Water-energy nexus: Condition monitoring and the performance optimization of a hybrid cooling system
The steel industry is one of the highest water-intensive sectors. To reduce water consumption in the cooling process of this sector, hybrid cooling systems are proposed. As these systems consume water and energy simultaneously, their operation management needs to be done dynamically, considering wat...
Guardado en:
| Autores principales: | , , |
|---|---|
| Formato: | article |
| Lenguaje: | EN |
| Publicado: |
KeAi Communications Co., Ltd.
2021
|
| Materias: | |
| Acceso en línea: | https://doaj.org/article/442f695369954b45bb313a1a4c383704 |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| Sumario: | The steel industry is one of the highest water-intensive sectors. To reduce water consumption in the cooling process of this sector, hybrid cooling systems are proposed. As these systems consume water and energy simultaneously, their operation management needs to be done dynamically, considering water-energy nexus. In the present research, considering regional water and energy scarcities, an operation optimization framework is proposed for the operation management of a direct reduction unit cooling system in a Steel Company. As the behavior of hybrid cooling systems varies over time under the influence of mechanical depreciation and change of environmental conditions, its modeling must be done in a dynamic and precise manner to optimize system performance. In the current study, system modeling is performed by using physical laws (white-box modeling) and machine learning techniques (black-box modeling). Machine learning has been used to modify the deviation of the white-box model from the system situation being caused by equipment degradation. Coupling a dynamic black-box model with the white-box model results in increased accuracy of about 53%. Application of the developed dynamic model, in combination with the proposed framework, has shown that water and energy loss rates could be reduced by 83%; and leads to an 85% saving in possible production reduction. This significant improvement is achieved by the hybrid model's precise prediction of outlet water temperature with 0.91 °C root mean square error; Therefore, using the developed model could help in the improvement of the hybrid cooling system's water and energy efficiency. It is also demonstrated that the model might act as a self-learning model which becomes more precise over time. |
|---|