E-BOOK

Motor current signature

analysis explained
Find out how the MCSA condition monitoring technique uses current and voltage to detect
upcoming failures at an early stage, and what makes it stand out from other technologies.
Contents MCSA + voltage for smart industry 03

How MCSA works: capture 04

How MCSA works: conversion 06

How MCSA works: analysis 07

AI = analysis, supercharged 09

Detect & localize 10

MCSA rules electric 11

Tracking terabytes over time 13

Automatic operating point classification 14

Real-time performance & energy metrics 16

Conclusion 17

Contact 17

E-book | Motor current signature analysis explained

 03
Induction Motor Fault Diagnosis by Motor Current
Signature Analysis and Neural Network Techniques

MCSA + voltage for smart industry

MCSA stands for motor current signature analysis. Where other

condition monitoring technologies analyze vibrations or oil or
temperature, MCSA analyzes AC current. Modern MCSA systems
add voltage to the basic technique, which raises sensitivity and
accuracy. For the rest of this explainer, we'll use the acronym MCSA
to refer to systems that capture both current and voltage.

MCSA is based on the fact that subtle changes in an asset's

operation affect the connected motor’s magnetic field, which
affects the incoming current (and voltage). The technique has a
long track record. It was first suggested in the 1970s, to reliably
monitor motors in nuclear reactors from a safe distance. Since then
a vast body of research has proven MCSA's value for a wide range of
industrial equipment, from agitators to yaw systems. Today,
companies in diverse industries are using MCSA to detect over 90
percent of critical equipment failures up to five months in
advance.

MCSA can't do everything, but it can do a lot—and there are things

only MCSA can do. That makes it a condition monitoring technique
every 21st-century company should have in its maintenance toolkit.
Figure 1. MCSA is proven technology.
Now let's zoom in on how MCSA works.

E-book | Motor current signature analysis explained

 04
Motor control

How MCSA works: capture

cabinet (MCC)

The first step in any smart condition monitoring program is

to install permanent sensors to continuously capture Motor starter

high-frequency data. Vibration-based systems stick

accelerometers on the equipment to measure how it’s
jostling around; temperature-based systems hang infrared
cameras near the equipment to take pictures of the heat
coming off it.

MCSA systems are different from all these other systems in

To the motor
that MCSA sensors do not need to be near the actual in the field
equipment. MCSA uses current transformers and voltage Power supply Gateway Switch Data acquisition
device (DAQ)
taps that install in the motor control cabinet, where they
capture all three phases of the current and the voltage at a
high frequency around the clock.

So, right there you have MCSA’s first major advantage. It

doesn’t matter whether your pump is submerged in
wastewater or you’re running 2,000-degree molten steel
over your rollers; MCSA sensors are easy and safe to Voltage Current
install, and they’re shielded from operational hazards. sensor sensor

Figure 2. MCSA sensors click onto the electrical wires in the motor control cabinet,
protected from the hazards of the production floor.

E-book | Motor current signature analysis explained

 05

MCSA sensors install in the motor

control cabinet, not on the asset,
making them easy and safe to
install, and shielding them from
operational hazards such as
chemicals, heat and water.

E-book | Motor current signature analysis explained

 06

How MCSA works: conversion

The next step is to translate the current and voltage

measurements from the time domain to the frequency
domain. This breaks down the incoming signal into all its
component frequencies. You can think of the resulting graph
as the signal’s “signature”—the S in MCSA.

Vibration analysis starts with the same conversion, by the

way. It's called the fast Fourier transform, or FFT, and it's
been around since the 1960s. It's a hugely valuable tool: the
world's largest association of engineers voted it one of the
top 10 algorithms of the 20th century, and it's even been
called "the most important numerical algorithm of our
lifetime."

The FFT produces a series of graphs that tell you how each
component frequency contributes to the total energy in the
current or voltage sine wave. Each graph is a snapshot of the
signal's "signature" at a single moment in time. As you take
more and more measurements, this signature will change
over time. Now let's see how MCSA analyzes those changes
to detect developing problems.
Figure 3. MCSA uses the fast Fourier transform to convert the current and voltage samples into a
frequency spectrum. In the bottom graph, the height of each point tells us how much energy that
frequency contributes to the corresponding sine wave in the top graph at one moment in time.

