For people associated with manufacturing business, the word ‘maintenance’ seems to create a familiar image in mind. It is certainly an image of a mechanic dragging a toolbox across the floor of production on wheels to fix a machine that has stopped working suddenly. Even after the concept of predictive maintenance developed, this stock image remained for many years.
Even with the introduction of predictive maintenance, opening the machine doors, guards and covers continued to meet with surprises as unexpected wear, the addition of downtime, and a rushed search for the suitable parts to fix or repair.
It is true that preventive maintenance’s concept improved downtime and minimized the long term costs back in its day. Still, the possibility of meeting with surprises remained unchanged along with the cost associated with downtime that accompanied them. This happened because it was dependent on the ability of the skills, experience, and intuitive understanding of the mechanic, combined with the hopefulness of the maintenance regime’s preventive element, to keep the machine performance under the optimum range. However, the lack of capability needed for assessment of a machine’s operability in real-time still left room for costly failures (surprise-of-the-week).
Predictive maintenance (PdM) monitors the condition and performance of equipment while it is operating normally to minimize the probability of failures. It is also called condition-based management, and it is being utilized in the industrial world since the 1990s.
Predictive maintenance existed even before the 90s. However, its history has not been documented formally.
The main goal of predictive maintenance technology is to predict when the failure can occur (based on certain factors), and its ability to prevent the failure from occurring via regularly scheduled corrective maintenance. Without condition monitoring, predictive maintenance cannot exist. Condition monitoring is defined as the uninterrupted monitoring of machines, to ensure the best use of machines, during the conditions of the process.
There are three facets of condition monitoring: remote, periodic, and online.
Online condition monitoring is the uninterrupted monitoring of production processes or machines, with data collection on various speeds and changing positions of the spindle.
Periodic condition monitoring is gained via vibration analysis. It provides insights into changing installations’ vibration behavior along with trend analysis.
Remote condition monitoring allows monitoring of the equipment from a remote location and uses transmitted data for analysis.
Predictive maintenance (PdM), for real-time assessment of the assets’ performance, depends upon the predictive maintenance equipment. By combining predictive formulas with condition-based diagnostics, along with some help from the Internet of Things (IoT), predictive maintenance develops a suitable tool for the asset’s data collection and analysis. Through this data, the areas that need attention can be easily identified.
Let’s create an in-depth understanding of some major elements involved in the process to understand the working of predictive maintenance.
Predictive maintenance involves the monitoring of each asset by using condition monitoring equipment. To be more specific, sensors are fitted in machines that help to collect data about the equipment to enable the asset’s efficiency evaluation and identify wear in real-time.
This is an essential step as though the equipment’s physical inspection has traditionally been the major practice by which maintenance personnel attempt to observe assets, there is a huge probability of failures occurring as most tear and wear happens on the inside of the machine so, to inspect them fully, you are required to take them apart.
However, by making use of predictive maintenance and condition monitoring sensors, you can achieve an accurate representation of an asset’s overall condition without any sort of interruption in productivity.
Various kids of parameters are measured by these sensors, depending on the machine’s type. Usually, parameters like noise, vibration, temperature, oil levels, and pressure are measured, but you can do more than that and measure things like corrosion and electrical currents.
Gathering data is one thing, but being able to analyze it and use it for the purpose that it is intended for, is entirely another. With IoT technology, the sensors mentioned above collect and share data. Predictive maintenance heavily depends on these sensors for connecting the central system, which stores all information, to assets.
The assets can analyze data, work together, communicate, and recommend necessary actions based on the setting of the system.
This information exchange occurs at the core of predictive maintenance and enables maintenance tech to form a sense of the overall situation and identify assets that require attention.
Here predictive maintenance travels beyond condition-based maintenance. All the previously collected data is analyzed through predictive algorithms that help in identifying trends with a motive of detecting any call for repair, replacement, or servicing.
These algorithms work according to a predetermined set of rules that helps in comparing the current behavior of the asset against the asset’s expected behavior. Deviations work as a signal of gradual deterioration that can later become a reason of asset failure.
Let’s learn the difference between preventive and predictive maintenance.
While various maintenance programs use them both, there are several differences between predictive maintenance and preventive maintenance.
Preventive maintenance involves inspection and maintenance of machinery, no matter if the equipment needed maintenance or not. This type of maintenance schedule is usually based on either time trigger or usage. For instance, a heater gets serviced every year before the winter season, or a car calls for scheduled maintenance after every 5,000 miles.
Preventive maintenance does not require a component of condition monitoring that is required by predictive maintenance. This is the main reason why the preventive maintenance program does not include as much capital investment in training and technology. Various preventive maintenance programs require manual gathering and analysis of the data.
Preventive maintenance is determined by using the asset’s average life cycle. Whereas, the identification of predictive maintenance is based on preset and condition of equipment’s specific pieces, by utilizing various technologies. Predictive maintenance also needs more capital investment in training, people, and equipment as compared to preventive maintenance, but in the long run, cost and time savings will be huge.
