Preventive vs. Predictive Maintenance - What is the difference? TS 16949 - 7.5.1.4.

[Rosana]

what is the difference between Preventive and Predictive maintenance?

can someone explain this to me with an example?


Thank you

Rosana

[Sporty]

Per ISO/TS16949,
Predictive Maintenance: activities based on process data aimed at the avoidance of maintenance problems by prediction of likely failure modes to prevent disruption of production.

Preventive Maintenance: action taken to avoid causes of equipment failure and unplanned interruptions to production.

Predictive would be looking at data and trying to predict what might go wrong, like a weld tip that fails at approximately 500 welds.....you would put in a program to replace weld tips at 450 welds.....

Preventive is a program, whether monthly, daily, yearly.....where you perform maintenance to keep your equipment running smoothly.....change oil, clean, etc....

I hope this helps.

[Rosana]

Thank you for your help

Rosana

[paulnguyen]

Hi Rosana,

IMO, both preventive and predictive maintenance are activities that involve inspection, repair if necessary, of a given subject.

The differences are how often and the sourced data that conclude the activity.

Predictive maintenance is resulted in known failure facts or historical failure data. From those data, we have known that the subject IS going to fail. and the frequency is set from there.

Preventive maintenance is regular frequency inspection activity set by recommendation, experience or common practice.

Example of this is the case of our car. Changing oil every 6 months and replacing tires every 80,000 Km are cases of predictive maintenace because we know that if we don't do that the car will give us trouble.

Checking the oil level every month (some cases, daily) to see if there is any leak, visual inspection on tires every morning to see if there is any nail, etc. is the case of preventive maintenance. Again, the frequency is based on each individual car condition, road condition in the driving area.

That was my thought, see if anybody has other opinions? or better way to explain it.

Paul N.

[Dean P.]

Paul,

I see it a little differently than you. Say you have a piece of equipment with a gearbox. You change the gearbox oil every 6 months to keep the oil fresh. This is preventive maintenance.

You also analyze the gearbox oil for the presence of metal each time you change it. You know that when you metal content reaches a certain point, it is time to replace the gears. This is predictive maintenance. Basically, predictive maintenance is something you can track on an X-Y graph, and look for a trend when there is a known point of failure.

In your car example, you rotate the tires every 10K as a preventive measure. You inspect and record the tread depth, and once you reach a certain limit, you replace the tires (this is predictive). Based on the data, you can predict the date of failure, and thus do your preventive maintenance before you get there. Does this make sense?

[paulnguyen]

Hi Dean,

I've received your point about the analysis, study, trend to predict failure for the predictive maintenance activity.

It is still a little cloudy for me regarding the changing oil every 6 months to keep it fresh. From what source of data that the frequency of 6 month arises? Is the gear box still functioning normally if I don't (or I forget to) change it until 12 months later?

If there is a study on the oil vicosity and conclude that it needs to be changed on the gearbox at a given interval, wouldn't you consider it a predictive maintenance?

Or, maybe we are talking about 2 different cases. If there is a prediction based on study, historical data, that is predictive mainternance.

Regards,

Paul N.

[Dean P.]

That's how I see it, Paul. Preventive is something you do because you know that if you neglect it, things will go wrong (ie. changing your oil, greasing equipment). However, predictive is based on actual data, and can predict the exact point of failure, which allows you to plan you activites to avoid any downtime.

The 6 months timeframe is usually a 'best practice' as provided by the OEM, and may be based on data, however we are not collecting the data so it is a preventive maintenance that we do. If we sampled our oil and had it tested for viscosity, etc., and saw a trend in the data, then we could predict when we need to change the oil - predictive maint.

Either way you look at it, I think it is one of the areas that everyone has trouble with during an audit (at least we always did!). Good luck.

[paulnguyen]

Thanks, Dean.

Paul N.