Condition monitoring as a foundation for successful asset reliability and maintenance
This guide goes into detail on everything we consider important to understand about condition monitoring. What is it, how does it work, understanding what makes for effective condition monitoring and what your organization needs to prepare? It also explains the role condition monitoring plays in achieving predictive maintenance and how Senseye’s services can help you reach your predictive maintenance goals.
- What is condition monitoring?
- How does condition monitoring work? (The past, the present and the future)
- Condition monitoring techniques
- What makes for effective condition monitoring? (Goals, metrics, culture, cost and measurement)
- How ready is your organization? (Data acquisition and management, evolving your maintenance culture and additional short-term investment)
- Adopting a predictive approach to condition monitoring (Making CBM more predictive, predictive CBM benefits and scalable, predictive CM for industry)
- Our services
How does condition monitoring work?
Condition monitoring is not a maintenance methodology. Instead, it's a technology enabler for other maintenance methodologies. Every organization's maintenance journey will be different, and to get started, you must understand your organization's maintenance mix.
A maintenance mix includes various methods, including corrective maintenance, preventative maintenance, condition-based maintenance (CBM), and predictive maintenance (PdM). But to understand how condition monitoring works, it's helpful to know where we have come from, where we are now, and what we should consider for future implementations.
Condition monitoring had very humble beginnings. Based on intuition, engineers would listen to a machine to try and understand if it ‘felt right or not’. But pure intuition isn’t a good way to scientifically implement predictive maintenance and avoid unplanned downtime.
And so, CM systems began to advance, especially in Aerospace and Defence industries, where rotating equipment such as engines, gearboxes and motors became the focus of implementing PdM. Ten years ago, a typical condition monitoring project would have followed a traditional engineering roadmap. It would have incurred a lot of expense and effort upfront, only delivering value later and ruling out the rapid return on investment (ROI) demanded by most potential users.
The picture has gradually improved over the past decade, as continuous, automated monitoring systems have become more affordable. But it’s only more recently that techniques and technologies have combined to tip the balance decisively favouring condition monitoring in a broader range of industries. Industry 4.0 - a term coined to describe the fourth industrial revolution, which focuses heavily on interconnectivity, automation, machine learning, and real-time data - has significantly impacted the adoption of automated monitoring systems, enabled by truly affordable and scalable products and technology.
Today, there are two broad categories of condition monitoring types, offline and online:
Offline condition monitoring
With offline methods, handheld equipment tracks and records readings, at scheduled intervals. There are various conditions to monitor, such as vibration, acoustic emissions, ultrasound, and thermography. The readings provide a snapshot of signals at a particular point that maintenance engineers can interpret to detect and diagnose current issues.
The handheld equipment can detect minute variations from the specification. For example, when detecting vibrations, your handheld equipment uses an IEC standard database (e.g. ISO 20816-1) to diagnose the bearing condition based on a single reading.
- Handheld equipment is typically costly
- Requires a high level of training to operate the equipment
- Requires a high level of training to acquire reliable and repeatable readings
- Requires a high level of training to interpret the raw data that these systems produce
- Systems must be highly accurate as readings could be infrequent (e.g. some readings are taken once every six months or as time permits)
Online condition monitoring
With online methods, permanently wired sensing equipment monitors the machinery. The sensing methods used will be based on parameters such as vibration and current.
See more about condition monitoring techniques.
Upwards of thousands of measurements can be taken per second to give extremely high-resolution signals that experts can interpret to detect and diagnose machine behaviour.
Advantages over offline monitoring:
- Real-time automated capture by dedicated equipment
- Does not require a maintenance engineer to place temporary sensors
- Issues with machinery are detected much sooner
- Much more appropriate for machinery designed to be safety or mission-critical
- No delays between readings
- Reduces investigation time
More comprehensive condition monitoring implementation will become widely available in the future. As sensors, connectivity and processing costs tumble, technology-driven capabilities that used to cost millions for Aerospace and Defence companies will be available at a fraction of the price, helping to reduce unplanned downtime and operational costs for far more organizations.
