell voltage and caustic current efficiency (CCE) in chlor-alkali industry, the necessity of accurate approach for prediction these parameters has become evident. In the current work, an extreme learning machine (ELM) approach is used to this end. Determination of the statistical qualities including R2 and different types of error reveals the fact that ELM method is suitable tool for calculation of CCE and cell voltage. The determined R2 values for CCE and cell voltage are equal to 1. Furthermore, RMSE values are 0.00002 and 1.3 × 10−6 for cell voltage and CCE, respectively. On the other hand, different graphical methods confirmed this acclaim. Moreover a sensitivity analysis is used to show effect of brine concentration, current density, operating temperature, electrolyte velocity, run time and pH on cell voltage and CCE. This analysis concluded to the fact that brine concentration and current density have the most effects on CCE and Cell voltage, respectively.