1.
What is Condition Monitoring?
1.1.
Common Types of Monitored Assets
2.
Why Invest in an Asset Condition Monitoring Solution
2.1.
Benefits of Condition Monitoring
3.
Main Condition Monitoring Methods
3.1.
Condition Monitoring Methods
4.
Key Technologies Powering Remote Condition Monitoring
4.1.
Industrial IoT (IIoT)
4.2.
Inspection Drones
4.3.
Machine Learning and Deep Learning
4.4.
Digital Twins
5.
Conclusion
After another operating incident, asset managers ask: What went wrong and why? It’s tough to answer because 90% of asset and equipment failures are event-driven, not time-based. The challenge is not knowing which event could be the next ‘tipping point’ for irreversible damage unless you have a mature condition monitoring process in place.
Learn how emerging technologies help asset owners gain hindsight and foresight into asset performance, reduce operation and maintenance (O&M) costs, and improve operating efficiencies.
Condition monitoring (CM) is the process of observing asset and machine conditions over time to detect early performance anomalies. Parameters like vibration, temperature, wall thickness, or acoustic emissions can be the harbingers of major troubles — cracking, breakage, corrosion, and more.
The simplest condition monitoring system has four elements:
Condition monitoring sensors can provide diagnostics on vibration, temperature, pressure, sound, moisture, and other parameters. You can also integrate data from sources like SCADA, asset management platforms, and industrial drones for an end-to-end view of the asset. The aggregated data can be then analyzed using myriad techniques and software tools.
| Industry | Monitored Assets |
| Oil & Gas | Pumps, compressors, turbines, pipelines, offshore drilling rigs, storage tanks, wellheads, separators, heat exchangers, gas processing units, valves, and rotating equipment. |
| Energy generation | Wind turbines, transformers, generators, power transmission lines, solar panels, inverters, substations, gas turbines, nuclear reactors, boilers, and hydroelectric plants. |
| Infrastructure | Bridges, tunnels, railways, roads, water treatment plants, electrical grids, HVAC systems, cranes, elevators, escalators, pipelines (water/gas), structural health monitoring. |
| Mining | Conveyor belts, crushers, haul trucks, grinding mills, underground equipment, ventilation systems, draglines, shovels, drills, hoists, pumps, and power distribution systems. |
Collecting pre-failure data helps understand the asset’s wear and servicing needs. Without it, you can’t recreate the events leading to failure, exposing your operations to more risks.
Historic conditioning data aids root cause analysis and improves future foresight with modeling techniques. In other words, condition monitoring helps switch from reactive to planned corrective maintenance and preventive maintenance strategies.
There are plenty of different asset monitoring methods, each best suited for specific types of equipment and data capture. Some are manual, others are compatible with automated data collection.
For example, 76% of asset owners use automatic dam monitoring methods, and only 24% — manual. Among the top 50 global mining companies, 94% have robust data infrastructure and 90% collect condition-monitoring data automatically.
Increased automation allows real-time measurement of more parameters with enhanced accuracy. You can track your system’s state with fewer field trips and schedule maintenance as needed to optimize downtimes.
Here’s a summary of the most-used condition monitoring methods for different types of industrial assets.
| Vibration analysis | Shock pulse analysis Broadband vibration analysis Time waveform analysis requency analysis (FFT) |
| Electrical monitoring | Alternating current field measurement (ACFM) Power signature analysis Surge testing Megohmmeter testing |
| Thermography | Infrared imaging Comparative thermography Absolute thermography Lock-in thermography |
| Ultrasonic testing | EMAT testing Phased array testing Ultrasonic thickness and gauging Ultrasonic backscatter technique |
| Acoustic emission monitoring | Acoustic emission testing Acoustic ranging Amplitude monitoring Backwall echo attenuation |
| Process variable and performance trending | Flow rates Pressure Temperature Visual inspection |
| Electromagnetic measurements | Magnetic particle inspection Magnetic flux leakage Metal magnetic memory method Pulsed eddy currents |
| Oil analysis | Wear particle detection Contamination testing Viscosity measurement Dielectric strength testing |
| Motor circuit analysis | Air gap Power quality Rotor Stator |
Effective asset condition monitoring management hinges on data availability and analytics — and new technologies enable access to deeper insights.
