On DSS Method for a Fast Identification of the Static and Dynamic Responses of Servovalves

TL;DR

The paper tackles rapid identification of both static and dynamic responses of electrohydraulic servovalves by modeling the valve as a quasilinear system $A \frac{d^{2}x}{dt^{2}}+B \frac{dx}{dt}+f(x)=u(t)$ with a monotone nonlinearity $f(x)$ and a static relation $x=f^{-1}(u)$. It introduces the DSS method to extract the static characteristic from short-time data via the parametric form $(x(t),f(x(t)))=(x(t),u(t)-A x''-B x')$, followed by an iterative refinement of $A$, $B$, and $f(x)$ using measured outputs. The approach leverages a near-linear transfer function $K(s)=1/(A s^{2}+B s+C)$ to achieve fast parameter identification with a small set of frequencies, and requires no numerical differentiation of noisy data due to its iterative scheme. The method is shown to produce excellent results in simulations and generalizes to higher-order systems, enabling mobile, low-cost testing and rapid design-cycle analyses for control engineers.

Abstract

In this work we consider a class of quasilinear systems of differential equations which allows to describe dynamics of electrohydraulic servovalves. A method for fast identification of static and dynamic responses, by a short-time experiment, is described.