Single plane graphical balancing tutorial

http://onlinevib.yolasite.com/balancing-tutorial-1-1.php
http://www.cbmapps.com/
Introduction

This is a step by step tutorial to explain the graphical method to perform single plane
balancing of a rigid rotor, applied to the online balancing simulator.

Why single plane graphical balancing?

Because it will give you insight on the balancing process, which is important even if you
later prefer or need to use other methods.
Because it's easy and can be performed with simple tools.
Because it works. In many real life situations you don't need anything more complex.

Warning

Balancing a real machine is a delicate job that requires training and experience. Be sure
to know all the risks involved before attempting to balance any particular machine.

Single plane balancing procedure

 Set up the simulation and a reference frame.
 Measure initial vibration, which is supposed to be caused by an unknown
unbalance weight.
 Add a (know) trial weight and measure its effect.
 Knowing the effect of the trial weight, calculate the weight necessary to cancel
rotor unbalance.
 Place calculated weight and check if it vibration is within tolerance.
 Repeat process to refine if necessary.

You will need:

Polar plot paper (download and print from here).

A ruler and a triangle or set square.
Color pens or pencils.
The balancing simulator.

Set up the simulation and reference frame

Simulation set up

Open the balancing simulator in a new window and set up your rotor.
The rotor geometry and other parameters are up to you, just make sure you select
counterclockwise turning sense and static initial unbalance. In this example we are going
to use the following set-up:

Balancing simulator set-up

For the simulation, you don't need to set up a narrow rotor like this. As long as you select
static initial unbalance the simulator will produce an unbalance suitable to be corrected
using the method described in this tutorial. Keep in mind that in real life, wider rotors
will more likely present a couple unbalance component and narrow rotors are the ones
most likely to be balanced in a single plane.

Reference frame

We are going to represent the vibration and weights position in the same plot, this helps
to interpret the physical location of the balancing weight in the rotor.

Your polar plot paper should start like this:

Initial polar plot

Mark the location of the vibration sensor relative to the rotor and set angles in the
circumference starting at this mark, increasing in the direction opposite to the turning
sense. Also mark the turning sense of the rotor with an arrow.

In the simulator, the vibration sensor is located horizontally at the left (side view) and as
we chose C.C.W. rotation, for this example the plot will look like this:
Polar plot with reference frame

Now the plot represents the rotor as viewed when the phase reference signal is generated,
which for the simulator means when the keyway is horizontal to the left.

Measure initial vibration

Start the rotor and measure initial vibration

Choose how to measure vibration. The simulator allows velocity or displacement units;
you can choose whichever you prefer but once chosen you should stick with it. For this
example we will be using mm/s rms.

Start the rotor and take vibration data, we will call this the "first run", in this example the
values are:

First run vibration values

Yours will be different even if you set up a similar rotor because the initial unbalance in
the simulator is random. You will notice that vibration values at both bearings are almost
in-phase (which is to expect from the chosen geometry) and that amplitude and phase are
not perfectly stable, this happens often in real life, specially with severe unbalances at the
beginning of the process.

As we are performing single plane balancing we need only one vibration measure.
Choose one bearing, the one with the highest vibration if they are not equal. In this
example that would be the right bearing with 13 mm/s rms @ 34º.

Plot initial vibration

Choose a scale for the plot. In this example the radius of the polar plot circumference (the
largest vector that can be drawn) is 8 cm, so we will use a scale of 1 cm for each 2 mm/s
(2:1), so 13 mm/s would correspond to 6.5 cm in the plot.

Using the ruler or triangle, draw the initial vibration vector and label it "O".
First run vibration data plotted

Remember this is just a vector representation of the vibration data and it doesn't mean
vibration is occurring in that direction nor nor that it is higher in that direction; vibration
plotted this way means that when the phase reference signal is triggered, the vibration
peak is occurring at the angular position indicated by the vector.

Add trial weight and measure its effect

Add trial weight

Now we will temporarily add a know weight to the rotor so we are able to measure its
effect on vibration. As we are simulating a static unbalance we will place the weight in
the central (axial) plane of the rotor. The exact weight and position are not important, for
now the goal is to achieve a significant vibration change.

