Scribd
Upload
273 views9 pages

EECS Maglev

This document summarizes a student project to design a PID controller to stably levitate a metallic ball using a maglev setup. The controller was able to levitate the ball with less than 10% overshoot and settling time under 3 seconds. The controller was further tested and modified to precisely track square wave inputs of varying amplitudes and mitigate external disturbances. The experiment demonstrated levitation control and provided experience applicable to control of maglev trains.

Uploaded by

Akhil Gupta
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
273 views9 pages

EECS Maglev

This document summarizes a student project to design a PID controller to stably levitate a metallic ball using a maglev setup. The controller was able to levitate the ball with less than 10% overshoot and settling time under 3 seconds. The controller was further tested and modified to precisely track square wave inputs of varying amplitudes and mitigate external disturbances. The experiment demonstrated levitation control and provided experience applicable to control of maglev trains.

Uploaded by

Akhil Gupta
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 9

EECS 460

Maglev Report

Achint Benoyjee Rohit


Akhil Gupta
B. Praveen Varma
Venkatesh Verma

1. Design of the controller


The purpose of this experiment was to stably levitate a metallic ball in our experimental setup. The
plant function for this experiment has been given to us as:

The specifications for this setup are a) the setup has to be asymptotically stable, b) zero steady state
error, c) with settling time less than 3 seconds and d) overshoot of less than 10%
In oder to achieve zero steady error wrt step input, we need atleast a type 1 system. In order to achieve
this in system, we need to integrate an integrator in our controller. Also our characteristic equation in
our plant, does not have s term. Without the s term, the roots will be lying on imaginary axis, which
would cause the output to keep oscillating. Hence to introduce the s term, we use a differentiator in
our controller.
To include all of the above mentioned parameters, we select a PID controller to stabilize the plant and
achieve the desired dynamics.
a) The PID controller design is attached in hand written copy.
b)Attached in the same

2) Performance analysis of the controller


As can be seen below, on using the controller for this plant, the output was stable.

Part 2 (a) Simulation studies with our original controller as :

Precompensator:

11.04

() :

0.613 2 + 3.68 + 11.04

0.613 2 + 3.68 + 11.04

(i) Step Input response for a linearized plant model.


Linearized plant:
Overshoot: 9.95%

3734

2 2180

Settling time: About 1.7 seconds (#compare against calculated value)


Steady State Error: 0

(ii) Step input response for nonlinear plant:

Overshoot: 9.5%
Settling time: 2 seconds
Steady state error = 0

2(b)
The maglev unit consists of a connection interface panel with a mechanical unit on which coil is
mounted. An infrared sensor is attached to the mechanical unit. The electrical units transfer the
measured signal to pc via I/O card. The mechanical and electrical system provide a complete system.

Since the phenomenological model is non-linear function, it has to be linearized. The linearized plant
function is mentioned in the beginning of report. The given plant function is only applicable within a
range of +/- 0.5 cm.
Following is the output with step input. The output is stable and the steady state error is zero after
integrator sets in 15 seconds.

3) Conclusions
The controller was able to track the input step function without any steady state error after 15 seconds,
when the integrator started working in the controller.

With the new controller, the output with step of 0.4 cm amplitude and 8 s period was achieved as
shown below.

4. In our experimental setup, we changed our input to square wave only after the system was stable
with integrator active for step input. The controller was not able to track the square wave precisely. This
could be because the integrator which was selected was quite high initially. Hence the tracking of square
could not be done. In order to achieve this, we calibrated the controller and changed the controller
values to ........
The initial controller had a very high integrator value. Hence the controller was not able to handle the
quick changes (in square wave input) and react to it quickly. We modified the controller by modifying
the integrator and differentiator values by also modifying the position of zeros of our controller to
increase natural frequency factor to improve response time.
Upon modifying the controller, square wave of amplitude 0.4 cm with period 8 seconds was precisely
tracked without any issues by the new controller developed.

For the next trial amplitude was modified to 0.45 cm. Output is as shown below.

For amplitude of 0.5 cm, output is

And with amplitude of 0.6 cm, output is

In order to check the robustness of system to external disturbances, the ball was given external impulse
and the output was plotted. Following is the plot with external disturbances. As can be seen, the
controller was able to mitigate most of the external disturbances without any problem.

Based on the ability of controller to handle these disturbances, this maglev can operate on a shaky base
as well efficiently.
One real life example of magnetic levitation is Mag Lev trains. Now, the trains have not just to levitate
magnetically but also propel forward and brake. In our experiment we designed a system which can
keep a metallic ball levitated and to some extent can endure slight horizontal oscillations. So, as far as
the levitation aspect is concerned the design concept and methodology used in our experiment should
be sufficient. Therefore, apart from our design considerations, following parameters will also have to be
considered for the Train:
Mechanism to accelerate and Brake
Varying mass of the train depending on the number of passengers
Aerodynamic drag consideration
The experiment overall was performed as a team. The initial design of the controller we started off with
was designed by Achint. The experimental analysis and tuning of the controller was jointly done by all
four i.e. Achint, Akhil, Praveen and Venkatesh. Akhil and Achint did most of the physical restraining of
the ball on the apparatus as per the experiments requirements. The final report was compiled by
Praveen and Venkatesh.

You might also like