E-book | Motor current signature analysis explained


How MCSA works: analysis
In the graph below, we’ve drawn a particular motor's current frequency
signature at two different points in time: the newer one in blue, the
older one in black. You can see they don’t precisely line up. That tells
you that something has changed.
Now sometimes that will be a normal process change—maybe you’re
running the motor faster, or you’re pumping a thicker fluid. But you’ll
also see changes when something inside the equipment changes: a
bearing starts to wear out, for example, or a coupling starts to loosen,
or the insulation covering the windings in the motor’s stator starts to
degrade.
Figure 4. Comparing the spectral energy for a healthy motor (black) and that same motor with a broken rotor bar (blue).
E-book | Motor current signature analysis explained
07
So the first way that MCSA detects developing problems is to track
these changes over time and compare them with a library of known
failure mechanisms. We call these “fingerprints of failure.” In the
graph we’re looking at here, the blue peaks at multiples of this
motor’s rotational frequency are the hallmark of a broken rotor bar.
(There's nothing unique going on here—this is what every condition
monitoring technology does. There are identifiable fingerprints of
failure for oil, heat, vibration, sound, and so on. MCSA just uses this
established technique on a different source of data, with different
“fingerprints.”)
 08

MCSA remotely detects over 90 percent of

developing failures up to five months in advance,
so you can schedule precisely the maintenance
that's needed, in optimally planned windows.

E-book | Motor current signature analysis explained

 09

AI = analysis, supercharged

Here’s a second asset, whose visual signature is quite

different from the one on page 5. So part of the challenge in
condition monitoring (no matter what kind of data you're
using) is to identify what a “fingerprint of failure” looks like
for each unique machine.

That's why a good MCSA system also performs data-

driven analysis, and not just model-based analysis. "Data-
driven" means the system learns what healthy behavior
looks like for that specific equipment, so when changes start
to happen, the system can distinguish right away between
normal variation and changes that signify developing
damage.

This is where you absolutely have to have a system based

on artificial intelligence. There’s no way any human being,
no matter how expert, can keep up with the volume of
information this depth of analysis requires.

Now let's look at a specific industrial asset for some

concrete examples.
Figure 5. Every motor and machine has a different healthy pattern, which the condition
monitoring system must learn if it's going to detect developing faults early enough to help.
Recent advances in machine learning have brought this level of detection within reach.

E-book | Motor current signature analysis explained

 10

Detect & localize

A rise in the noise floor around A rise at the motor's rotational

Here we see some potential failures in a the supply frequency is typical of frequency and its harmonics
specific industrial asset—a pump that’s pump cavitation. plus a rise in the noise floor is
typical of a coupling fault.
cavitating, a misaligned coupling, some
damage to a bearing. Each fault has a specific
effect on the current and voltage waveforms.

For example, when a bearing starts to degrade,

the energy will start to rise at one or more of
the frequencies associated with the bearing’s
physical dimensions: the fundamental train (or
cage) frequency, the ball pass inner and outer
race frequencies, and the ball spin frequency.

These changes cause vibrations that influence

the air gap between the motor’s stator and
rotor, causing the magnetic field disturbances
we mentioned on page 3. That means MCSA
can detect mechanical faults throughout the
entire drive train, just by measuring the
current and voltage to the motor. And because
these failure fingerprints are all distinct, MCSA
can also localize mechanical faults to the
component that's degrading.
asset in the field motor control cabinet
Figure 6. MCSA can detect and localize mechanical
faults in diverse parts of the connected asset.
A rise at a bearing's cage frequency
is typical of bearing degradation. If
the rise were at a different frequency,
the MCSA system would know the
problem was in a different bearing.

E-book | Motor current signature analysis explained

 11

MCSA rules electric

A rise in the noise floor around A rise at the motor's rotational

So far, so good: MCSA is great at detecting the supply frequency is typical of frequency and its harmonics
and localizing mechanical faults. Now what pump cavitation. plus a rise in the noise floor is A stator winding short circuit
typical of a coupling fault. typically shows a rise at the third
about electrical faults, which are
current harmonic.
responsible for roughly 30% of industrial
motor failures?

Here's where MCSA really stands out. MCSA

will detect electrical faults far earlier than
any other condition monitoring
technology.

That's because electrical changes directly

affect the motor’s magnetic field, long
before you can measure a physical change
asset in the field motor control cabinet
in vibration (or a rise in temperature, or a
change in oil composition). By measuring
current and voltage, MCSA has direct access
to the very first sign that damage is starting
to occur.

A rise at the motor's rotational MCSA systems can instantly identify supply-side problems
frequency and its harmonics such as voltage and current unbalance, harmonic distortion,
Figure 7. MCSA will detect electrical faults far sooner plus a rise in the noise floor is and power quality issues.
than other techniques. typical of a coupling fault.

E-book | Motor current signature analysis explained

 12

MCSA is uniquely suited to

catch electrical faults as
soon as they start to
develop.

