TY - JOUR
T1 - Efficiency and NOx emission optimization by genetic algorithm of a coal-fired steam generator modeled with artificial neural networks
AU - da Rocha, Bárbara Pacheco
AU - de Assis Brasil Weber, Natália
AU - Smith Schneider, Paulo
AU - Hunt, Julian David
AU - Mairesse Siluk, Júlio Cezar
N1 - Generated from Scopus record by KAUST IRTS on 2023-09-23
PY - 2022/5/1
Y1 - 2022/5/1
N2 - High level of efficiency is pursued whenever an energy convertion system is operated, such as steam generators, but emission contents must be kept within legal and reasonable values at the same time. The interdependence of these two factors often limit the system performance, and optimization procedures allow for finding combinations of parameters that may satisfy that anthagonism and help technical teams to operate actual systems. This study aims to find operating points of an actual superheated steam generator that combine high efficiency with low NOx emissions, based on the system data analysis. The novelty presented in the paper comes from the assessment of a power plant placed on a hot, humid and seasonal wheather, witch differs from previous works. The steam generator efficiency and NOx emission are represented by two independent Artificial Neural Networks (ANNs), based on a common database. The influence of the ANNs selected input paremeters are assessed with the aid of Design of Experiment, which points out that 2 out of 10 inputs to efficiency can be removed, but none to NOx emissions. The objective function aims to find the combination of input values that allow to operate the steam generator with the highest efficiency and the lowest NOx emission, covering three weighting combinations: 50 to 50, 75 to 25 and 90 to 10 in percentage of efficiency in respect to NOx emission. The genetic algorithm optimization procedure identified input values that guaranty 97.95% efficiency and 222.28 mg/mN3 of NOx emission based on the reference values of 98% and 220.00 mg/mN3, when assuming the 90 to 10 ponderation.
AB - High level of efficiency is pursued whenever an energy convertion system is operated, such as steam generators, but emission contents must be kept within legal and reasonable values at the same time. The interdependence of these two factors often limit the system performance, and optimization procedures allow for finding combinations of parameters that may satisfy that anthagonism and help technical teams to operate actual systems. This study aims to find operating points of an actual superheated steam generator that combine high efficiency with low NOx emissions, based on the system data analysis. The novelty presented in the paper comes from the assessment of a power plant placed on a hot, humid and seasonal wheather, witch differs from previous works. The steam generator efficiency and NOx emission are represented by two independent Artificial Neural Networks (ANNs), based on a common database. The influence of the ANNs selected input paremeters are assessed with the aid of Design of Experiment, which points out that 2 out of 10 inputs to efficiency can be removed, but none to NOx emissions. The objective function aims to find the combination of input values that allow to operate the steam generator with the highest efficiency and the lowest NOx emission, covering three weighting combinations: 50 to 50, 75 to 25 and 90 to 10 in percentage of efficiency in respect to NOx emission. The genetic algorithm optimization procedure identified input values that guaranty 97.95% efficiency and 222.28 mg/mN3 of NOx emission based on the reference values of 98% and 220.00 mg/mN3, when assuming the 90 to 10 ponderation.
UR - https://link.springer.com/10.1007/s40430-022-03481-3
UR - http://www.scopus.com/inward/record.url?scp=85129234287&partnerID=8YFLogxK
U2 - 10.1007/s40430-022-03481-3
DO - 10.1007/s40430-022-03481-3
M3 - Article
SN - 1678-5878
VL - 44
JO - Journal of the Brazilian Society of Mechanical Sciences and Engineering
JF - Journal of the Brazilian Society of Mechanical Sciences and Engineering
IS - 5
ER -