Monitoring-Informed Progressive Decoupled Inversion of Instantaneous and Creep Parameters in High Concrete Dams

Material parameter inversion for high concrete dams is essential for updating numerical models with in-service monitoring data and supporting deformation-based safety assessment. However, the joint inversion of instantaneous and creep parameters is often affected by high-dimensional coupling and parameter response ambiguity, which can reduce the stability of the identified parameters. This study proposes a monitoring-informed progressive decoupled inversion method for the instantaneous and creep parameters of high concrete dams. Combined with orthogonal experimental design, Latin hypercube sampling, and multi-objective particle swarm optimization, surrogate models and multi-objective functions are constructed for the inversion of instantaneous and creep parameters. Based on the water pressure and temperature-related deformation components extracted from monitoring data, the instantaneous parameters are identified first, followed by the inversion of creep parameters using the total deformation response. A progressive decoupled inversion procedure is then implemented by iteratively updating the parameter ranges until no further improvement in the overall fitting accuracy is achieved. Application of a case study shows that the progressive inversion reduces the mean relative error at representative monitoring points and loading conditions from 4.56% to 1.79%. The calculated displacement histories agree well with the measured deformation trends, and the calibrated numerical model produces physically reasonable displacement and stress distributions. The proposed method supports parameter identification and model updating for long-term dam safety assessment.