The kinetics of heat transfer in hardening concrete is a key issue in engineering practice for erecting massive concrete structures. Prediction of the temperature fields in early age concrete should allow for proper control of the construction process to minimize temperature gradients and the peak temperatures, which is of particular importance for concrete durability. The paper presents a method of identification of the thermophysical parameters of early age concrete such as the thermal conductivity, the specific heat, and the heat generated by cement hydration in time. Proper numerical models of transient heat conduction problems were formulated by means of finite-element method, including two types of heat losses. The developed experimental–numerical approach included the transient temperature measurements in an isolated tube device and an in-house implementation of an evolutionary algorithm to solve the parameter identification task. Parametric Bezier curves were proposed to model heat source function, which allowed for identifying such function as a smooth curve utilizing a small number of parameters. Numerical identification tasks were solved for experimental data acquired on hardening concrete mixes differing in the type of cement and type of mineral aggregate, demonstrating the effectiveness of the proposed method (the mean-squared error less than 1 °C). The proposed approach allows for the identification of thermophysical parameters of early age concrete even for mixtures containing non-standard components while omitting drawbacks typical for classical optimization methods.

硬化混凝土中的传热动力学是竖立大型混凝土结构的工程实践中的关键问题。对早期混凝土温度场的预测应允许对施工过程进行适当控制，以使温度梯度和峰值温度降至最低，这对于混凝土的耐久性特别重要。本文提出了一种识别早期混凝土的热物理参数的方法，例如热导率，比热和水泥水化产生的热量。通过有限元方法建立了包括两种类型的热损失在内的瞬态热传导问题的正确数值模型。开发的实验数字方法包括在隔离管设备中进行瞬态温度测量，以及在内部实施进化算法来解决参数识别任务。提出了参数贝塞尔曲线来对热源函数建模，这允许使用少量参数将函数识别为平滑曲线。解决了针对硬化混凝土的实验数据的数值识别任务，该混凝土混合物的水泥类型和矿物骨料类型不同，证明了该方法的有效性（均方误差小于1°C）。