@article{01K0420KKENB2A9NEGN2SABGBW,
  abstract     = {{The present study examines the utility of a hybrid model based on the Butler-Volmer Eq. for simultaneously optimizing dosages of multiple corrosion inhibitors. The model is trained to predict corrosion rates measured by linear polarization resistance (LPR). It offers a cheaper and quicker alternative to predicting corrosion inhibition efficiency, which is generally measured by electrochemical impedance spectroscopy (EIS) and potentiodynamic polarization. For the first time, the current study demonstrates how a hybrid model can be used to capture the relationship between corrosion inhibitors and the corrosion rate measured by linear polarization resistance. The hybrid model facilitates separating the benefit of co-occurrence and direct linear correlations from the impact inhibitors have on the parameters of the Butler-Volmer Eq.. Purely data-driven models such as support vector regression (SVR) rely on co-occurrence and direct correlations to make predictions. Therefore, the Sobol sensitivities of inhibitors in purely data-driven models are diminished in the presence of more correlated variables. This study demonstrates how an SVR model fails to predict the corrosion rate when variables less or similarly correlated as the inhibitors are used. An interrupted analysis illustrates how the response of the corrosion rate to each inhibitor varies according to the Sobol sensitivity of the inhibitors in the model. Optimization of inhibitor dosages was carried out using the NSGA II algorithm, facilitating the simultaneous minimization of chemical usage and the corrosion rate. The hybrid model where the surrogate cathodic component (Cbc) was modeled with an SVR and trained with variables selected by magnitude-based partial mutual information, was deemed most suitable for optimizing inhibitor dosages. Thus, the hybrid model is the best candidate for the optimization of multiple corrosion inhibitors added to a cooling water system, enabling the effective mitigation of the corrosion rate.}},
  articleno    = {{146767}},
  author       = {{Jayaweera, Chamanthi Denisha and Hitsov, Ivaylo Plamenov and Van Haeverbeke, Maxime and Tumlos Solon, Kimberly and Gómez Cortés, Cristian Camilo and Depover, Tom and Diekow, Thomas and Verliefde, Arne and Nopens, Ingmar}},
  issn         = {{0013-4686}},
  journal      = {{ELECTROCHIMICA ACTA}},
  keywords     = {{modelling,optimization,corrosion inhibitors,model-based optimization,linear polarization resistance,MILD-STEEL,MECHANISTIC MODEL}},
  language     = {{eng}},
  pages        = {{19}},
  title        = {{Model development and model selection for simultaneous optimization of corrosion inhibitor dosages for a cooling water system}},
  url          = {{http://doi.org/10.1016/j.electacta.2025.146767}},
  volume       = {{537}},
  year         = {{2025}},
}

Altmetric

View in Altmetric

Web of Science

Times cited: