cancel
Showing results for 
Search instead for 
Did you mean: 
cancel
Showing results for 
Search instead for 
Did you mean: 

Predict time to failure using the ThingWatcher

LR_9796586
10-Marble

Predict time to failure using the ThingWatcher

 

The below image is part of a demo “ThingWorx Analytics Enablement Step 2: Use Case Discussion”

 

LR_9796586_0-1623942715035.png

 

 

The prediction of time to failure, in this case 356 days, is among the functionalities shown in this demo for the ThingWatcher, but in the documentation of this module the time to failure prediction is not reported.  In summary, the question is, it is possible to predict the time to failure using only the anomaly detection module? If so, which functionality of the module is used for it and if case that is not possible, which module of thingworx was used to get the time to failure prediction?

1 ACCEPTED SOLUTION

Accepted Solutions

@LR_9796586 ,

 

Unfortunately I do not have any other advice on how to use Anomaly Detection for your use case then, as its not design for TTF.

 

Using the Predictive Analytics method is the typical approach that is recommended.

 

I can see if one of our Analytics Consultants has any additional feedback, but I would also recommend that you reach out to your ThingWorx sales representative to discuss use case design options available for your scenario.

 

Regards,

 

Neel

 

View solution in original post

4 REPLIES 4

@LR_9796586 ,

 

Thank you for posting your question to the PTC Community.

 

Anomaly Detection is not the optimal method of performing a Time-To-Failure monitoring or prediction, as its more geared to live failure and alerting on a data stream (looking for anomalies as they happen).

 

Have you had the opportunity to review our Community Topic on TTF?

 

https://community.ptc.com/t5/IoT-Tech-Tips/Predicting-Time-To-Failure-with-ThingWorx-Analytics/m-p/5...

 

If that post provides the information you need, please mark this comment as Solution Accepted.

 

Regards,

 

Neel

Thanks @nsampat  I read the post you linked but the solution is not what I’m looking for, in the post the TTF is approached as a classification problem with only four possible values. In the “ThingWorx Analytics Enablement Step 2: Use Case Discussion” the prediction is a continuous value with granularity of days, a classification model as used in the link you provided does not accomplish that outcome .

@LR_9796586 ,

 

Unfortunately I do not have any other advice on how to use Anomaly Detection for your use case then, as its not design for TTF.

 

Using the Predictive Analytics method is the typical approach that is recommended.

 

I can see if one of our Analytics Consultants has any additional feedback, but I would also recommend that you reach out to your ThingWorx sales representative to discuss use case design options available for your scenario.

 

Regards,

 

Neel

 

View solution in original post

slangley
23-Emerald I
(To:LR_9796586)

Hi @LR_9796586

 

If you feel your question has been answered, even if it wasn't the answer you wanted, please mark the appropriate response as the Accepted Solution for the benefit of others with the same question.

 

Regards.

 

--Sharon

Announcements