E-book | Motor current signature analysis explained

 13

Tracking terabytes over time

So far, we've only seen graphs that show the electrical

signature at two reasonably distant points in time. If it were
up to us humans to process the data, that's what we'd have
to do—throwing out mountains of valuable data in between,
and hoping that we check in often enough to stop trouble
while it's still small.

But machine learning algorithms have three superpowers:

first, they're at work 24/7/365, processing terabytes of data
without cease. Second, they're able to detect tiny changes
the human observer would miss. Third, they're constantly
improving their predictions based on new data. All this
means that AI-based systems consistently close in on
perfection—catching all true positives and flagging no
false positives—the longer they run. AI systems never have
a bad day or forget what they've learned.

This is what makes 21st-century condition monitoring

systems so scalable. The AI software does all the heavy
lifting, only alerting the monitoring system's data scientists
when there's an actual anomaly that needs to be reviewed
Figure 8. Three MCSA graphs tracking changes at a specific fault frequency over time, for three
and communicated to the customer. different situations: bearing damage, an elevator guide rail failure, and a broken rotor bar.
The colored vertical lines show where the AI flagged a new status (see legend). A return to
green indicates the client resolved the problem through maintenance.

E-book | Motor current signature analysis explained

 14

Automatic operating point classification

Many industrial assets run at different operating points. These changes

can run the gamut from occasional and planned (a fan that runs faster
when more ovens are in production) to constant and unpredictable (a
wind turbine subject to changing wind speeds and directions).

Changes in load and speed distort pretty much every kind of data you
can use to monitor asset health. That makes it important to know
whether a change in the incoming data reflects a developing problem,
or just a change in operational parameters. Most condition monitoring
techniques must get this information elsewhere: from the client's
production schedule, for example, or from a separate system of devices
that can track the relevant data.

With MCSA, that information is built into the box. MCSA systems
directly measure the frequency supplied to the motor—which means
you automatically have the speed for every data point you collect.
MCSA also directly measures the voltage, which gives you the Figure 10. Each color represents the instantaneous spectral energy (top) and
instantaneous load at every data point. active power (bottom) at a different combination of speed and load.

The AI software does the rest, automatically comparing each new
measurement with the right set of data: the asset's history of healthy
behavior for that specific combination of speed and load.
Note how much the data vary—if we didn't know they represented different
situations, we would either flag nonexistent damage (because behavior has
visibly changed) or miss real damage (because our definition of "healthy
behavior" is too broad). Only systems that measure current + voltage can
automatically eliminate these risks.

E-book | Motor current signature analysis explained

 15

MCSA is the only condition

monitoring technique that
can automatically provide
metrics to raise asset
performance and energy
efficiency in real time.

E-book | Motor current signature analysis explained

 16

Real-time performance & energy metrics

The presence of both current and voltage data lets you track a host of
additional metrics* that can help raise efficiency, lower costs and shrink
your company's environmental footprint, such as:

a real-time pump monitor to help you steer a pump back to its

best efficiency point, reducing cavitation and raising bearing and
seal life. Over time, the data can tell you where redesign or (a) sample real-time pump curve and performance metrics
replacement would score major cost and efficiency gains.

an energy monitor to track an asset's operational efficiency. Over

time, the data can tell you where redesign or replacement would
score major cost and efficiency gains. (Read more in our
sustainable industry white paper.)

a power quality monitor to identify and solve supply-side issues

such as voltage unbalance and harmonic distortion.
(c) sample energy efficiency metrics

All these metrics require current and voltage information; they can't be
calculated from vibration, thermal, acoustic or oil-based data.
(b) sample power quality metrics

* Since these metrics are extras that go beyond monitoring for developing failures, Figure 11. Only an MCSA system has the necessary data to calculate extras
not every MCSA vendor will provide them. Be sure to compare features for different such as quality, performance and energy efficiency metrics.
systems.

E-book | Motor current signature analysis explained

 17

Conclusion

So those are the facts on MCSA! We hope this explainer has given
you useful insight into this condition monitoring technique and
how it can strengthen your company's predictive maintenance
strategy.

If you're ready to start comparing apples to apples, we’d be happy

to tell you more about our MCSA solution, SAM4. Just drop us a
line to book a no-obligation demo at your convenience.

Contact Passionate about solving the

problem of unplanned
Lars Ligtenberg downtime, Semiotic Labs
Sales development representative uses AI-driven electrical
lars@semioticlabs.com waveform analysis to create
smart predictive maintenance
+31 653 325 256
solutions that increase
productivity while saving
www.semioticlabs.com
customers' time and money.

E-book | Motor current signature analysis explained