Concrete Hydration Model Characterization Using Evolutionary Optimization

Abstract

Pile thermal integrity assessment by means of temperature measurement has received increasing attention in recent years. The thermal integrity testing method measures temperature changes during the concrete curing process; using an appropriate concrete hydration model together with tracking temperature development during the curing process, defects within piles could be detected. However, the implementation of thermal integrity testing in practice faces, potentially, many uncertainties including undocumented concrete mixes, lack of knowledge of ground thermal properties, uncertain boundary conditions for pile, etc... These uncertainties increase the complexity of determining appropriate parameters for the hydration model, which directly affects the defect detection capacity of the method. This paper presents an inverse approach using differential evolution (DE) algorithms to determine the concrete hydration model. With this approach, the finite element (FE) analysis is integrated into the DE algorithm to generate approximate solutions that match a controlled dataset instead of approximating the concrete hydration parameters with limited prior knowledge as currently used in practice. Firstly, a field test temperature dataset with a well-defined boundary condition is selected. The temperature development corresponding to the selected dataset is then numerically simulated using an uncalibrated general hydration model. Finally, the hydration model parameters are determined using DE algorithms based on the measured and simulated temperature development as inputs. A field case study is presented in the end of this paper. The results indicate that the proposed inverse approach using DE Pile thermal integrity assessment by means of temperature measurement has received increasing attention in recent years. The thermal integrity testing method measures temperature changes during the concrete curing process; using an appropriate concrete hydration model together with tracking temperature development during the curing process, defects within piles could be detected. However, the implementation of thermal integrity testing in practice faces, potentially, many uncertainties including undocumented concrete mixes, lack of knowledge of ground thermal properties, uncertain boundary conditions for pile, etc... These uncertainties increase the complexity of determining appropriate parameters for the hydration model, which directly affects the defect detection capacity of the method. This paper presents an inverse approach using differential evolution (DE) algorithms to determine the concrete hydration model. With this approach, the finite element (FE) analysis is integrated into the DE algorithm to generate approximate solutions that match a controlled dataset instead of approximating the concrete hydration parameters with limited prior knowledge as currently used in practice. Firstly, a field test temperature dataset with a well-defined boundary condition is selected. The temperature development corresponding to the selected dataset is then numerically simulated using an uncalibrated general hydration model. Finally, the hydration model parameters are determined using DE algorithms based on the measured and simulated temperature development as inputs. A field case study is presented in the end of this paper. The results indicate that the proposed inverse approach using DE algorithms can be used effectively in thermal integrity testing.