Thanks to improvements in wireless core sensor transducer technologies (and substantial cost reduction), asset owners can collect real-time data about gyroscopic motion, pressure changes, vibration patterns, temperature, and humidity. Industrial drones, in turn, speed up photogrammetry, visual testing, and non-destructive testing at heights. Cover greater distances and inspect hard-to-reach assets faster using specialized drone payloads for thermography, ultrasonic testing, dry film thickness measurements, gas or radiation detection.
You can use data engineering to create custom health indicators for assets for effective diagnostics and predictive monitoring. Machine learning algorithms excel in anomaly detection. Data-driven diagnostics can also reveal key factors affecting asset conditions and help implement proactive maintenance. By applying data modeling methods using historical and real-time data, you can forecast asset maintenance needs and model the impacts of degradation mechanisms.
Overall, condition monitoring capabilities widen significantly with the deployment of the next four technologies.
IIoT refers to the growing medley of sensors, edge devices, and automated industrial systems with wireless connectivity and real-time data exchanges. They offer a path to connect legacy machinery and non-digital assets with modern technology for data analysis.
IoT gateways enable secure data transmission, processing, and filtering from all connected assets to cloud-based and on-premises asset condition monitoring solutions.
By consolidating data from multiple sources — condition monitoring sensors, asset management systems, and maintenance applications — you can gain end-to-end visibility into your asset portfolio and apply deeper analysis to optimize its performance.
Eldorado Brasil, one of the largest single-line kraft pulp operations, chose Barker Huegs condition monitoring system for its facilities. The system identified early signs of bearing failure on a bottom digester cooling booster pump motor. The team tracked the failure progression and mitigated its effects until scheduled corrective maintenance during planned downtime. Eldorado Brasil avoided emergency repairs and the loss of about 2,740 tons of pulp production, saving approximately $1.8 million during just one incident.
Beyond advanced monitoring capabilities, IIoT also brings in edge computing capabilities.
Machine data is streamed to the cloud for analysis, but control decisions can be made locally (e.g., adjusting temperature or rotor speed). System controllers (like PLCs) and condition monitoring controllers can share real-time sensing. This provides a functional interface for remotely tuning machine performance and reliability.
For instance, you can combine data from a drone-based wind turbine inspection with sensor-based insights to remotely diagnose rotor imbalance or yaw system issues. Then make adjustments remotely to mitigate risks before dispatching a maintenance crew. Greenko Group, for example, has digitized condition monitoring of over 2,200 wind turbines with AWS services. The system streams telemetry data to a monitoring dashboard every minute. In the background, the system automatically triages low-performing wind turbines, based on parameters like power-performance curve; pitch-angle versus wind speed versus power output; ambient, nacelle, generator, and radiator temperature.
Real-time information is available to engineering teams 24/7 and allows them to take timely action. The team can apply remote corrective action or, if this fails, issue a work order for an on-site physical inspection.
Drone inspections help collect a lot of valuable data at a fraction of the time and cost, compared to manual methods. Rather than sending personnel rope-climbing up the flare stacks or descending deep into shafts, you can dispatch nimble drones to map the terrain, collect thickness measurements, or inspect for visual damage.
Thanks to specialized payloads, drones can bring extra efficiencies to processes like:
For infrastructure inspections, drones remove the need for commissioning aircraft, bringing in lifting equipment, or scaffolding construction, dramatically reducing costs. An oil refinery saved over $150K in crane rental and labor costs by opting for a drone-based stack inspection with a Voliro drone.
In the UK, drone bridge inspection saved the local authority about £1 million in costs and greatly improved the data quality. A drone with an RGB camera and LiDAR sensors provided accurate footage of hard-to-reach areas, leading to early identification of issues like corrosion, cracking, and cable wear.
Apart from making inspections less cumbersome, drones also bring extra speed and accuracy to measurements. For instance, with Voliro ultrasonic transducer payload, inspection crews can collect up to 120 thickness readings per hour.
