Design of Planar Nonlinear Springs for Prescribed Load-Displacement Functions

2007 ◽  
Author(s):  
Christine V. Jutte ◽  
Sridhar Kota
Keyword(s):  
Parametric Model ◽  
Nonlinear Finite Element ◽  
Algorithm Optimization ◽  
Optimization Scheme ◽  
Element Analysis ◽  
Single Plane ◽  
Nonlinear Spring ◽  
Vehicle Suspensions ◽  
Nonlinear Springs ◽  
Load Displacement

Nonlinear springs can simplify and improve the performance of a variety of devices, including prosthetics, MEMS, and vehicle suspensions. Each nonlinear spring application has unique load-displacement specifications that do not correspond to one general spring design. This limits the use of nonlinear springs and thus compromises the performance of these applications. This paper presents a generalized methodology, including topology, size, and shape optimization, for creating nonlinear springs with prescribed load-displacement functions. The methodology includes a new parametric model that represents nonlinear springs as a single-plane, ‘fractal’-like network of splines. The parametric model and the objective function are incorporated into a genetic algorithm optimization scheme. Nonlinear finite element analysis evaluates the large displacements of each spring design. Three nonlinear spring examples, each having uniquely prescribed load-displacement functions including a “J”-shaped, an “S”-shaped, and a constant-force function, generate designs that demonstrate the methodology’s effectiveness in designing nonlinear springs.

Design of Nonlinear Springs for Prescribed Load-Displacement Functions

2008 ◽  
Vol 130 (8) ◽  
Author(s):  
Christine Vehar Jutte ◽  
Sridhar Kota
Keyword(s):  
Displacement Function ◽  
Curved Beams ◽  
Element Analysis ◽  
Nonlinear Load ◽  
Nonlinear Spring ◽  
Nonlinear Springs ◽  
Bending Stresses ◽  
Bending Loads ◽  
Load Displacement ◽  
Synthesis Methodology

A nonlinear spring has a defined nonlinear load-displacement function, which is also equivalent to its strain energy absorption rate. Various applications benefit from nonlinear springs, including prosthetics and microelectromechanical system devices. Since each nonlinear spring application requires a unique load-displacement function, spring configurations must be custom designed, and no generalized design methodology exists. In this paper, we present a generalized nonlinear spring synthesis methodology that (i) synthesizes a spring for any prescribed nonlinear load-displacement function and (ii) generates designs having distributed compliance. We introduce a design parametrization that is conducive to geometric nonlinearities, enabling individual beam segments to vary their effective stiffness as the spring deforms. Key features of our method include (i) a branching network of compliant beams used for topology synthesis rather than a ground structure or a continuum model based design parametrization, (ii) curved beams without sudden changes in cross section, offering a more even stress distribution, and (iii) boundary conditions that impose both axial and bending loads on the compliant members and enable large rotations while minimizing bending stresses. To generate nonlinear spring designs, the design parametrization is implemented into a genetic algorithm, and the objective function evaluates spring designs based on the prescribed load-displacement function. The designs are analyzed using nonlinear finite element analysis. Three nonlinear spring examples are presented. Each has a unique prescribed load-displacement function, including a (i) “J-shaped,” (ii) “S-shaped,” and (iii) constant-force function. A fourth example reveals the methodology’s versatility by generating a large displacement linear spring. The results demonstrate the effectiveness of this generalized synthesis methodology for designing nonlinear springs for any given load-displacement function.

Generalized Synthesis of Nonlinear Springs for Prescribed Load-Displacement Functions

2006 ◽  
Author(s):  
Christine M. Vehar ◽  
Sridhar Kota
Keyword(s):  
Shape Function ◽  
General Scheme ◽  
Stress Constraints ◽  
Displacement Function ◽  
Algorithm Optimization ◽  
Load Range ◽  
Element Analysis ◽  
Nonlinear Load ◽  
Nonlinear Springs ◽  
Load Displacement

A spring’s nonlinear load-displacement function is described by three factors, the (i) shape function, (ii) load-range, and (iii) displacement-range. The shape function encompasses the nonlinear relationship between the load and displacement, and therefore, is the most difficult factor to match. In this paper, we present a general scheme for topology, size, and shape optimization of nonlinear springs for prescribed load–displacement shape functions, while simultaneously meeting manufacturing, space, and stress constraints. This paper presents the objective function and a novel, floating point parametric model used within a genetic algorithm optimization scheme. The nonlinear springs all undergo large deformations and are evaluated by nonlinear finite element analysis. Two examples are included to demonstrate the effectiveness of the methodology in synthesizing nonlinear springs that match a prescribed load-displacement shape function.

scholarly journals The information complexity of learning tasks, their structure and their distance

2021 ◽  
Author(s):  
Alessandro Achille ◽  
Giovanni Paolini ◽  
Glen Mbeng ◽  
Stefano Soatto
Keyword(s):  
Kolmogorov Complexity ◽  
Large Scale ◽  
Parametric Model ◽  
Training Dataset ◽  
Optimization Scheme ◽  
Learning Tasks ◽  
Asymmetric Distance ◽  
Special Cases ◽  
Scale Models ◽  
Real World Datasets

Abstract We introduce an asymmetric distance in the space of learning tasks and a framework to compute their complexity. These concepts are foundational for the practice of transfer learning, whereby a parametric model is pre-trained for a task, and then fine tuned for another. The framework we develop is non-asymptotic, captures the finite nature of the training dataset and allows distinguishing learning from memorization. It encompasses, as special cases, classical notions from Kolmogorov complexity and Shannon and Fisher information. However, unlike some of those frameworks, it can be applied to large-scale models and real-world datasets. Our framework is the first to measure complexity in a way that accounts for the effect of the optimization scheme, which is critical in deep learning.

Sign in / Sign up

Export Citation Format

Share Document