Identification of critical parameters influencing resistance performance of amphibious vehicles based on a SM-SA method

The design of high-speed amphibious vehicles needs to consider more factors compared to ships. The efficiency of design and optimization of parameters will not be realized in the absence of a feasible parametric model and design criterion. This paper provided a recognition method for the critical factors that significantly affect amphibious vehicles' resistance. Firstly, an initial parametric model of amphibious vehicles was established, and the resistance coefficients were acquired through numerical simulations. The principal component variables of initial data were extracted by principal component analysis (PCA). Then the functional relations between resistance and principal component variables were obtained respectively through artificial neural network (ANN) and nonlinear polynomial fitting (NPF). Next, two surrogate models were employed to analyze the sensitivity of the resistance to initial parameters. The identified sensitive parameters include the trim angle, loss of waterplane area, and some principal dimensions coefficients. The variation of parameters' sensitivity and their interactions were recognized when parameters are located in different regions. Ultimately, the resistance surrogate model was constructed with critical parameters, enabling the rapid optimization of parameter scheme. Compared with the initial scheme, the optimized scheme achieved significantly reduction of resistance. The extraction and optimization method for critical parameters in this paper provides reference for the design of amphibious vehicles.