What is condition monitoring? (The 2024 guide)

Condition monitoring is an important tool in the predictive maintenance of machines. By collecting and analyzing certain signals from motors, developing faults and inefficiencies can be identified, and unplanned downtime can be avoided.

There are a number of different signals that can be taken into account when monitoring mechanical assets. Traditional techniques were mostly based around vibration analysis, but more modern, innovative techniques focus on ESA (electrical signature analysis).

Condition monitoring can differ both in the way that it is performed (i.e. handheld vs online), as well as the type of motor data that is monitored.

Handheld vs online

Although the practical differences between handheld and online techniques might seem obvious, the way in which each method is used can differ significantly.

Simply put, handheld condition monitoring involves the use of a handheld sensor, which is applied to a machine when necessary to determine the condition of the machine, and ultimately determine when maintenance must be scheduled.

Online condition monitoring requires sensors to be installed permanently (either on the asset itself or near the asset) and provides ongoing real-time insights into the health of the motor.

The handheld use case

An important initial point to make is that the handheld approach typically offers a less complete picture of asset health, and so is often applied to assets of medium criticality. The reason that it offers a less complete picture is that by definition, readings are taken at intervals instead of continuously. And fewer readings translate into a less precise picture of asset health. If the assets that need to be inspected are far apart, in an ATEX zone, in a hazardous location or in a difficult-to-reach place, then this can result in even fewer readings being taken. Whether this incomplete picture provides appropriate protection against unplanned downtime is determined largely by the criticality of an asset (and the downtime costs associated with that asset), and as such an incomplete picture is typically applied to assets with a medium criticality level.

As well as monitoring assets of medium criticality, handheld condition monitoring is also a popular monitoring method for assets that do not run regularly. If an asset only runs half the year (for example, the motor powering a ski lift), then readings may only need to be taken for half of the year. So it may not make financial sense to install permanent online condition monitoring on assets that only run periodically (however, this obviously depends on the criticality of the asset when it is running).

The online use case

In comparison with the handheld approach, online condition monitoring provides a much fuller picture of asset health. As the online variant involves the use of permanently installed sensors, it is used to automatically take regular measurements from the asset without needing to dispatch an engineer to inspect the asset. This in turn means developing faults can be quickly detected and resolved before unplanned downtime occurs. It follows that online condition monitoring is often deployed on high-criticality assets, where preventing unplanned downtime is a priority.

Data source

Condition monitoring can also vary when it comes to the type of data that is collected. Below is a brief introduction to some of the more popular methods.

Vibration analysis

One of the oldest and most trusted forms of condition monitoring, vibration analysis monitors the vibrations emitted from a piece of machinery, and can detect developing faults if the vibrations measured differ from that of a healthy machine. Vibration measurements can be collected through either a handheld device in close proximity to the asset or through sensors installed directly on the asset.

Electrical signature analysis (ESA)

ESA-based systems measure current and voltage signals from within the motor control cabinet. By deploying sensors inside the MCC (which is a dry and hospitable place for sensor deployment), ESA is well suited to monitoring assets located in dangerous environments (such as hot roller table in a steel mill) or difficult to reach places (such as a sewage pump located deep underground). And because ESA systems measure current and voltage data, it is able to offer energy usage insights which can be used to improve the efficiency of the asset itself.

Oil analysis

The composition of the lubricant is analysed to determine the presence of a developing fault. For example, increased levels of iron, copper or lead in the oil can be a sign of wear. Oil was historically analyzed through samples taken periodically, but more modern systems allow for continuous monitoring through on-asset sensors.

Thermal analysis

Temperature sensors, such as thermocouples or infrared cameras, are used to analyse the heat coming from a machine, which can be used to detect the presence of a developing fault.

Acoustic analysis

Acoustic emission analysis involves the measurement of transient elastic waves emanating from an asset. This enables maintenance teams to detect the presence and precise location of faults such as cracks, leaks and fiber breakage in a piece of equipment.

The condition monitoring comparison guide

To learn more about the different types of condition monitoring, including areas in which each method is strongest (and weakest), download the comparison guide below.