Trial weight estimation can be a complex issue when balancing certain machines.
Whenever possible, consult previous balancing reports for the same or similar machines.
If unsure, be safe and use a light weight.

In our example, we add a weight of 50g @ 120º in the central plane in the periphery of
the rotor.
Adding trial weight to rotor

Remember to follow the established safety procedures and to properly isolate, lockout
and tagout all sources of hazardous energy when working in a real machine.

In a real life job it is extremely important to firmly attach the trial weight to the rotor, a
high speed flying trial weight represents a serious safety hazard.

Register the trial weight location in the polar paper, in the exterior of the circumference,
like this:

Plotting trial weight location

Measure the effect of the trial weight

Start the rotor and register the new vibration value (at the same bearing you chose for the
first run), we call this the "test run", in this example it is 9.4 mm/s rms @ 110º.

Now plot the new vibration vector (4.7cm length in 2:1 scale in the example). This
vibration is the response to the original unbalance and the trial weight, so we label it
"O+T".

Trial run vibration data plotted

Now we are going to determine the contribution of the trial weight itself, which is the
vector difference between (O+T) and O:

T = (O+T) – O

Graphically we trace a vector from O to O+T, and that would be T.

Effect of trial weight plotted

To quantify the magnitude of T, we measure its length with the ruler, for the example it is
7 cm, that is to say 14 mm/s rms in our 2:1 scale.

If the magnitude of T is not at least, say, 20% of O, you should try with a heavier weight.

Now, using the ruler as a guide and sliding the triangle as shown in the picture, translate
T to the origin of the plot to measure its angle.
Translating effect of trial weight to the origin

In the example, the measured angle was 174º, so we have determined that the effect of a
50g mass at 120º is a vibration of 14 mm/s rms at 174º.

Calculate balancing weight

Desired effect of balance weight

Normally we would like to completely eliminate rotor's unbalance, so our objective is to

find a balancing weight that causes an effect just opposite to that of the initial unbalance,
one that would produce on its own a vibration of the same amplitude of O and 180º apart.

Let's call this "−O" and plot it with the same magnitude of O but opposite angle (in the
example, we add 180º to the angle and get 214º).
Effect needed to cancel initial unbalance

Calculate balancing weight

The trial weight of 50g produced a response (change in vibration) of magnitude 14mm/s,
so if we want now a response of magnitude 13 mm/s, we should use a balancing weight
of 50g * 13mm/s / 14mm/s = 46.4g. Use a similar cross multiplication to solve for the
balancing weight in your simulation.

As for the angle, the response T of the trial weight placed at 120º (in the rotor's reference
frame) had an angle of 174º (vibration phase), so if we want a response at 214º we must
place the balancing weight at 214º − 174º = 40º from the position of the trial weight, that
is 160º. Find the angular distance between T and −O for your data and add it to the trial
weight's angle to calculate the required balancing weight angular position.

At this point it is important to validate these calculations with the plot. In the example it
is apparent that we need a slightly lower weight (T is slightly longer than −O) and that we
need to move the weight clockwise from its current position, this is consistent with the
calculated weight of 46.4g @ 160º.
Calculate balancing weight

Place balancing weight and verify

Remove trial weight

It is easy to calculate the balancing weight to attain the desired effect of canceling the
initial unbalance and also to cancel the effect of the trial weight (so there is no need to
remove the trial weight, this can save time if placing or removing balancing weights
takes too long). The calculations we just did do not take this into account, if you
calculated the balancing weight as outlined above, you absolutely need to remove the trial
weight.

Add balancing weight

Add the new correction weight, W. For the example 46g @ 160º in the same axial plane
and radius we added the trial weight.
Remember to follow the established safety procedures and to properly isolate, lockout
and tagout all sources of hazardous energy when working in a real machine.

Start rotor and verify

Start the rotor again to test the result of this balancing weight. In the example the
vibration was reduced to 0.8 mm/s rms @ 137º in the right bearing, and half of that in the
left bearing, which corresponds to zone A of ISO 10816 standard (typical of new
machinery).

Vibration with balance weight

Repeat process to refine if necessary

Chances are by now your simulation is also in tolerance, but you might still need (or
want) to refine the balance of your rotor, in this case you can place a new trial weight and
start the process again or use the last weight as trial weight and the last vibration value as
trial run, T.