We found a match
Your institution may have access to this item. Find your institution then sign in to continue.
- Title
BLAST FURNACE SLAG FOR SO<sub>2</sub> CAPTURE: OPTIMIZATION AND PREDICTION USING RESPONSE SURFACE METHODOLOGY AND ARTIFICIAL NEURAL NETWORK.
- Authors
KOHITLHETSE, ITUMELENG; EVANS, SUTER KIPLAGAT; BANZA, MUSAMBA; MAKOMERE, ROBERT
- Abstract
The main reaction parameters examined were the amount of blast furnace slag, the hydration duration, ammonium acetate concentration, and temperature. The Response surface methodology was employed to quantify their impact on the sorbent's surface area. Using a central composite design, the surface area of the resulting sorbent corresponding to Brunauer-Emmett-Teller (BET) was investigated. The sorbents produced range in surface area from 49.89 to 155.33 m2/g. Additionally, the effectiveness and response prediction capacities of the Response Surface Methodology (RSM) and Artificial Neural Network (ANN) modeling methodologies were investigated. The models were assessed using various statistical metrics, including (MSE) mean squared error, (ARE) average relative errors, the (SSE) sum of squared errors, (HYBRID) Hybrid fractional error function, (SAE) Sum of the absolute errors, (R2)coefficient of determination, and Root means square. According to statistical evidence, the ANN approach surpassed the RSM-CCD model approach. The surface area of the sorbent was shown to be significantly influenced by interactions between variables in addition to all the individual variables examined. The sorbent was made from a material with substantial structural porosity based on SEM. The functional groups were identified using FTIR. The XRF determined the elemental composition of BFS-based sorbents.
- Subjects
ARTIFICIAL neural networks; RESPONSE surfaces (Statistics); SLAG; BLAST furnaces; AMMONIUM acetate; ERROR functions
- Publication
Chemical Industry & Chemical Engineering Quarterly, 2024, Vol 30, Issue 4, p349
- ISSN
1451-9372
- Publication type
Article
- DOI
10.2298/CICEQ230717006K