Zoned elasticity modulus inversion analysis method of a high arch dam based on unconstrained Lagrange support vector regression (support vector regression arch dam)

An inverse modeling process is proposed to estimate the actual elasticity modulus of a high arch dam based on unconstrained Lagrange support vector regression (ULSVR) and monitoring data. The proposed ULSVR eliminates the inequality and equality constraints to improve the iteration speed, especially for large repeated calculations, and the process takes advantage of the culture genetic algorithm (CGA) by importing the belief space, which utilizes social knowledge to guide the generations to accelerate the evolution speed. Then, CGA is used to optimize the parameters of ULSVR and seek the optimal elasticity modulus combination. The inverse modeling procedure is successfully applied to optimize the zoned elasticity modulus of the Jinping high arch dam and foundation in its initial impound period. The inversion analysis shows that the actual modulus is larger than the experimental results, showing that the actual situation is safer than the design condition.