What are the benefits of condition monitoring?

Simply put, condition monitoring uses a number of signals to predict three things. First, whether an asset will break. Second, how it will break, and third, the time you have to fix or replace the asset before it functionally fails. Armed with this information, you can schedule maintenance at a time that suits production.

Avoid unplanned downtime

The ability to plan downtime in an industrial environment is hugely beneficial, as the true cost of unplanned downtime due to a failed asset is often wildly underestimated. There are a number of cost factors which are routinely ignored, such as:

• the true cost of an unplanned delay in production
• the need to pay overtime to maintenance staff to replace the asset
• depending on the severity and type of machine break, other machines may be damaged as a result of the asset fault
• the cost of needing to store large numbers of spare assets in case any one of your assets breaks. Condition monitoring means you will be forewarned of any asset break (sometimes up to 5 months in advance), meaning safety stock for faulty assets can be bought when needed

Apart from the avoidance of downtime due to machine breakage, condition monitoring contributes to a well-run plant in a number of other ways:

Maximize ROI

Predictive maintenance using condition monitoring allows you to maximize the return on investment in your mechanical assets. By monitoring the actual condition of your machine, you can inspect, fix or replace the machine only when it’s necessary, and not before.

Conversely, preventive maintenance requires the replacement of all machines after a certain period of time (or number of running hours) regardless of whether they have started to show signs of a fault. By keeping your machines in action until it’s necessary to change or replace them, you can get more out of your machine, improving TCO (total cost of ownership), and maximize initial capital ROI.

Maintenance engineers can act more efficiently

In a scenario where there has been a breakage, maintenance engineers are able to act faster using condition monitoring. Different motor signal patterns are indicative of different developing faults in the asset the motor is driving—so condition monitoring will help the maintenance engineer to focus on the right fault, and not waste time checking parts of the asset that are not broken. This ultimately makes the maintenance engineer faster and more effective at his or her job.

Safeguard employee safety

By being able to determine when an asset will break, maintenance personnel can ensure safer work practices. Depending on the nature of the asset, a breakdown could be quite destructive, and could pose a threat to the safety of employees working around the asset. By using condition monitoring, maintenance personnel can plan maintenance before a motor break poses a potential threat to safety.

Improve motor efficiency with ESA

SAM4 by Samotics uses electrical signature analysis, meaning it can also detect when a motor is beginning to run less efficiently. As a result, you can focus your efficiency improvements on specific motors.

Future-proofing your plant

Statistically, 20-40% of your maintenance personnel are likely to retire in the next 5 years. That means your ability to react to future unplanned downtime could suffer. SAM4 helps your maintenance team avoid unplanned downtime and maximize plant productivity in the future.

Condition monitoring is an important part of any industrial maintenance strategy, and has a wide range of uses in a range of different environments, including:

• Oil and gas
• Transport
• Food and beverage
• Healthcare
• Local communities
• Water and wastewater
• Airports
• Pulp and paper
• Chemical production
• Power generation

The specific assets that condition monitoring is used for include:

• AC induction motors
• Pumps
• Compressors
• Conveyors
• Blowers and fans

View the video below for 3 examples of how condition monitoring is used in the steel industry.

Selection guide: which solution is right for your plant?

The best way to learn about a specific solution is to book a demo with the supplier. Below are a list of questions and topics to discuss with the supplier during the demo to help you identify whether the condition monitoring software in question is right for you.

Part A: How does this system work?

Systems can differ in a number of ways depending on the use case. Ask the supplier the following questions to get a better idea of whether this solution is the right one for you.

Which type of data does this system collect?

Condition monitoring involves the analysis of asset metrics. But the type of metric measured can vary from supplier to supplier.

Before the demo takes place, consider making a list of the types of asset metric data your plant could generate. This will give you an idea of whether this system will work for your plant. Data types often used include: current, voltage, vibration patterns, motor acceleration and thermal data.

How will this system collect that data?

Different systems collect data in different ways. For example, ESA allows sensors to be installed in the motor control cabinet, whereas handheld thermal sensors require the maintenance professional to physically inspect the machine (which might not be an option if the asset is located in an ATEX zone).

