Optimal design of computer experiments for surrogate models with dimensionless variables

Abstract

This paper presents a method for constructing optimal design of experiments (DoE) intended for building surrogate models using dimensionless (or non-dimensional) variables. In order to increase the fidelity of the model obtained by regression, the DoE needs to optimally cover the dimensionless space. However, in order to generate the data for the regression, one still needs a DoE for the physical variables, in order to carry out the simulations. Thus, there exist two spaces, each one needing a DoE. Since the dimensionless space is always smaller than the physical one, the challenge for building a DoE is that the relation between the two spaces is not bijective. Moreover, each space usually has its own domain constraints, which renders them not-surjective. This means that it is impossible to design the DoE in one space and then automatically generate the corresponding DoE in the other space while satisfying the constraints from both spaces. The solution proposed in the paper transforms the computation of the DoE into an optimization problem formulated in terms of a space-filling criterion (maximizing the minimum distance between neighboring points). An approach is proposed for efficiently solving this optimization problem in a two steps procedure. The method is particularly well suited for building surrogates in terms of dimensionless variables spanning several orders of magnitude (e.g. power laws). The paper also proposes some variations of the method; one when more control is needed on the number of levels on each nondimensional variable and another one when a good distribution of the DoE is desired in the logarithmic scale. The DoE construction method is illustrated on three case studies. A purely numerical case illustrates each step of the method and two other, mechanical and thermal, case studies illustrate the results in different configurations and different practical aspects.

Electromechanical Resonant Ice Protection Systems: Analysis of Fracture Propagation Mechanisms

“…fr,1 for the initial state without fracture, fr,2 after cracking on the entire thickness, fr,3 after delamination of 25% of the surface. The VPLM method [41] and its associated computer DoE plans [42] are used to obtain the approximate analytical expressions of ! for states 2 and 3.…”

Surrogate modeling: tricks that endured the test of time and some recent developments

“…While non-dimensionalization is by itself a variable transformation technique, it can further benefit by a combination with other techniques. Non-dimensionalization combined with exponential scaling transformations has been, for example, shown to be extremely effective at both reducing the dimensionality and improving the accuracy of surrogate models of mechatronic components over a large domain of variation of the input variables (multiple orders of magnitude) [26,92,142,175].…”

Efficient sizing and optimization of multirotor drones based on scaling laws and similarity models

“…π pi = L a pi,0 j p a pi, j i (5) In the case where the dimensionless numbers π y and π di can be assumed constant, such as respectively the one representing the magnetic saturation and the one representing the diameter-length ratio for the motor, then the function f can be approximated by using scaling laws or similarity models [23,24]. Otherwise, when data is available, regression models [25,26] are used such as for modeling the propeller performance or landing gear structural analysis. A detailed description of the hypotheses and methods associated with these sizing models of drone components can be found in [14].…”

“…After several iterations, the final mass, M total, f inal calculated as sum of the different components must converge to this initial guess M total . This is expressed by the following inequality: M total ≥ M total, f inal (26) This condition is added to the global constraints, so that both results tend to a unique solution.…”