Pre-trained analytics model - Predictive Maintenance

[GN_10324745]

Hello. I work as a PTC partner and we want to understand more about the journey of Predictive modeling. We have some clients that want the use case of predictive maintenance and we have some doubts about the best way to do this approach and the PTC suggestion to the partners about skills and the professional profile to provide this use case.

Beyond that, we want to know if PTC (or some partner ) has some pre-trained model for predictive maintenance (through vibration/temperature data) or if it will be necessary to build and train the model from the beginning.

[nsampat · Accepted solution]

@GN_10324745 ,

Thank you for posting your question to the PTC Community.

I can only point you to the Developer Zone for reference material, but if you wish to have additional enablement I would recommend that you work through your PTC Partner Contact to have engagement with the PTC Analytics Consultants Team who can help you on your journey.

If you do not know who that is, please let me know and I can see if one of those team members can reach out to you directly.

From the PTC Developer Zone: Build a Predictive Analytics Model

Regards,

Neel

[DmitryTsarev]

Just to supplement Neel's @nsampat  great answer, I'd also suggest to check a few more things on the ThingWorx Developer Portal...

1) Entire Predictive-oriented Learning Path, where you'd build an end-to-end system, starting from collecting data from sensors and all the way down to building a predictive model https://developer.thingworx.com/en/resources/learning-paths/vehicle-predictive-pre-failure-detection-thingworx-platform

2) The "Analytics Training Videos" series of videos https://developer.thingworx.com/resources/guides/analytics-training-videos

Generally speaking, the Developer Portal is a really-really great resource to obtain ThingWorx-related knowledge. Guys and gals @ ThingWorx team within PTC are doing a great job on sharink knowledge and best practices via this portal.

Other than that, while the resources above provide a great starting point and a "30,000 foot view", if you're new to Analytics / Machine learning / predictive maintenance, I'd strongly advise against trying to do a project on your own. While the tools (ThingWorx-based solution in this case) are more or less clear and understandable, the topic itself is pretty multidimensional and interdisciplinary so having someone with some experience under the belt would be a huge boost.