Professional Documents
Culture Documents
Industrial Internet of Things-Based Predictive Maintenance For Equipment Protection
Industrial Internet of Things-Based Predictive Maintenance For Equipment Protection
1st Yasser N Aldeoes 2nd Dr. Prasad Gokhale 3rd Dr. Shilpa Y. Sondkar
dept. Computer Engg, dept. Computer Engg, dept. Instrumentation Engg,
Vishwakarma University Pune, Vishwakarma University Pune, VIT College
India India Pune, India
Prasad.gokhale@vupune. shilpa.sondkar@vit.edu
yasser.aldeoes-043@vupune.ac.in
ac.in
Abstract: This article provides a brief overview of each revenue streams, and Industrial Internet of Things (IIoT)
approach that should be included in a predictive maintenance technology can help them do so[3]. Machines on the
program for typical facilities, as well as basic inspection manufacturing floor or appliances in a customer's home can
procedures that can be used to reduce field time. Often,
even evaluate their performance and order their replacement
improper and unnecessary maintenance may result in wasted
resources, time, and money. Industry 4.0 has developed a components when there are indicators that equipment needs
methodology that can be used to lead maintenance from failure to be repaired when predictive maintenance is in place.
to success. In this paper, a safe and fast method of analysis for Predictive maintenance can even employ big data techniques
predictive maintenance is used. This method is called infrared to forecast future equipment breakdowns. Predictive
imaging. It is based on image processing and statistical maintenance is condition-based, meaning it only performs
methods, in which the temperature of the hot spots of the maintenance when the component requires it. Industrial
bearings during work is compared with the reference
temperature. The application of predictive maintenance has maintenance, in particular, contributes significantly to this
become an essential part of all production and industrial fields, competitiveness by ensuring the dependability and
including general machinery and the petroleum sector, and gas availability of production equipment[4]. The maintenance
and power generation, and water plants, petrochemicals. All costs/added-value product ratio is considerably higher than
these industries need to successfully implement predictive 25% in continuous production businesses (for example, the
maintenance programs through a maintenance process that energy, chemical, food, cement, or paper sectors). Process
begins by collecting data from sensors and historical data and
failures, it appears, can halt all production and have a
then displaying the solutions. Data analysis will then be shown
through statistical modeling and analysis techniques. Finally, substantial impact on the manufacturing industry's
data and results are sent and displayed in control panels and competitiveness. As a result, predictive maintenance has
smart devices that provide employees with information to become a crucial issue in recent years.
make sound decisions about the state and maintenance of their
system.
to produce models and forecasts. Reports are greatest public uses of vibration Analysis, infrared images,
generated, work orders are generated, and and ultrasound signals show in Table 1.
maintenance activities are recommended as a Table 1: Common predictive technology applications.
result of this stage. Leadership, operators,
Equipment/Cran
Electric Motors
Applications ↓
Tanks, Piping
Transformers
maintainers, and engineers are examples of users
Condensers
Exchangers
Pumps
Valves
who might use these results to make decisions.
Heavy
Generators
Heat
Technol
es
ogy
Diesel
C. Present Vibration # # # # # # #
Monitoring/A
The final stage is to display the results of data in the nalysis
panel after finding the underlying cause of the problem Infrared # # # # # # # # #
and knowing the state of the equipment. These Thermograph
dashboards are tailored to the user's requirements, y
displaying only the data that is relevant to them. Users Lubricant, # # # # # #
Fuel Analysis
may see the sensor values on a graph on the screen,
Visual # # # # # # # # #
which can be used to predict when a failure will Inspection
happen. They can observe temperature, vibration, rpm,
and other important readings there. These panels permit
users to see real data in real-time, track trends, and keep Vibration can be used to detect mechanical defects in
an eye on system security. rotating equipment. Infrared is used to detect faults in
electrical circuits. Ultrasound is used to find leaks in
compressed air and gas lines [21].
V. PREDICTIVE MAINTENANCE EQUIPMENT & TESTING
TOOLS
Predictive maintenance uses non-destructive testing
A. Vibration Analysis
methods like infrared, vibration analysis, sound level
measures, oil analysis, and other particular online tests to Vibration analysis is conducted to identify the
assess equipment status [20]. Predictive maintenance operational and mechanical condition of the equipment.
software consists of vibration control, infrared imaging, and Vibration analysis provides several benefits, including the
visual examination. To maintain all the equipment, we potential to detect developing issues before they become
need a successful predictive maintenance program. It can
too serious and cause unscheduled downtime. Most
logically reduce costly downtime for machines by scheduling
maintenance only when needed. With the right tools, you can predictive maintenance solutions include vibration
predict when maintenance is needed. For example, the monitoring as a fundamental component. Vibration
monitoring can detect worn or broken gears, deteriorated or wear particles, water content, and other factors. Oil analysis
damaged bearings, and mechanical looseness. Vibration is often used in industrial facilities to check essential
analysis can also reveal misalignment and unbalance before equipment such as compressors and gearboxes. Diesel
they lead to shaft damage[22]. Vibration analyses are engine tests are regular and necessary in the transportation
particularly useful on high-speed rotating machinery, and business. Diesel engines generally work in harsh and
they can be the most costly part of a PdM program to difficult operating conditions, such as train tracks, heavy
implement. The vibration's velocity, displacement, and transport trucks, ships, and others. Oil analysis allows us to
acceleration properties are assessed to define the severity of precisely track the health of these engines in a way that
the vibration, and these are generally referred to as the other predictive technologies may not be able to do in terms
vibration's " amplitude ". of detecting bearing problems caused by metal particles or
chemical changes to the metal caused by high temperature,
and sealing problems caused by pumping fluids into the oil
[26].