One outside storage tank inspection takes about 75 minutes, meaning teams can inspect at least five storage tanks per day. Faster inspection times mean lower downtime and more efficient labor use. Moreover, you can schedule more frequent inspections to monitor infrastructural changes over time.
All the drone-collected data can be streamed into an asset management system for historical benchmarking, real-time assessments, and even predictive decision-making,
Machine learning (ML) and deep learning (DL) methods add a ‘predictive’ or even a ‘prescriptive’ component to condition monitoring.
Instead of just getting a 48-hour service alert, you can forecast service degradation over time and model different scenarios to optimize TCO and minimize downtime. For instance, in the mining industry, many draglines have duty monitors that predict weld failures on the boom, using fatigue models and strain measurements. Traditionally, these models have been statistical. But lately, most predictive maintenance models are powered by AI.
Razor Labs, for example, built an all-in-one automated predictive maintenance platform for the mining industry. It uses AI sensor fusion — a technology for integrating machine condition monitoring data from multiple sources and applying AI models to produce more reliable insights. The tool can run advanced failure root cause analysis, accurately predict mining equipment failure, and suggest prescriptive servicing actions. The company claims its system has prevented over 1,000 yearly downtime hours on tracked conveyors and reduced downtime by 25% on stockyards.
Caterpillar, in turn, developed a machine learning model for rapid detection of engine oil dilution, which can lead to engine damage. It can detect dilution in just 2.4 hours (down from 10 days) and has saved one customer over $360K in maintenance costs.
Another company’s model can detect wheel slippage, which can lead to brake failures and reduced life of the tire and axle assembly. Based on condition monitoring data, the team could timely notify the customer saving them about $500K in downtime costs.
In addition to early issue detection, a PdM-empowered schedule can also reduce manual inspections by over 50% without sacrificing operational quality. With remote condition monitoring, you don’t need to constantly have personnel in the field, reducing the time employees are at risk.
Digital twins are a new generation of predictive maintenance platforms, offering 3D asset visualization, real-time monitoring, and advanced predictive capabilities for modeling different asset usage scenarios.
For example, a digital twin of a solar farm can visualize all facility components in 3D — from panels to inverters — using footage, captured by drones and remote cameras. It allows crews to zoom in on the tiniest details without conducting multiple on-foot inspections and continuously analyze the assets’ health using data from sensors.
Moreover, they can further wrangle with the data to evaluate how different parameter changes will affect power output, for instance. Or rely on recommendations, provided by AI algorithms. In Germany, researchers at Fraunhofer Institute for Solar Energy Systems ISE have developed a digital twin that can automatically position solar panels throughout the day to provide sufficient light to nearby plans, while also optimizing yields based on grid conditions and tariffs.
Danish Sund & Bælt Holding, in turn, created a digital twin system for maintaining the Great Belt bridge. Previously, the team had to visually inspect over 300,000 square meters of concrete every six years. That was extremely labor-intensive.
Developed in partnership with IBM, the new digital twin streamlines ongoing monitoring and maintenance. It’s based on data from data maintenance records, design documents, and recent drone footage to identify cracks, rust, corrosion, and ongoing stress, allowing teams to better anticipate and predict repair work. Engineers can model how the traffic loads affect the most critical areas and gauge the change effects on the bridge.
Moreover, with access to the latest data, Sund & Bælt Holding can accelerate work order management. There are very narrow windows for bridge closure, and the model allows engineering teams to prepare most of the work before descending a five-mile tunnel.
Condition monitoring helps O&M teams gain deeper visibility into the assets’ health and performance. Thanks to reduced sensor costs and commoditized access to the cloud, you can aggregate data from a wider range of systems, seamlessly extending a digital analytics pane over physical assets.
Specialized inspection drones like Voliro also help improve the frequency of aerial asset inspections without augmenting the costs. Purpose-built for NDT at heights, our drone comes with six different payloads for collecting wall thickness measurements, taking dry film thickness measurements, lighting protection system testing, and corrosion under insulation detection. Collect conditioning data at a faster pace with greater predictability and fewer risks to your personnel.