A lot of confusion has existed for a long time over the correct way of running inspection to identify the existence of a given failure mode. It included several doubts and concerns such as; should a type of sensory inspection be performed? Should a quantitative inspection be performed? Should more than one condition monitoring technologies be applied? Etc.
Predictive maintenance techniques involve:
Sensory inspection has always been considered as the base of any good maintenance and inspection program. An inspector would be sent to perform sensory inspection and use touch, sound, and sight to determine whether there has been any change since the last inspection. It would help it identifying repair and fixing issues in time.
While sending someone to perform a sensory inspection can prove to be very beneficial. However, there are many loopholes in this process, and it must not be perceived as an inspection program’s backbone. As sensory inspection only helps to identify problems that are obvious and cannot identify internal issues of equipment.
Enhanced sensory inspection fills the holes left by sensory inspection. Both of these inspections are sensory and a quantitative measurement combined with characteristics of condition monitoring. Such inspection use tolls like strobe lights, radiometers, simple ultrasonic meters, and handheld vibration pens that help to detect potential defects. These tools certainly enhance the power of the human senses. However, they, too, have a limit. All of these tools play an important role in identifying failures, but they cannot be compared to a comprehensive condition monitoring process.
Quantitative inspection can have their fair share when collecting data regarding a failure mode. Such inspection requires someone to measure something. Some common quantitative inspection involves temperature measuring of a pump’s seal or measuring a pump impeller’s backplate. These measurements help to provide data to the engineer and the planner and allow them to determine whether there is a need for any maintenance. If you design a quantitative inspection properly, you will be able to receive minimum, maximum, and typical values along with conditional tasks explained for when these limits exceed. However, a quantitative inspection that is performed at an accurate frequency of inspection will automatically show measurements that are exceeding the limits.
Predictive maintenance (PdM) is also known as condition monitoring. It is the application of technologies that are condition-based, equipment performance, or statistical process control for detecting defects early and eliminating them to avoid additional downtime or costs.
This monitoring must be conduction while the equipment is operating normally and with almost zero interruption in its process. The main purpose of these tools (infrared thermography, vibration analysis, motor circuit analysis, and so on) is to identify certain defects that were not found in earlier inspections methods, while the equipment is normal operation.
Making use of the technology that is available for your ease, allows you to run a complete assessment on all parts of the equipment and identify the failures that have been impossible to detect previously.
When you are using predictive maintenance is a maintenance strategy, the maintenance can only be performed on machines when required, i.e., before any failure occurs. This results in various cost savings:
Predictive maintenance programs have proved to be the reason for the tenfold increase in ROI, a twenty-five to thirty percent reduction in the cost of maintenance, a seventy to seventy-five percent decrease in breakdowns, and a thirty-five to forty-five percent reduction in downtime.
This cost-saving surely comes at a price. However, various condition monitoring techniques are very expensive and need experienced personnel for performing effective data analysis.
Here are the types of predictive maintenance:
This type of analysis can be referred to as an on-the-go predictive maintenance analysis. It is performed on manufacturing plants that have high rotating machinery. A vibrational analysis is a cost-effective way of detecting defects. Moreover, vibrational analysis can allow you to identify any imbalance, bearing wear, or misalignment.
It calls for less money for the implementation, and it is commonly used for high and low rotating machinery. Among lubrication technicians, such type of analysis is specifically popular.
Sonic acoustical analysis stands on the borderline of predictive and proactive maintenance. Whereas, ultrasonic acoustical analysis is only used for predictive maintenance. It helps to identify sounds relating to stress in the ultrasonic range and mechanical friction. It is used for mechanical equipment as well as for electrical machinery that releases subtler sounds.
The infrared analysis does not depend upon an equipment’s loudness or rotational speed. Therefore it has proved to be suitable for various types of equipment and assets. When the temperature indicates potential issues, infrared analysis plays the role of the most cost-effective predictive maintenance tool. It is commonly used to detect problems that are related to airflow, motor stress, or cooling.
It is known as a nonintrusive or nondestructive testing technology. Infrared thermography is widely used in predictive maintenance. Personals are enabled to learn about high temperatures by using IR cameras. Components that are worn, such as malfunctioned electrical circuits, will release heat that, on the thermal image, will appear as a hotspot.
Through the identification of hotspots, these inspections can highlight problems and allow you to avoid downtime and costly repairs.
Through acoustic technologies, it can be possible to detect liquid, gas, or vacuum leaks in equipment on an ultrasonic or sonic level. Sonic technology is considered as a cost-effective technology as compared to ultrasonic technology, but its use is limited. Ultrasonic technology, however, offers more applications, and it is reliable for the detection of mechanical issues.
It enables a technician to listen to stress or friction in rotating machinery and help in detecting deterioration earlier than any other technique.
