Abstract: Affected by factors such as ambient temperature, cooling water, and surface insulation, the actual thermal parameters of arch dam during construction differ remarkably from laboratory test results. Based on temperature data obtained by using distributed optical fiber sensor, we employed the cross-global artificial bee colony (CGABC) algorithm determine the concrete thermal parameters of Baihetan double-curvature arch dam and capture their real-time variations. To address the slow convergence and susceptibility to local optimals encountered by traditional artificial bee colony (ABC) algorithm in obtaining the optimal function value, we developed the CGABC which integrates the concept of global optimal solutions from particle swarm optimization (PSO) and the cross-operation strategy of genetic algorithm (GA). By considering the influence of multi-stage cooling water and environmental temperature, we employed CGABC for the inversion of concrete thermal parameters of Baihetan arch dam. The inversion results demonstrate a favorable agreement between CGABC-calculated values and measured temperatures. In conclusion, CGABC exhibits excellent adaptability in the thermal parameter inversion of arch dams.

Key words: artificial bee colony algorithm, arch dam, thermal parameters, inversion analysis, pipe cooling