Building the machine health matrix: baselining and trending data
Out of necessity, many plants run a patchwork of strategies to maintain systems, relying on run-to-fail strategies or scheduled preventive calibration along with periodic troubleshooting of glitch-prone components.
Short of refitting the whole plant with new devices, there is an Industrial Internet of Things (IIoT) strategy suitable for nearly any plant: bootstrapping a condition-based program to identify anomalies.
This kind of strategy employs wirelessly connected hand-held test instruments connected to smartphones and to the cloud. Managers, engineers, and technicians can then employ software to organize the data into a machine health matrix that, along with associated analytical tools, aids in preventive and predictive maintenance. This kind of condition-based program can also aid troubleshooting and root cause determination.
The following control applications benefit from baseline and trending analyses:
- Variable frequency drives (VFDs) and associated rotating equipment (motors, pumps, etc.)
- Control loops
- Programmable automated controllers (PACs) and programmable logic controllers (PLCs)
- Pneumatic components.
The sheer volume of data produced in plants can pose a challenge, and on-premise storage can risk data loss from corrupted hard drives, fire, or other disasters. Plus, what company hasn’t struggled with how to organize all the data and securely provide access to the people who need it? As time marches on, software evolves, as does data storage.
Even with a computerized maintenance management system in place, it’s difficult to analyze data collected by the typical range of technologies-voltage, temperature, current, and vibration. It’s a time-consuming task to assemble a holistic view of failure patterns for all plant assets.
For example, failing to isolate the source of an intermittent voltage spike on a PLC input at a critical line may cause troubleshooting teams to overcompensate by recommending the replacement of the whole device, while only one element was at fault. Yet that will still not solve the underlying problem.
Consider how technologically advanced and wirelessly connected tools, such as thermal imagers, power quality analyzers, digital multimeters, handheld oscilloscopes, resistance testers, and more would enable systematic data collection along the line.
- Create baselines of critical equipment during normal operation, then
- Continue to capture the same measurements over time so when failure occurs, the data can be used to determine failure mode patterns.
How to create a baseline with connected test instruments
A team is assigned to create and catalog test points of each individual piece of equipment in web-based software—painted targets on motors and near belts where infrared images are taken, voltage inputs and outputs on the PLC, motor couplings, and power to the compressor. Those test points are associated with their respective equipment in the software.
Technicians are then equipped with test instruments-vibration meters, infrared cameras, voltage or current modules, handheld oscilloscopes-even power quality analyzers-as well as software that can organize, store, and correlate data around multiple test points and multiple pieces of equipment.
When the line is running at normal operating conditions; technicians take baseline measurements at each test point with the connected test instruments and move on down the line. The initial test samples could be averaged across several identical pieces of equipment operated in the same application to create a standard deviation (one sigma).
The technician saves data to the specific test point and machine and automatically uploads data from the test instruments to the cloud, with a smartphone. (In some cases, engineers may choose to log measurements, with a voltage module or power quality analyzer, for example, over longer periods of time.) Those measurements are also immediately sharable anywhere via the app or web-based desktop software.
Four tips to capture data in one place
- Each test point is associated with equipment for identification online, and all data from individual equipment is stored in cloud-based, secure data servers.
- At a specified time, the technician takes the same measurements at the same point on the PLC and the other connected equipment. For the best results to create comparisons, the measurements are made under consistent, repeatable, operating conditions-ideally at the same time of day.
- When there is a problem on the line, record measurements not only from the device under test, but also other machinery that might be impacting that device.
- In the software, use analytical tools to compare the measurements and asset performance over time and use that information to make more informed predictive decisions.
Comparison and analysis
- Check status changes and thresholds set that were exceeded and check change in state from baseline.
- Compare the measurement trends, identifying any abnormal conditions such as overcurrent or events that correlate to a system stress condition.
- Use the software’s ability to compare data from the various technologies side-by-side and correlate it to specific events occurring with the PLC and associated devices.
Answer questions such as:
- Were the voltage waveforms on the PLC outputs malformed?
- Did infrared inspection test points show temperature spikes?
- Did the level of vibration increase on the motor powering the conveyor?
The creation of a matrix of time-stamped data from multiple test instruments over time facilitates finding the root cause of failures in a system. Rather than viewing one measurement on one machine, data can be correlated and analyzed with multiple types of measurements from individual machines, as well as compared with other critical equipment in the same system.
Wireless test tools and cloud-based software
- Aggregate all data without the errors prone with manual data collection or the cost of labor associated with collecting and organizing the data.
- Gather historic performance data and establish valuable performance baselines.
- View all the data on one screen, trying to avoid managing multiple pieces of software.
- Correlate multiple data points, such as voltage, current, and thermal images, to allow the spotting of irregularities quickly.
- Get multiple data points and trend graphs to compare each piece of equipment, which helps preventive maintenance planning and meets uptime goals.
- Access immediate updates on asset status from the field because techs can use their mobile phones to change the status of suspect equipment from normal to at risk.
- No matter what shift works on a problem every technician will, with consistency, collect system characteristics and supporting information.
Over the long-term, having access to this kind of smart data on file in an easily accessible place will help everyone on the team. Questions such as the following will become answerable: Has this happened before? Has anyone looked into this previously?
– Glenn Gardner is business development manager for Fluke Corp. and Jeff Black is managing editor, global content, for Fluke Corp. Edited by Eric R. Eissler, editor-in-chief, Oil & Gas Engineering, email@example.com.
- Create a matrix of time-stamped data from multiple test instruments over time, to facilitate finding the root cause of failures in a system.
- The sheer volume of data produced in plants can pose a challenge, and on-premise storage can risk data loss from corrupted hard drives, fire, or other disasters.
How do you currently collect data on machinery? Is it effective?
– See additional stories about cloud computing and the Industrial Internet of Things (IIoT) linked below.