Development of an Elbow Actuator
The revolute joint at the elbow of the soft robot was mimicked using an elbow actuator. This actuation was not possible by merely scaling the network actuator because:
- The section modulus of network actuator is very less.
- The load-bearing area of PNA is the entire lower surface, which will not be the case in the elbow joint.
- The stroke required is less, but a considerable amount of force is required.
For this purpose, PNA was reinvented to Elbow actuator. Elbow actuator uses the principle of expansion of bellow for its operation. It contains a sector of 4-fold bellow in one segment and a glass fibre element stretching across the axial direction. This element gives gets in tension and restrains motion in one direction while becomes slack allowing motion in another direction.
Figure 1 Elbow Actuator
While other assumptions from PNA still hold, for simplicity, the fibreglass member was neglected from a simulation. The geometry and simulation setup is shown in the Fig. 2-3
Figure 2 Simulation Setup of Elbow Actuator Figure 3 Simulation setup geometry
The features in the control volume were meshed using quadratic triangular elements as large deformations are expected. The quadratic elements capture curved surface deformations and relatively larger size elements can be used as compared to linear elements for same computation capacity and accuracy. The material for chambers is HT33 (see Chapter 4). To constraint the elongation of actuator a thin polyethene film is introduced in base components (HT33) during molding (see chapter 6).
The simulation was run using a direct solver to reduce the computation time and as only Force and Displacement convergences were needed to be achieved. Due to this, the computation was done in 2 steps to reduce the pressure ramping gradient.
Figure 3 Converged Mesh of Elbow Actuator Isometric View
The results of the simulation are shown in figure 4-5-6
Figure 4 Von-Mises stress in Elbow Actuator
The stroke is limited by the bellow folding dimensions. The below was so designed as it will be completely unfolded by the time it reaches about 90-degree rotation about X-axis. It was intended as such because a fully unfolded bellow provided maximum holding force owing to the high section modulus.
Figure 5 Total deformation at t=0.79458 Figure 6 Total deformation at t=1.111 sec
The below starts inflating thereafter, to avoid this problem structural ribs are needed. Though this will increase resistance to rotation, it will reduce uncontrolled inflation.
The actuator is needed to be analyzed in the time domain to analyses the damping characteristics upon force and jerk loads. For that, a thorough investigation of material properties is needed.
Considering bending in one direction enough for elbow actuator, Single chamber elbow actuator was designed.
It has a single chamber as shown in the figure with a semi-cylindrical curved surface and the flat face assisted with a stiffening Teflon tape layer. There is also a fibreglass tube attached for load-bearing in the opposite direction.
As shown in the figure, the mold designed for the actuator was very intricate in nature and the removal of the inner core of the mold was materially impossible, and thus the inner core of the mold had to be sacrificial in nature.
The material of the 3D printed mold, which was Poly Lactic Acid (PLA) dissolves in Acetone and PDMS rubber, which is non-reacting in nature, has no effect on treatment with Acetone. Thus, after casting the actuator part, the external removable part of the mold was separated and the remaining cured part with the mold core in it was immersed into acetone and allowed to dissolve overnight.
Figure 7 Elbow Actuator after casting of End connectors
Figure 8 Elbow actuator Mold with Inner core
Figure 9 Elbow Actuator after casting of a single end connector
A setup similar to the test setup for the PNA was made for the testing of the elbow actuator. The calibration and force transmission through one point and subsequent measurement on the weighing scale is also similarly shown in the following figures. Different sets of readings were taken and the Time cycle for the actuation was also measured for 20N of actuation force and plotted on the graphs as follows. The maximum deflection achieved in the actuator was observed to be 70 degrees when actuated.
Figure 10 Elbow actuator Bending test
Figure 11 Elbow Actuator test Setup Figure 12 Elbow Actuator bending after Mounting on Robot
Figure 13 Graph of Actuation time vs Pressure for 20N Force
Test results for Elbow Actuator are plotted as shown in the graph (figure 13), on the abscissa pressure is varying from 0 bar to 3 bar and on ordinate Actuation Time is indicated in seconds. Actuation time is higher for the low pressure and as pressure increases actuation time is increasing. For the pressure 1 bar to 3 bar angular velocity of the elbow actuator is varying between 0.21 rad/sec to .40 rad/sec this implies the velocity of arm is varying from 80.5 mm/sec to 160 mm/sec.