Intelligent On-Line Vibration Monitoring
An in-depth analysis of one of the most advanced analysis
tools for vibration analysis. By Jesus. R Sifonte, P.E., Level III
Vibration Analyst
|
T |
oday, most maintenance professionals are aware
of the benefits of a comprehensive vibration analysis program. The level of success of such a program
depends on the practices used for analysis, including analyst’s ability of
interpreting vibration spectra to determine machine condition. Accurate vibration analysis requires knowledge
of machinery, vibration theory and a lot of field practice. Vibration analysis is one of the most
important technologies in Predictive Maintenance, but it requires time to
master. The entire process, which ranges
from data collector configuration and the collection of data itself, and goes all
the way to alarms setting and data analysis, is remarkably detailed, and if any
of the steps between is not performed properly, a good diagnostic will never be
attained. Some predictive maintenance
programs experience a certain level of failure in critical equipment largely because
data is not taken frequently enough to detect incipient problems and take
corrective actions on time. Keeping critical
equipment from causing halts in plant production is the aim of Intelligent
On-Line Vibration Monitoring System.
On-Line Vibration Monitoring
The
process of automatic collection of vibration and/or process data from
permanently installed transducers is known as On-Line Vibration
Monitoring. Through this process, data
is aquired from these sensors and sent directly to software, all without the
need of human interaction. A continous
stream of data from critical machines is sent to a PC and trend plots of
overall values are provided for analysis and allows the setting of flag alarms
when a preset vibration level is detected.
The end user can then interpret the data and take corrective measures. Vibration data could be trended as just one
overall value for the selected frequency range of each measurement point or it
can also be trended as a multiple band of overall values, usually six, which
monitor different frequency ranges within a particular spectrum.
Intelligent On-Line Vibration Monitoring
Advanced on-line systems are able
to gather data, analyze its spectral information by comparing current data with
baseline data of a healthy similar machine, and prepare a diagnostic and maintenance
recommendation report with an exceptionally high level of accuracy. This report can then be sent to the end user
right after the data is collected.
Besides collecting data automatically, these systems have an integrated
automatic diagnostic system based on spectral analysis rather than overall
values. This means that even if the
overall vibration does not reach its preset value, the system is still able to
monitor small changes within spectrum, correlate it to machine rotating
components and flag specific incipient problems with amazing accuracy.
How does an Intelligent Automatic System work?
The following flowchart shows the main steps of an automatic diagnostic procedure used by an Advanced On-Line Monitoring System.
The following picture shows a typical report from an Intelligent On-Line Vibration Monitoring System
Modern Automatic diagnostic
systems are able to justify diagnostics and recommendations with vibration data
allowing vibration experts to judge the results (see vibrations details in the
figure above). The report shown is
issued every time data is taken from the machine and sent electronically to
designated users when problems are found or the machine status has changed.
How accurate are the results?
Accuracy
is the most important aspect when diagnosing problems to a very critical
machine in any plant. To reach a high
level of diagnostic accuracy when dealing with vibration data you must do
vibration signature analysis in order to pinpoint problems from specific components
inside the machine.
The US
Navy carried out an experiment with a particular high end system in which
diagnostic accuracy and overall performance was compared to Vibration Analyst
Engineers with 2-4 years of experience.
The results were astounding. The automatic diagnostic system was more
then 40 times faster than the engineers when analyzing data. It was also 8% more accurate than these
engineers as reported by a 15 years of experience analyst engineer chosen as
the referee for the test. The automatic
system proved to be 99% accurate when diagnosing a machine as OK and 96%
accurate when detecting problems as compared to the experienced analyst. The software they tested emulates the
analysis procedure followed by a human vibration analyst with the help of a
rule based procedure to determine machine problem and severity.
Requirements for a Successful Implementation
There are
many successful Intelligent Online Systems implemented which have being
delivering outstanding results to end users.
But the key to make it all work for you is in the implementation. The following is a list of key elements when
implementing these systems:
1. Train
users in the basics of vibration analysis.
2. Choose
appropriate transducers, sensors locations and mounting techniques.
3. Determine
the machine operation mode for data collection and program it into the
Intelligent Software.
4. Determine
data collection frequency according to machine criticality.
5. Enter
the machine characteristics (both dynamic and physical) into the software.
6. Select
baseline data. Make sure there are no
defects on the average data.
7. Review
the expert system report and carry out repair recommendations.
Final Notes
Critical
machines are often subjected to extreme operating conditions due to very
demanding production requirements. Wear
acceleration and failure can occur often, even in machines monitored
periodically in predictive maintenance programs. Such machines are ideal candidates for
Intelligent On-Line Vibration Monitoring.
Protect your product and machinery investment through Intelligent
On-line Monitoring, but be smart about it, only choose high end systems that
deliver answers, not just more data.