The Short Version: Plants use both names, and the official public pages do not all split them the same way. You will see condition-based maintenance (CBM) and predictive maintenance (PdM) used for overlapping work, so the label on a slide is a weak guide to what a program actually does.
Here is what the public pages say. DOE and PNNL define predictive maintenance against time-based preventive maintenance: PdM schedules work on quantified condition, while preventive maintenance schedules on a preset time interval. NASA names condition-based maintenance as work performed when the need arises, with asset performance monitored and compared to baselines so maintenance follows an indicator of deterioration or impending failure. ISO 17359:2018 is a general-guidelines standard for procedures to consider when setting up a machine condition-monitoring programme. Connecting that ISO scope to CBM or PdM strategy names is editorial synthesis.
So the useful question is not which of the two words a vendor or a team prefers. Ask what is being measured, and ask what triggers the work order. Those two answers tell you far more than the name does.
Independence and scope
Reliable Magazine does not sell reliability software, sensors, consulting, or certification. This guide is independent editorial analysis. It is not a DOE, NASA, PNNL, or ISO publication, and it does not speak for any of them.
Everything below is drawn from public pages: the PNNL O&M Best Practice maintenance approaches page, the DOE FEMP O&M Best Practices guide, the NASA Reliability Centered Maintenance guide, NASA NPR 8831.2F Chapter 7, and the ISO catalogue entries for ISO 17359:2018 and ISO 13372:2012. Where a public page uses a specific term, we keep that page’s term. We do not merge the three names into a single invented definition.
What each name refers to on the public pages
PNNL and FEMP foreground predictive maintenance. NASA uses overlapping predictive-maintenance, condition-monitoring, and condition-based-maintenance terminology. ISO covers condition monitoring and diagnostics.
Predictive maintenance (PdM), per PNNL and FEMP. The PNNL O&M Best Practice page lists four maintenance approaches: reactive or corrective (run-to-failure), preventive (time-based), predictive (fix it before it breaks), and reliability centered maintenance as a combination of the others. It describes predictive maintenance as measurements that detect the onset of system degradation, a lowered functional state, so that causal stressors can be eliminated or controlled before significant deterioration in the component’s physical state. Those measurements and their analysis indicate current and future system functionality. The page states the fundamental difference plainly: predictive maintenance bases the maintenance need on the actual condition of the machine, while preventive maintenance schedules solely on time. Named technologies on that page include thermography, oil analysis, ultrasonic analysis, vibration analysis, motor analysis, and performance trending with root-cause analysis. The DOE FEMP O&M Best Practices guide carries the same definition family.
Condition-based maintenance (CBM), per NASA. NASA NPR 8831.2F, Chapter 7, section 7.7.6, defines condition-based maintenance as maintenance performed when the need arises. The performance of the asset is generally monitored continuously during operation, and the results are compared to asset baselines. Maintenance is performed after one or more indicators show that equipment is going to fail or that equipment performance is deteriorating. The NASA RCM guide (section 2.2) presents condition monitoring (CdM) or condition-based maintenance as alternatives to traditional interval-based preventive maintenance, and it names failure precursors, quantified equipment condition, and improved repair scheduling as what this approach provides. That same guide is careful to say condition-based maintenance should not replace all interval-based maintenance: interval-based tasks still fit abrasive, erosive, or corrosive wear, material-property change, or a clear age-to-reliability correlation.
Condition monitoring and diagnostics, per ISO. ISO 17359:2018, “Condition monitoring and diagnostics of machines: General guidelines,” is the current edition of the general procedure standard (the 2011 edition is withdrawn). ISO 13372:2012, “Condition monitoring and diagnostics of machines: Vocabulary,” is the companion terminology standard, confirmed in 2023, intended to give users and manufacturers a common vocabulary. The public ISO catalogue abstracts describe those documents as condition-monitoring and diagnostics guidelines and vocabulary. Connecting them to CBM or PdM strategy names is editorial synthesis. The full procedures and definitions are paywalled, so this guide relies only on the public catalogue titles and abstracts.
Where the public pages overlap
Read the three sets of pages side by side and the overlap is obvious.
The DOE and PNNL description of predictive maintenance is already a condition-based idea. It bases the maintenance need on the actual measured condition of the machine, which is the same underlying logic NASA describes for condition-based maintenance.
NASA uses condition-based maintenance and condition monitoring for quantified-condition work, monitoring performance and acting when indicators cross away from baseline. That is the same trigger logic PNNL attaches to the word predictive.
ISO 17359:2018 gives guidelines for general procedures to consider when setting up a machine condition-monitoring programme, and ISO 13372:2012 is the companion vocabulary. Connecting those ISO documents to a CBM or PdM program is editorial synthesis. Neither ISO document declares a winner between the two strategy names, and neither DOE nor NASA publishes a single joint table that maps CBM against PdM term for term. The names travel in different documents from different organizations, and the underlying measure-then-act logic is shared.
Fit questions a plant can ask
The label matters less than the answers to a short list of concrete questions. Run a candidate program or a proposed one through these.
- What is measured? Name the parameter. Vibration, temperature, oil condition, ultrasonic emission, motor current, or a performance metric. If no condition evidence is being observed or measured, the setup does not match these source descriptions.
- What triggers the work order? A calendar or run-hours interval, or a measured indicator crossing a baseline. This is the single most useful question, because it is the split PNNL and FEMP draw between predictive maintenance based on quantified equipment condition and preventive maintenance based solely on time.