For more information, why not read our white paper: Condition Monitoring: Past, Present and Future.Download white paper
Condition monitoring techniques
Whether offline or online, there are a variety of conditions and parameters to measure when it comes to condition monitoring. Below is an outline, with an easy comparison of time, cost, and complexity.
|Vibration is most often used for rotational components (bearings, shafts, gearboxes) and is well suited to complex assets. Unfortunately, it requires significant post-processing to extract value and needs a high sample rate to be most effective. These factors make it one of the most expensive methods of determining a machine's condition.||
|Measuring temperature can be as simple as bolting a thermocouple onto something or monitoring using thermal imaging. It’s well suited for monitoring solid-state electrical components but inferior for mechanical parts - by the time a temperature delta shows, the damage is likely too profound for any notifications to be helpful.||
|Current monitoring is effective when monitoring electrical motors and for understanding the systems to which they are attached. The motors' behaviour and any mechanical issues (gearbox, bearing) can be identified through current monitoring if taken at a high enough sample rate. The cost is attractively low, but accuracy and specificity lag.||
|At its most basic, it is inexpensive (some microphones and a system to capture data). But truly accurate acoustic emission detection requires similar post-processing to vibration monitoring. The time horizon is excellent, and it performs best in high-frequency applications.||
|A sample of gearbox/motor oil is taken and inspected for debris. Oil debris monitoring can be done online using special sensors but is most commonly manual and labour-intensive. The cost drivers are time (manual labour and potentially downtime) as well as any lab analysis. Unfortunately, once the debris is visible in the oil, significant damage has already occurred.||
|Ultrasound is particularly well suited to inspect pumps / sealed systems for leak detection and monitors things like bearings and valves. It’s versatile but requires significant processing and analytics to produce valuable information for end-user maintainers and operators. The detection speed is rapid, with NASA stating that ultrasonic bearing monitoring can provide the earliest warning of failure.||
What makes for effective condition monitoring?
Condition monitoring can be a minefield. You must have a clear idea of what you want to achieve and its viability. Without this, it will be impossible to monitor the correct assets with suitable sensors. The following tips are essential for anyone who is considering implementing a condition monitoring programme for predictive maintenance.
Understand what it is you are looking to achieve. Have a precise aim to ensure it delivers what you want. For example, are you looking to achieve regulatory compliance, a cost reduction, or improve build quality? Your goals will be unique to your organization and sector. Once you have established a plan, the next step is to verify its validity.
Having a metric to measure success is key to establishing the validity and Return on Investment (ROI) of condition monitoring. Your crucial metric could be one or a combination of: Reduce the cost of production, reduce downtime, reduce wastage, reduce spares inventory, reduce planned maintenance, extend asset lifetime, ensure uptime, ensure quality etc.
How ready is your organization to implement, maintain and optimize condition monitoring? Consider the level of commitment needed from senior management to operations and engineering. Consider the current knowledge level of CM in the organization and the level of investment to upskill. And finally, consider the appetite for change. Your organization will need to be ready for condition monitoring. See ‘How ready is your organization’ for more information.
Ensure the overall cost of implementation and ongoing monitoring is less than the gains (e.g., financial savings, cultural benefits). To do this:
- Perform a cost/benefit analysis.
- Take the overall cost of implementation and ongoing monitoring, e.g. technology, training, upskilling, processes etc.
- Determine the potential savings and gains, e.g. downtime avoided, extending the lifespan of your assets, improved production quality etc.
Decide what asset to measure. We recommend starting simple. Start with known issues and establish easy ways to collect the data. Then build from there. You should measure all assets in time, but beginning with a pilot will make measuring and proving an ROI easier.
For more information, read our white paper 'Effective Condition Monitoring: An Enabler For Predictive Maintenance'.download white paper
How ready is your organization?
Automatic condition monitoring requires a shift in organizational culture to support significant changes in process, attitude, and skillset. These changes can bring challenges, but an organization can enable factory and organization-level scalability by moving from manual to automatic condition monitoring. A half-hearted or poorly planned implementation could see a substantial investment result in disappointing results. The following is a three-point checklist outlining the idea groundwork to implement effective automated condition monitoring:
Data is the foundation of condition monitoring programs, so the correct sensors need to be in place on the right equipment to fuel the analysis engine. The sensors must generate meaningful data that is influenced by the health of the asset being monitored. Acquiring data doesn’t need to be complicated or advanced, but some level of monitoring should exist.
Ask yourself the following:
- What is the monetary value of improving our maintenance?
- How reliable do we need to be?
- Which are our critical assets?
- Are we collecting the correct data in an appropriate format?
- Do we have the right people to help us transition to predictive maintenance?
Data needs to be accessible and secure. Effective mechanisms of separating IT and OT networks need to be employed - a simple IOLink interface could suffice, or solutions like Siemens MindSphere can take all security and connectivity worries away.
Suppose you have historical data in a compatible format – from an IoT middleware platform such as ThingWorx, Predix or MindSphere, or a more traditional factory Historian such as Kepware, Wonderware or OSI PI. Although you could start with something as simple as a CSV file regularly exported – this can fast-forward the path to meaningful diagnostic and prognostics.
