An intelligent sensing system for estimation of efficiency of carbon-capturing unit in a cement plant. (1st December 2022)

• Record Type: Journal Article
• Authors: Jadoon, Usman Khan; Ahmad, Iftikhar; Noor, Tayyaba; Kano, Manabu; Caliskan, Hakan; Ahsan, Muhammad
• Is Part Of: Journal of cleaner production. Volume 377(2022)
• Publication Date: 2022-12-01
• Subjects: Artificial intelligence -- Greenhouse gases -- Process flow sheeting -- Soft sensors -- Sustainability; Factory and trade waste -- Management -- Periodicals; Manufactures -- Environmental aspects -- Periodicals; Déchets industriels -- Gestion -- Périodiques; Usines -- Aspect de l'environnement -- Périodiques; 628.5
• DOI: 10.1016/j.jclepro.2022.134359
• Languages: English
• ISSNs: 0959-6526

Abstract

Abstract: In this study, an artificial intelligence-based framework was developed to monitor the efficiency of the carbon capture unit, i.e., the Sour Compression Unit (SCU) and a cryogenic unit, in a cement manufacturing plant. Initially, an Aspen Plus-based model of the SCU and the cryogenic unit was developed. The process model was then transformed into dynamic mode through the interfacing of MATLAB and Excel with Aspen Plus to capture real-time process behavior. Five hundred fifty data samples were generated by varying the process conditions, i.e., inlet water flow rate, temperature, pressure, etc. The dataset was then used to develop ensemble models. Output compositions of the models were CO2, SO2, and NO. The ensemble models were used as surrogates in Sobol and Fourier Amplitude Sensitivity Test (FAST) frameworks to find the most sensitive process conditions that affect the efficiency of the process. The correlation coefficients based on the ensemble models for estimation of the composition of CO2, SO2, and NO, were 0.9888, 0.9663, and 0.9970, respectively. Based on the sensitivity analysis, the inlet flue gas temperature and pressure were the dominant variables that affect the mole fraction of SO2 and NO in the flue gas. Besides, the CO2 recovery was found heavily dependent on temperature of the flash tanks. The proposed framework was highly accurate and will provide a base for real-time implementation of the concept of smart factory in cement manufacturing plants.