Think about how your plant could collect data, and discuss this with the supplier.

Is this an online or offline solution?

Not all condition monitoring solutions are online. Some systems are purely on-premise, where the maintenance information does not leave the premises.

Although on-premise solutions may sometimes be the only option (think about ships, where network connectivity is limited), online systems are beneficial in most other ways. For example, if you have a company with plants in multiple locations, a central maintenance crew can monitor the health of any motor from any location. Discuss with the sales consultant whether offline or online would suit your use case best.

Part B: How effective is this system?

Once you have established whether this condition monitoring system will work for your plant, it’s time to find out how effective the system is. By asking the sales consultant the following questions, you will gain an understanding of how effective the system is.

What is the failure detection rate?

If the solution routinely misses failures, the ROI for your project will suffer. Anything above 90% detection is considered to be high.

How does the system identify a fault?

How the system will actually identify a fault is an important question. Traditional condition monitoring has required manual data analysis to determine a developing fault. But more modern systems (such as SAM4) use AI and data science to automate analysis and automatically determine if there is a developing fault.

How much detail can the system give on the type of the developing fault?

Advanced systems can not only identify a developing fault, but can also identify the specific type of fault and the severity of the fault. Ask the sales consultant if this system also has this functionality.

If a fault is detected, how will the system alert the maintenance team?

This could be completely manual, completely automated, or a mix of both. Often a mix of both can be beneficial, as a set of human eyes can double-check that the fault is really a fault before the maintenance team is alerted and maintenance is scheduled.

Part C: Installation

Installation is your first real interaction with the tool. A painful or drawn-out installation period can kill your team’s enthusiasm for a new tool or new way of working, while a quick and simple installation period can help build support for your tool.

How easy is installation?

This might seem like a subjective question, but different condition monitoring systems can differ significantly when it comes to ease of installation.

Is installation support offered?

Depending on the complexity of the installation, you might require support. Support can come in the form of onsite consulting, over-the-phone support or online support materials and documentation.

Samotics also offers installation and technical support if needed. However, we find most of our clients are able to install SAM4 without any problems.

AI training time

More modern condition monitoring systems will contain an artificial intelligence element, which will usually take some of the data analysis burden off the maintenance engineer, allowing the maintenance engineer to focus on conclusions and actions.

However, the AI system will often first require a learning period so that it can learn how your motor behaves, and the different operating points at which the motor typically runs. Once the system has captured this information, it can determine changes in future performance metrics which are indicative of a developing fault.

The aforementioned AI training time will vary depending on the system; however, it’s important to ask for an indication of how long this learning period will usually be. SAM4 typically requires only 2–6 weeks before the AI system has learned what it needs to learn in order to effectively monitor your system.

Part D: Continued ease of use

Does this system have an easy-to-use interface?

The only way to really get a feel for this is to ask about it during the demo. You might ask if you can explore the interface by yourself for a few minutes. Without asking the sales rep for help, see if you can find:

• Current motor health
• Fault history
• Is there a way to compare assets (possibly useful when comparing motor energy efficiency)?
• Is there an in-dashboard tour? (For non-tech-savvy users, an in-dashboard tour can be very useful when demonstrating basic functionality.)
• Does the user interface integrate with your existing CMMS system?

SAM4 has an intuitive dashboard which helps you to visualize performance data in a helpful way, and take action faster. By making our dashboard as intuitive as possible, very little training is needed to use SAM4, meaning you can start monitoring your assets as soon as possible.

Ease of ongoing system maintenance

Depending on the makeup of your system, ongoing maintenance can become expensive.

As mentioned above, if your sensors are installed in difficult-to-reach places, then ongoing sensor maintenance can be more costly and can take longer. Additionally, if your sensors are installed on assets which themselves are situated in hazardous environments, they are likely to be damaged and break more often, which in turn increases the costs of ongoing maintenance.

Condition monitoring is a valuable tool for any industrial organization. To learn more about how it could benefit your organization specifically, sign up for a SAM4 demo today.