Temperature Prediction of Mass Concrete During the Construction with a Deeply Optimized Intelligent Model

In the construction of ultra-high voltage (UHV) transformation substations, mass concrete is prone to temperature-induced cracking due to the temperature difference between internal and external environments. These cracks significantly affect the load-bearing capacity of the foundation. Therefore, accurate temperature prediction and effective temperature control are critical challenges that must be addressed. To better capture temperature fluctuations and enable real-time prediction and control, this study proposes a novel hybrid model named CPO-VMD-SSA-Transformer-GRU for predicting concrete temperature. Firstly, sine functions with different sample sizes were simulated using three models: Transformer-GRU, VMD-Transformer-GRU, and CPO-VMD-SSA-Transformer-GRU.It can be observed that the CPO-VMD-SSA-Transformer-GRU model achieves higher accuracy and faster convergence toward theoretical values, and then five performance metrics have been evaluated named Mean Absolute Error (MAE), Mean Square Error (MSE), Root Mean Square Error (RMSE), and Coefficient of Determination (R²). Furthermore, the proposed deep-optimized model was applied to predict temperature variations in mass concrete. For a single time series in the lab, the evaluation metrics for Checkpoint 1 and Checkpoint 2 were respectively 0.033736, 0.0018812, 0.0013051, 0.036127, 0.98832 and 0.016725, 0.00091304, 0.00036536, 0.019114, 0.96773, indicating corresponding reliability in predicting high-dimensional and nonlinear temperature behavior. In addition, the proposed model was extended to multi-variable time series for practical cocrete construction. The MAE, MSE, RMSE, and R² values for Checkpoint 1 were respectively 0.56293, 0.34035, 0.58339, 0.95414 and 0.85052, 0.78779, 0.88757, 0.91385. These results show improved performance compared to the single-variable model, further validating the high accuracy and reliability of the multi-variable CPO-VMD-SSA-Transformer-GRU.Therefore, the multi-variable model effectively captures fluctuation characteristics under complex conditions, obtaining accurate temperature prediction. The deep investigations provide an effective solution for temperature prediction and corresponding optimized control in real construction environments.