Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                

ME270 Final Project Kaneko

Download as doc, pdf, or txt
Download as doc, pdf, or txt
You are on page 1of 5

Jesse Kaneko

Final Project ME270


Prof. Granda

NASA CDSL 6-DOF Platform

Figure 1: NASA CDSL 6-DOF Platform

Physical System and Problem


The 6-DOF Stewart platform at the CDSL (contact dynamics simulation
laboratory) has proven records of docking and berthing simulations over 20
years. The general shape and dimensions can be seen in the picture above
and the payload can be seen in the picture below. However, even with its
proven record, there is still room for improvement in controls so that the
mechanism can simulate contact dynamic motions more accurately. Some of
the controls issues include increasing the bandwidth of the system and
implementing a predictor to compensate for transport and dynamic response
relay. The implementation of an additional feedforward control element to
the existing PD control system is for improving the controller performance.

Even before the PID systems are attained there are many steps in between
that can be done relatively quickly with the use of CAMP-G.

Figure 2: Payload mount of NASA CDSL 6-DOF Platform

Objectives
The objective of this project is to utilize CAMP-G, MATLAB, and
SIMULINK to gain control of the physical motion of the NASA CDSL 6DOF Platform. The bond graph of the NASA CDSL 6-DOF Platform can be
seen in Figure 3. The objective of my thesis is to implement an additional
feedforward control element to the existing PD control system with
minimum modifications (the existing PD servo cards will be used with their
tuned parameters as they are). The additional input current from the added
feedforward element to the hydraulic servo valve of each actuator leg is
superimposed to the input current to the valve from the existing servo card.
For pratical implementation of the added element, the original input current
from the element is multiplied with a weight factor. The weight factor can
vary from 0.0 to 1.0, depending on the effectiveness of the additional

element. This can be a risk-free approach when the weight factor is set to
0.0 because there will be no effect from the added element.

Figure 3: Bond graph of platform

Specifications
The docking and berthing simulator mechanism is composed of two major
hardware components. The first is the Stewart platform where the chaser
vehicle is mounted and the second is the stationary mount where the docking
mechanism of the target vehicle is mounted. There is a force sensor of the
stationary mount to measure contact forces and torques between the two
docking mechanisms. The motion of the two halves is digitally simulated
using the computer. Through the simulated motion, the computer also
calculates the motion of the platform and displacements of the hydraulic
actuator legs.

Figure 4:Physical system in NASTRAN

Procedure
The procedure for this project was to gain full movement control of this
physical system with the use of CAMP-G, MATLAB, and SIMULINK.

Figure 5:MATLAB GUI for platform

The first step was to develop the bond graph (figure 3). Once the bond
graph is created CAMP-G can interface with MATLAB and develop four

programs that, with a little programming, allow one to control the desired
system. Once the interface has been completed the programming must take
place in the m-files. Once the m-files have been altered all of the desired
systems can be controlled. Using these m-files the block diagrams can also
be attained. An example of the m-files can be seen in figure 6.

Figure 6:Mod m-file

Results
For this physical system it was necessary to run the simulation in the time
domain. The use of the m-files gave the necessary differential equations.
With the differential equations now found the motion is also known. The
step response reacts the way one would expect it to along with the bode
diagrams.

You might also like