Application of a physics informed neural network to optimization of chlorine production by membrane-less electrolysis
DOI:
https://doi.org/10.62638/ZasMat2106Abstract
This paper shows how a physics-informed neural network (PINN) can connect Faraday's balance, changing electrolyte conductivity, and experimental data in a model of available-chlorine production. First, a steady-state component balance yields a baseline linear relation between concentration and active electrode length. The paper then shows why assuming constant current density overestimates outlet concentration in a longer cascade electrolyzer: electrolyte conductivity declines during conversion. Six Aqualor H measurements are used to derive the correlation , inspired by semiempirical models of free-ion concentration and mobility. The PINN approximates the spatial concentration profile and learns by reducing the differential-equation residual and inlet-condition error. The plots show approximate agreement between the PINN and theoretical model for the short H25 pilot electrolyzer, while experimental yield is lower because of vessel-volume effects. For the H1000, the PINN follows the experimental profile more closely than the constant-current theoretical model. The enhanced model uses the composite conductivity correlation; numerical solution of its governing equation gives an outlet concentration of about 9.48 g/L for the adopted parameters, versus 10.10 g/L from the baseline model and 9.4 g/L in the displayed experimental seriesThis paper shows how a physics-informed neural network (PINN) can connect Faraday's balance, changing electrolyte conductivity, and experimental data in a model of available-chlorine production. First, a steady-state component balance yields a baseline linear relation between concentration and active electrode length. The paper then shows why assuming constant current density overestimates outlet concentration in a longer cascade electrolyzer: electrolyte conductivity declines during conversion. Six Aqualor H measurements are used to derive the correlation , inspired by semiempirical models of free-ion concentration and mobility. The PINN approximates the spatial concentration profile and learns by reducing the differential-equation residual and inlet-condition error. The plots show approximate agreement between the PINN and theoretical model for the short H25 pilot electrolyzer, while experimental yield is lower because of vessel-volume effects. For the H1000, the PINN follows the experimental profile more closely than the constant-current theoretical model. The enhanced model uses the composite conductivity correlation; numerical solution of its governing equation gives an outlet concentration of about 9.48 g/L for the adopted parameters, versus 10.10 g/L from the baseline model and 9.4 g/L in the displayed experimental series
Keywords:
PINN; artificial intelligence; available chlorine; membraneless electrolysis; Faraday's law; electrolyte conductivity; Aqualor H.References
J. Newman, K. E. Thomas-Alyea (2004) Electrochemical Systems, 3rd ed., Wiley-Interscience, Hoboken, NJ, USA.
A. J. Bard, L. R. Faulkner (2001) Electrochemical Methods: Fundamentals and Applications, 2nd ed., Wiley, New York.
T. F. O’Brien, T. V. Bommaraju, F. Hine (2005) Handbook of Chlor-Alkali Technology, Springer.
D. Pletcher, F. C. Walsh (1990) Industrial Electrochemistry, 2nd ed., Chapman & Hall.
Black & Veatch Corporation (2010) White’s Handbook of Chlorination and Alternative Disinfectants, 5th ed., Wiley, Hoboken, NJ, USA.
H. B. Beer (1980) The invention and industrial development of dimensionally stable anodes, Journal of the Electrochemical Society, 127, 303C–307C.
S. Trasatti (1987) Electrocatalysis in the anodic evolution of oxygen and chlorine, Electrochimica Acta, 32, 369–382.
J. Jeong, C. Kim, J. Yoon (2009) The effect of electrode material on the generation of oxidants and microbial inactivation in electrochemical disinfection processes, Water Research, 43, 895–901. https://doi.org/10.1016/j.watres.2008.11.033
A. Kraft (2008) Electrochemical water disinfection: A short review, Platinum Metals Review, 52, 177–185. https://doi.org/10.1595/147106708X329273
D. Ghernaout, M. W. Naceur (2011) Ferrate(VI): In situ generation and water treatment—A review, Desalination and Water Treatment, 30, 319–332. https://doi.org/10.5004/dwt.2011.2217
X. Chen, G. Chen, P. L. Yue (2000) Separation of pollutants from restaurant wastewater by electrocoagulation, Separation and Purification Technology, 19, 65–76.
C. A. Martínez-Huitle, E. Brillas (2009) Decontamination of wastewaters containing synthetic organic dyes by electrochemical methods. A general review, Applied Catalysis B: Environmental, 87, 105–145. https://doi.org/10.1016/j.apcatb.2008.09.017
C. Comninellis (1994) Electrocatalysis in the electrochemical conversion/combustion of organic pollutants for wastewater treatment, Electrochimica Acta, 39, 1857–1862.
C. Comninellis, G. Chen (Eds.) (2010) Electrochemistry for the Environment, Springer
J. C. Morris (1966) The acid ionization constant of HOCl from 5 to 35°C, Journal of Physical Chemistry, 70, 3798–3805.
S. Kolić, M. Džomba, M. N. KrstajićPajić (2024) HLOROGEN – In Situ Electrochemical Production of Available Chlorine: Characteristics and Optimization, 9th Regional Symposium on Electrochemistry – South-East Europe, Novi Sad, Serbia, June 3–7, p. 21.
HLOROGEN development documentation (2026) Semiempirical Model of the Dependence of Aqualor H Conductivity on Available-Chlorine Concentration: Proposed Equation, Theoretical Rationale, and Statistical Validation, Predlog_ modela_ elektro-provodljivosti_Aqualor_H_Korigovan”, 22 June.
HLOROGEN development documentation (2026) Electrolysis PINNOptimiser – H1000.
W. Zhang, X. Chen, Y. Wang, L. Wu, Y. Hu (2020) Experimental and Modeling of Conductivity for Electrolyte Solution Systems, ACS Omega, 5, 22465–22474. https://doi.org/10.1021/acsomega.0c03013
K. L. Gering (2017) Prediction of Electrolyte Conductivity: Results from a Generalized Molecular Model Based on Ion Solvation and a Chemical Physics Framework, ElectrochimicaActa, 225, 175–189. https://doi.org/10.1016/j.electacta.2016.12.083
Y. Zhang, I. Bier, V. Viswanathan (2022) Predicting Electrolyte Conductivity Directly from Molecular-Level Interactions, ACS Energy Letters, 7, 4061–4070. https://doi.org/10.1021/acsenergylett.2c01947
E. M. Raissi, P. Perdikaris, G. E. Karniadakis (2019) Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations, Journal of Computational Physics, 378, 686–707. https://doi.org/10.1016/j.jcp.2018.10.045.
F. Gamry Instruments, Electrolyte Resistance, technical documentation. https://www.gamry.com/support-2/






