| Popis: |
Abstract Artificial intelligence has revolutionized various industries, including chemical process optimization. Artificial intelligence (AI) can be applied to various ethylene glycol (EG) production aspects to improve efficiency, quality, and overall process optimization. Process optimization of dimethyl oxalate (DMO) hydrogenation to ethylene glycol (EG) is carried out to minimize/maximize conversion, energy, productivity, bare module cost (CBM), and side product in the presence of pressure and temperature variables. A non-dominated sorting-based multi-objective neural network algorithm (MONNA) is applied to tackle problem optimization for EG production. The results show that the highest productivity, minimum energy cost, side product, and highest conversion are 172 Million RM/year, 0.00600 Million RM/year, 0.0320 kmol/hr, and 99.6%, respectively. The intermediate points in the Pareto Front PF for various three-objective situations provide lower energy costs than the bi-objective function. While bi-objective optimization might seem more straightforward, the intermediate points on the Pareto Front in three-objective optimization can provide lower energy costs due to more effective balancing of trade-offs, richer exploration of the solution space, capturing complex interdependencies, and offering more robust and flexible solutions. Decision makers can use the resulting pareto to decide on the most acceptable alternative according to their preferences. The decision variable plots show that the pressure highly affected the optimal solution with the opposite action. This work’s analysis is predicted to provide insight into optimization literature for energy savings and sustainability and promote commercial growth for the industry. Identifying optimal trade-offs between various objectives, such as energy cost and productivity, can help reduce overall production costs, making the production process more competitive. |
Full text from SpringerLink