Scaling of planar dielectric elastomer actuators in an agonist-antagonist configuration

Abstract

One of the great advantages of dielectric elastomers (DE) is their scalability. This study investigates the influence of size on their performance and the applicability of theory and established models on large-scale actuators. The testing is based on an active hinge configuration made with planar dielectric elastomer actuators as described by Lochmatter in 2007 [16]. The actuators are scaled to five times the size of the small hinge which results in planar membrane actuators of 0.95 × 0.25 m. Theoretical aspects of the up-scaling are discussed. The issue of scaling in design and manufacturing of actuators is addressed. A planar test rig for agonist-antagonist activation, comparable to the hinge-configuration, was set up and the actuators were experimentally characterised in terms of strain (deflection angle), blocking force (moment) and input energy. The behaviour of the large actuators could be predicted accurately with an Arruda-Boyce model and they scale according to theory.

Introduction

As a subgroup of the electro-active polymers, the soft dielectric elastomers have generated much interest because of their high potential for active deformation [1]. Dielectric elastomer (DE) actuators are basically compliant capacitors, consisting of a thin elastomeric film as dielectric, coated on both sides with compliant electrodes. Some of the benefits of this type of actuator include high energy density, low cost, ease of handling and manufacturing [2]. Their simple working principle allows a wide variety of possible designs and applications. Planar membrane actuators especially have been studied in a variety of publications [3], [4], [5], [6], [7], [8]. On many occasions the easy scalability is mentioned as an advantage of these membrane actuators, but has not been proved. Possible applications for large planar actuators are mentioned in [9], [10], [11], [12], [13], [14] including e.g. energy harvesting devices, pumps and space applications. So far, with a few exceptions [12], [14], only small to medium-sized objects have been fabricated and thoroughly characterised. In this paper the issue of scaling is addressed not only in terms of performance, but also in terms of design, manufacturing and modelling. One of the major challenges encountered from the scale-up has been maintaining the mechanical pre-stretch whilst avoiding impairment of the deformation. A different (slower) time-dependent behaviour is expected due to the larger area of the capacitor. The Arruda-Boyce model with material parameters adapted for VHB by Wissler et al. [15] was applied to verify the validity for large-scale actuators.

Large-scale and multi-layered actuators are necessary when large displacements and forces are required. The presented work is embedded in a project with the goal of transferring the method of propulsion of a fish into air to an airship, propelled by moving its body in an undulating way and by flapping a caudal fin. The fish is a very efficient swimmer, being manoeuvrable even at slow speeds whilst moving almost completely noiselessly. DE actuators are ideal for this application because the soft, flexible, light-weight membranes can easily be integrated onto the airship hull. In order to have enough buoyancy the lighter-than-air vehicle must be of a certain size, and in order to create thrust given deformations and forces are required. These were part of the motivation and boundary conditions for the up-scaling of the DE actuators.

The present study is based on the active hinge as published in [16]. The actuators on the small hinge measured 0.19 m in height (H) and 0.05 m in length (L). These actuators are scaled in the planar dimensions by a factor of five to 0.95 x 0.25 m. The thickness of the dielectric membrane or the electrodes is not scaled in order to ensure that the electric field remains unchanged. The total thickness of the actuator is scaled by the number of layers with a factor 3.

In this paper an overview over the theoretical scaling of different parameters is given. The scaling factors S and n are introduced for planar dimensions and number of layers respectively. The actuators are experimentally characterised on a planar test-rig that can be compared to the hinge configuration. Since most applications of agonist-antagonist actuators work with a hinge configuration and for easier comparison with [16], the deflection angle and blocking moment of an equivalent hinge are calculated with simple trigonometry from the evaluated strain and blocking force. Furthermore the input voltage and current is monitored and some electro-mechanical properties (e.g. the efficiency) are listed.