Obtain as much information on the monitored equipment as possible, such as nameplate information, equipment location, safety concerns and rotating speed. This information may be available from various sources in your facility, such as the Computerised Maintenance Management System (CMMS), plant library, or various other resources. You can use it to identify specific faults in the monitored equipment and significantly increase the accuracy of your analysis efforts.
Effective automated condition monitoring requires a shift in company culture. You can achieve company-wide adoption with a planned communications programme, outlining the big picture, the project goals, the benefits and provide updates along the way. Select a strong project team from around the business to help with communication, feedback, keep the project moving forward, and carry out any required training in their respective departments.
Security should be high on your priority list. Involving IT at the beginning of the process will make them supporters rather than a roadblock later. Too often, the most significant security threats such as the Meltdown and Spectre calamities are internal. Ensuring staff are mindful of security, updating passwords, ensuring antivirus software is up to date, encrypting data, keeping on top of permissions, and maintaining a firewall is critical in maintaining a secure network.
Continually assess any addition to a network as this carries additional risk. Solution providers should provide security documentation, including third-party audits, and answer any questions posed by your IT team. Your IT department may have valuable suggestions that can contribute to the overall success of the condition monitoring technology or may have concerns to address before implementation.
Any significant business project requires an implementation budget for staff training, sensors, data storage and infrastructure costs. Also, when moving from a preventative to a proactive approach to maintenance, there is inevitably a cross-over ‘bulge’ period where existing scheduled maintenance is taking place alongside new preventive jobs/actions.
As confidence in the predictive approach grows, you can reduce the scheduled maintenance. In the meantime, there is a bulge in the workload. Managing this requires flexible workforce planning and an upfront communications plan to avoid nasty surprises and reassure maintenance teams.
Start with something manageable - try a pilot program. Identify a known machine that requires constant monitoring and set up a small test to capture and analyse the data continuously. The test will help you develop an appropriate workflow and build up the internal culture necessary to take advantage of this change in practice before you scale it up to the entire factory.
Making CBM more predictive
CBM methods are crucial but lack foresight – typical diagnostics focus on current conditions and early detection of degradation based on predefined thresholds without providing a future horizon. CBM output is knowing that something will happen at some point, but not necessarily when.
Prognostics is the science of forecasting when assets will stop being able to perform their intended functions. Unlike CBM alone, adding prognostics capabilities enables an intelligent system to enhance the value of CBM by predicting when failure or degradation is likely to occur, meaning you can adequately perform PdM with the correct information. It is undoubtedly the future of condition monitoring.
Predictive CBM benefits
The overall goal of a PdM solution is to identify and rectify problems at early stages to improve machinery reliability and reduce the cost of ownership whilst improving availability.
PdM provides wide-ranging benefits
PdM generates significant savings
In other industries, the extensive deployment of PdM has generated significant operational and maintenance savings and contributed to increased operational efficiency. Reductions in maintenance costs of between 10% and 50% and increases in asset availability of 50% are not uncommon. Depending on the scale of the asset, one-off savings from a successful failure detection can be huge – several hundred thousand dollars per instance in the case of aerospace or energy.
Scalable, predictive CM for industry
By introducing an easy-to-use diagnostics and prognostics software tool, multiple different sectors can feel the benefits of condition monitoring.
A ‘data-driven prognostics’ approach uses real-world data that is more readily available in an industrial setting. It allows us to build a picture of ‘normal’ machine behaviour within two weeks (if no historical data is available). It can alert when abnormalities occur, in some cases providing fully automated diagnostics.
Maintenance data is then automatically correlated with this information to build prognostic models used by the system to identify RUL (Remaining Use of Life) on specific and similar assets. This happens without needing extensive user input or configuration.
Condition monitoring services
Senseye has extensive knowledge and experience in condition-based maintenance. Our founders honed their expertise in the Aerospace, Defence and Transportation industries, world-leading industries in safety, maintenance practices and condition monitoring technologies.
We believe in making these maintenance practices and technology easily accessible and scalable for the industrial and manufacturing sectors. Developed by experts, our Senseye PdM product and service provides deep support, consultancy, and ongoing input in the maintenance field.
Our team includes industry specialists, condition monitoring experts, mechanical engineers, and leading data scientists who offer expert condition monitoring and PdM services that include:
- General condition monitoring and maintenance consultancy
- Support to business case development
- Sensor selection and installation
- Maintenance record analysis
- Data development and pre-processing
- Integration projects (including SSO)
- Predictive Maintenance training