- Who interprets the data? A technician, an analyst, an automated rule, or a model. Interpretation is where a measurement becomes a decision.
- In the CMMS, is the task calendar or condition? Open the actual task record. A task that fires every 30 days is time-based regardless of what the program is called. A task that fires on a reading is condition-based.
- Is prognosis actually being done? Detecting a current out-of-baseline reading is one thing. Estimating how the condition will evolve is another. Check whether the program only detects present degradation or also projects future functionality, which is language the PNNL page uses for predictive maintenance.
No industry-average number is offered here, because the public pages this guide relies on do not support one for these questions.
One worked example
The names below are fictional and are used only to make the trigger logic visible.
Two plants, Cedar Line Bottling and Northgate Paper, run the same class of pump and collect the same vibration readings on it, on the same schedule, with the same sensors.
At Cedar Line Bottling, the vibration reading is logged, but the work order that actually replaces the bearing is generated by a fixed 12-month interval in the CMMS. The reading is reference information. The trigger is the calendar. On the fit questions above, this task is time-based scheduling with condition data collected alongside it.
At Northgate Paper, the same vibration reading is compared to an established baseline, and a work order is generated only when the reading crosses the alarm threshold. No reading crossing means no work order, even at month 12. On the fit questions, the trigger is the measured indicator, which matches the condition-based and predictive descriptions on the DOE and NASA pages.
Same asset, same readings, two different triggers, two different programs. Neither plant has bought a better product here, and this example does not declare a purchase winner. It shows that what you call the program is downstream of what actually fires the work order.
Honest limitations
A few things this guide cannot do, stated plainly.
The ISO 17359:2018 and ISO 13372:2012 procedures and definitions are paywalled. This guide uses only the public ISO catalogue titles and abstracts and does not reproduce paywalled procedures.
DOE and PNNL do not define condition-based maintenance as a separate labeled term on the maintenance approaches page. Their public pages define predictive maintenance and contrast it with time-based preventive maintenance. The CBM term, as a named category, comes from the NASA documents here.
NASA and DOE do not publish one joint, official CBM-versus-PdM comparison table. Anyone presenting a single tidy two-column table sourced to those agencies is combining documents, not citing one.
Vendor-specific splits, for example a claim that condition-based maintenance means fixed thresholds while predictive maintenance means an AI forecast, are out of scope. None of the public pages cited here state that exact split, so this guide does not assert it. If a specific product draws that line, that is the product’s framing to evaluate, not a definition from DOE, NASA, or ISO.
FAQs
Is condition-based maintenance the same as predictive maintenance?
On the public pages, the two names describe strongly overlapping work: measure the condition of the asset, then act when an indicator shows deterioration or impending failure. NASA uses condition-based maintenance for that idea, while DOE and PNNL use predictive maintenance for a very similar idea and contrast it with time-based preventive maintenance. There is no single official page that declares the two terms identical, so it is safest to compare the triggers rather than the labels.
How does DOE define predictive maintenance?
The PNNL O&M Best Practice page describes predictive maintenance as measurements that detect the onset of degradation so that stressors can be controlled before significant physical deterioration, with the analysis indicating current and future functionality. It states that predictive maintenance bases the need on the actual condition of the machine, while preventive maintenance schedules solely on time.
How does NASA define condition-based maintenance?
NASA NPR 8831.2F, section 7.7.6, defines condition-based maintenance as maintenance performed when the need arises. Asset performance is generally monitored continuously and compared to baselines, and maintenance follows once one or more indicators show that the equipment is going to fail or that performance is deteriorating.
What does ISO 17359:2018 cover?
ISO 17359:2018, “Condition monitoring and diagnostics of machines: General guidelines,” is the current general procedure standard for condition monitoring. Its public abstract says it gives guidelines for general procedures to consider when setting up a machine condition-monitoring programme. Connecting that scope to CBM or PdM is editorial synthesis. The detailed procedures are paywalled, so this guide uses only the public catalogue title and abstract.
Does condition-based maintenance replace all preventive maintenance?
No. The NASA RCM guide states that condition-based maintenance should not replace all interval-based maintenance. Interval-based tasks still fit abrasive, erosive, or corrosive wear, material-property change, or cases with a clear age-to-reliability correlation.
Which name should my plant use?
Pick the term your standards, your CMMS, and your team already use consistently, then verify the program against the fit questions. What is measured and what triggers the work order tell you what the program does. The name on the slide does not.
Where does condition monitoring fit relative to these two names?
Condition monitoring, as framed by the public ISO 17359:2018 and ISO 13372:2012 catalogue entries, is the measurement and diagnostics activity those standards name. Connecting that activity to CBM or PdM strategy names is editorial synthesis.
Related guides
- How to Start a PdM Program
- Reliability vs Availability
- How to Perform RCM
- How to Perform FMEA
- How to Calculate Asset Criticality
- Preventive Maintenance
- How to Calculate MTBF and MTTR
- Maintenance and Reliability Glossary
Sources
- PNNL O&M Best Practices, Maintenance Approaches
- DOE FEMP O&M Best Practices Guide (PDF)
- NASA Reliability Centered Maintenance Guide (PDF)
- NASA NPR 8831.2F, Chapter 7
- ISO 17359:2018 catalogue entry
- ISO 13372:2012 catalogue entry
Last updated: August 19, 2026. This guide is editorial analysis by Reliable Magazine.
