Speeding up reactive transport simulations in cement systems by surrogate geochemical modeling: deep neural networks and k-nearest neighbors

This study investigates how reactive transport (RT) simulation can be accelerated by replacing the geochemical solver the RT code by a surrogate model or emulator, considering either a trained deep neural network (DNN) or a k-nearest neighbor (kNN) regressor. We focus on 2D leaching of hardened cement paste under diffusive or advective-dispersive transport conditions, a solid solution representation of the calcium silicate hydrates and either 4 or 7 chemical components, and use the HPx reactive transport code as baseline. We find that after training, both our DNN-based and kNN-based codes, called HPx-DNN and HPx-kNN, can make fairly (7-component cement system) to very (4-component cement system) accurate predictions while providing either a 4 to 7 (HPx-DNN) or 2 to 5 (HPx-kNN) speedup compared to HPx with parallelized geochemical calculations over 4 cores. Benchmarking against single-threaded HPx, these speedups become 13 to 25 and 8 to 18 for HPx-DNN and HPx-kNN, respectively. Defining the maximum possible speedup as the computational gain in RT simulation that would be obtained if the geochemical calculations would be for free, we find that our HPx-DNN code allows for a close to optimal speedup while our HPx-kNN code provides half the maximum possible speedup. We further detail the achieved accuracy and speedup and provide suggestions as how to further improve both aspects.