1) Velocity (vibration velocity.)
D. Visual Inspections
2) Displacement (the degree to which the
measured component travels with each wave.) A plant's machinery and systems must be visually
inspected regularly as part of any predictive maintenance
3) Frequency (cycle of occurrence for vibrations). program. In many circumstances, a visual inspection will
Vibration amplitude (acceleration, displacement, reveal potential issues that other predictive maintenance
velocity ) is a measure of the severity of a machine's procedures will miss. Many potentially serious problems
problem, as previously stated. The determination of whether can go undiscovered even with the prediction tools
the vibrations are acceptable to allow safe machine outlined. Visual inspection of all essential plant systems
operation is a common problem for vibration analysts. regularly will supplement the other procedures and
These readings reveal the type of vibration that was guarantee that any faults are identified before they cause
identified. The accuracy of the analysis ensures that the major damage[27]. IIoT Predictive Maintenance
correct scale or measurement is chosen. accomplishes the same goal by analyzing massive amounts
of sensor data to see if the underlying gear is
malfunctioning. Root Cause Analysis (RCA) and Time to
B. Infrared thermography Failure are provided based on this analysis.
C. Oil Analysis
Approximately 80% of mechanical failures are due to
contamination, and oil analysis is one of the most simple
techniques to monitor contamination [25]. Another powerful
technology with numerous uses in several industries is oil
analysis. A sample of oil can be examined for viscosity,
Fig. 2. IIoT for PdM
oil, and pharmaceuticals. Predictive maintenance can temperature sensing systems, and ultrasonic sensors, while
address the issues that cause breakdowns and reduce the hydrocarbon sensor cables are used to detect hydrocarbon
amount of unplanned and needless maintenance. leaks. [29].
Maintenance of general machinery, and oil & gas
components is expensive and growing, necessitating more VII. DISCUSSION AND RESULTS
efficient maintenance techniques. Most predictive
maintenance may be performed while equipment is running
to minimize disturbance. Traditionally, the equipment has
been serviced on a time-based schedule, which has resulted In this work, The Vishwakarma Institute of Technology
in wasteful maintenance reparations and a raise in Lab has implemented the bearings for typical machinery to
maintenance expenses. Because these sectors play such an apply Predictive Maintenance and monitoring for other
essential role in our lives, it is critical to developing their machines similar with just the component being monitored
maintenance procedures. Figure.2 shows three steps explain changes. The bearing is fitted with sensors that collect
using IoT for PdM The first is “data collection”, “model & temperature and vibration data. almost in predictive
optimize” and then “present”. Show in the figure.3 maintenance centers, special programs are employed to
process sensor data and store it in a database. The
1. Typical Machines MATLAB program was used to examine the signals and
images obtained in this study, and the data were processed
Pumps, Motors, fans, gearboxes, compressors, and analyzed. The data collected by the sensors is stored in
conveyors, turbines, rollers, steel and paper industries, excel (.csv) spreadsheets for use in developing models and
engines, and machine tools with rotational parts are applying machine learning algorithms. Figure.4 shows three
examples of typical machines. The majority of these parts to explain what the users can see during this work. The
industries are concerned with gear, bearing, shaft, and left of the figure presents the bearing signals that are normal
chassis maintenance. Previously, these sectors used a time- and abnormal for sensor readings. in the middle of the
based preventive maintenance plan. As a result, figure, we see During bearing under normal operating
maintenance is one of the most significant cost drivers in conditions nothing unusual is noted the photo is green and
many businesses, accounting for up to 50% of overall there are no actions that need to be taken, and the bearing is
expenses. [28]. unhealthy show the device photo is the red color we should
do the measures to prevent failure.
2. Oil and Gas Industry
In Figure.3, continuous monitoring of the thermal
behavior of the bearings under evaluation is performed by
Predictive maintenance solutions powered by IoT assist
capturing vibration signals and infrared images associated
gas and oil firms to foresee equipment faults before they
with temperature values. When there is a thermal error we
have a substantial impact on their safety and bottom lines.
will see that the temperature readings start to rise compared
The IIoT works to ensure that pipelines, crude oil
to the time the bearing was operating normally. Through the
processing systems, and gas processing equipment are safe
new data, fault models will now detect a thermal fault-based
and reliable. Sound variations that indicate fluid leakage
it on which the necessary measures are taken directly.
(such as crude oil) are detected using fiber-optic sensors,