It is primarily employed for rotating equipment of high-speed. Vibration analysis helps a technician in monitoring vibrations of a machine via real-time sensors that are built into the machinery or through the handheld analyzer. A machine that is operating at its best emits vibrations in a particular pattern. When shafts, bearings, or other components begin to wear or fail, the vibration pattern gets disturbed, and this will indicate that the equipment requires maintenance and fixing.
For performing successful vibration analysis, technicians must be well-trained and skilled in understanding and identify the problems occurring.
It can be an effective tool for predictive maintenance. It allows a technician to run checks on the condition of the oil and determine whether or not other contaminants and particles are present. Few tests of oil analysis can also tell about the oil’s viscosity, base number and acid number, particle counts, wear metals, or presence of water.
Among many benefits of using oil analysis, one most important one is that the early test can put forward a baseline for new equipment.
Some other technologies include motor condition analysis, eddy current analysis, borescope inspections, data integration, and condition monitoring, etc. While many different technologies can help you in performing predictive maintenance, it is important to pick the suitable ones to find out potential issues in real-time successfully.
Predictive maintenance assists you in entirely improving your reliability and maintenance program. By using best practices along with technology to streamline processes and boost productivity. Some major benefits offered by predictive maintenance involve:
There are a variety of uses cases for predictive maintenance while using an industrial model of the Internet of Things. There are many risks and issues involved with the use of this technology.
In a nutshell, making use of cloud-based predictive maintenance has proved to be an effective way of minimizing overall costs. As most developing technologies involve a transition period in which a dual maintenance system might be implemented. As the IoT model establishes, a maintenance handover gets easier to achieve with improved results for outcomes and the bottom line of the company.
Begin with identifying essential equipment and systems that must be included in the program. Assets having high replacement/repair costs that are important for production are usually the best candidates for a predictive maintenance program.
For a predictive maintenance program to function successfully, one major factor to take into account is the existence of enough information that can provide actionable and essential insights into the behavior of the machine. Historical data for every pilot equipment will be provided from various sources like hard copy files, CMMS, enterprise software from different departments, charts and maintenance records, and so on.
Here, the predictive maintenance company will require performing an analysis of the critical assets that were previously identified for the establishment of failure modes.
With failure modes and critical assets identified, the net level includes designing the suitable approach of modeling that will be responsible for forming the basis for predictions of failure.
The outcome of this stage is a delivery of a fully automated system which:
Monitors conditions of operations through installed sensors.
Predicts and understand patterns that are created by data anomalies.
Creates alerts when the established thresholds show a deviation.
In this step, predictive maintenance is tested and validated by deploying the technology to a certain pilot equipment group.
Let’s have a look at predictive maintenance tools:
They collect data in real-time.
For industrial sensors, the market is constantly growing in quality and variety, with a large number of vendors who are offering sophisticated products at low costs.
Major parameters covered by industrial sectors involve:
Gateway devices serve as intermediate connection points that maintain connections between sensors and controllers to a computing platform, whether on-cloud or on-premise. Gateway devices protect the data that is being transported and the Internet of Things, providing an extra layer of security.
Even a simple line of production can generate large amounts of data that need to be collected, normalized, aggregated, and analyzed.
For securing the network, maintaining a low level of data communication, edge nodes are usually deployed close to the line of production. The edge nodes are responsible for handling part of processing and analytical workload. This makes network scaling much easier.
A centrifugal pump motor involved in a coal preparation plant is an important asset for daily operations. To reduce unscheduled downtime, the maintenance team decides upon using predictive maintenance technology. As it is a large component of mechanical equipment that helps in performing heavy rotations, the prominent choice is to perform monitoring on vibrations and vibration meters.
The vibration meter is attached near to the inner bearing of the pump and established a simple baseline measurement, that is visualized via waveform graph. After a few months, a spike in acceleration is identified by the vibration meter. The maintenance team reviews the newly collected data remotely, and an inspection is scheduled. The technician who is responsible for performing the inspection learns about a loose or damaged ball-bearing and begins to repair it.
The team then connects CMMS to the vibration meter. When a similar spike is identified, an issue is predicted with the ball bearing, and a work order is triggered automatically for performing repair.
The internet of things is known to have a huge effect on the manufacturing sector. It leads to increased automation, more improved operations, and creations of valuable business models. The application of digital technologies certainly offers various benefits across the value chain. However, it is still arguable when it comes to predictive maintenance that it can drive the most significant impact.
By using data analysis and sensors, companies can identify patterns in conditions of equipment and their performance and correctly predict the occurrence of a failure. This is how unplanned downtimes can be eliminated along with the delivery of substantial benefits of productivity.
Predictive maintenance software is a technique that provides great help to any business in learning about certain crucial impending maintenance and carrying it out as soon as possible, i.e., before the equipment breaks down or wears.