A novel inversion approach for seepage parameter of concrete face rockfill dams based on an enhanced sparrow search algorithm

Abstract

In order to rapidly and accurately determine the dam body and foundation permeability coefficients of concrete face rockfill dam, an improved sparrow search algorithm and the Support Vector Machine are combined to propose an ISSA-SVR model. In the ISSA, the Circle mapping initialization population position, periodic convergence factor, and Levy flight are employed to deal with uneven initial population distribution and easily fall into local optimum in the traditional sparrow search algorithm. The effectiveness of the proposed ISSA is verified by six classical test functions. The ISSA has been improved in terms of convergence speed, convergence accuracy, and stability. The simple seepage calculation of the earth-rock dam section proves that the improved ISSA-SVR model has smaller discreteness and higher accuracy than the SSA-SVR model and SVR model in parameter inversion. The ISSA-SVR model is applied in a practical project, and the relative error between the calculated value and the measured value of the water head at each measuring point is within 5 %. The seepage field distribution of the dam body is investigated, and the anti-seepage system blocks 91.9 % of the total water head. The anti-seepage sensitivity analysis of the peripheral joint shows that when the width of the joint reaches 20 mm, the penetration gradient of the cushion exceeds its allowable value. The research results confirm the superiority of the inversion model in the inversion of seepage parameters.