Temperature Control Measures' Intelligent Optimization for Quasi-Mass Concrete Structure

Abstract

Related to temperature control measures and material parameters, crack prevention by temperature control in quasi-mass concrete is an optimization problem including multiple complicated factors. This study tries to do some optimization of temperature control measures in the given concrete thermodynamics parameters circumstance. The quasi-mass concrete structure’s minimum value of the relationship between principal tensile stress duration curves of internal and surface and the tensile strength growth curve are taken as the input, and the gate pier surface heat preservation effect, pouring temperature, water cooling temperature, water cooling time are taken as the output, we establish the optimal temperature control measures by the neural network model. Applying the uniform design principle to combine the temperature control parameters, and using the finite element method to analyze the temperature field and creep stress field in the quasi-mass concrete structure containing cooling water pipes, we obtain the samples, for training the network, to describe the nonlinear relationship between the principal tension stress duration curve and the tension strength growth curve and the different temperature control measures. After inputting the fitting minimum value of relationship between the principal tension stress duration curve and the tension strength growth curve to the trained network, the system is able to automatically select the optimal temperature control measures for crack prevention. The example shows that the proposed optimal neural network model for temperature control measures is feasible.