JMP Reliability Club – Kickoff Meeting

The first meeting of the JMP Reliability Users Group was held online on Thursday November 14th. Read a summary of the meeting and view the PowerPoint slide deck for JMP Reliability resources.

Summary from November 2024 Reliability Club’s First Meeting

Attendees:

Actions:

1. If you are interested in being on the Reliability Club Steering committee to help with planning and running meetings, contact Don McCormack @DonMcCormack , Jason Wiggins @Jason_Wiggins , Jerry Fish @JerryFish , or Charlie Whitman @charlie_whitman  at JMP.  Mehul Shroff @MShroff  from NXP has agreed to be one of the committee members.
2. Suggest topics for the January meeting and let Don McCormack know if you have data to share on any topic you suggest.

In late 2023, we sensed a need to have resources available to JMP users who have an interest in reliability. We decided that having a reliability club would be beneficial for JMP users online.  So, we decided to provide a space in the JMP Community where users can just learn more about reliability, ask questions about reliability, and let users identify reliability capabilities they would like to see in the software and resources they will need.  The Club will also be a way to share best practices and ask/answer practical questions from other JMP users.

The attached slides provide a list of resources available on the web for those who want to learn a little bit more about reliability.  After reviewing the resources, please start a dialog of what additional resources you may need and, if possible, give an example of what problem you would like the resources to address.

At each meeting, we will go over old and new business and someone will give a talk about a topic of interest and, if possible, a JMP demo using non-confidential data to solve a problem. Mehul Stroh from NXP has agreed to be one of the organizers of the of this group.  We are looking for a few others to help organize meetings. It will be minimal commitment on your part – help identify topics and speakers and run the meetings.  One of the JMP SEs will host the meetings on Zoom and work with the User Enablement Team at JMP to publish the link to the meetings and update the Reliability Club area after the meetings.

The attendees decided:

1. Meeting once a month, beginning in January 2025 is the preferred meeting frequency.
2. Topics will include how to address reliability challenges like:
   1. How do I do a degradation analysis?
   2. How do I do accelerated life testing?
   3. Specific topics related to things we run into like how to evaluate parts that are stressed until they fail, but they're not stressed constantly. And, we don't know what the stress is until they fail and after we measure the stress that they saw. So, it's sort of an independent variable, but we don't know how to take that into account. Also, our modeling it's also has two different categories, but should we model them separately. Don Lifke has an example he can share to describe the problem
   4. Interesting use of reliability techniques and transactional environments. We have tons of data where we could look at that, but we have to be creative on how we use time and I just haven't seen as much written about that lately. I don't know if you can use the same distributions or not as what we used in the old manufacturing days. Don thinks this might be treated as a survival problem, as is common treatment for these types of problems in the insurance industry. It may apply to recidivism in the banking industry or billing errors. Kevin S. may be able to describe the use case for such a problem.
   5. Handling cyclic stressing or ramp stressing.
   6. Life studies life studies with not just temperature, but having another continuous X-variable that is not time-based and trying to model that to get a shelf life study. Rob Pinshot may be able to describe use case for such a problem.  Charlie Whitman thinks parametric survival may be the approach.  He has used that for accelerated testing when you have multiple variables, like relative humidity and voltage and current and temperature – whatever the accelerators are. Don McCormack said that often works with hard failures. With soft failures, you're probably going to have to use some kind of non linear modeling which is available in JMP. But that hasn't been written about a whole lot hasn't been demonstrated a whole lot within JP. But it is possible if you're looking at soft failures.
   7. There is a lot of different platforms inside of JMP. So, Matheus Sguissardi suggested separating the topics like life distribution, Fit Y by X, Life by X degradation. One topic for each meeting. Announce the topic that you would cover so we could prepare ourselves. In some cases, you could bring some cases based on that specific platform. For example, say sample size explorers for reliability testing. Members could prepare themselves to bring some industrial cases, some situations that they are suffering, perhaps sending in the industrial cases related to those topic in advance.  Don suggest demonstrating using real data is more useful, so we would need people willing to share anonymized or non-confidential data.
   8. Jerry fish suggested that some people are very experienced and have some very advanced questions and some that are fairly new to the reliability platform, maybe we have one meeting for newer folks, one meeting for more advanced or we split a single meeting into an introductory topic and a more advanced topic to try to meet everybody's needs.
   9.  Igor Tavrovsky said his telecommunications industry goes through major innovations. OEM manufacturing doesn't have information about reliability so he started new approach and it's for four years asking them to do reliability demonstration tests that I design. But I would like to have more knowledge about handle this because it's new equipment, so they have no past data and have only few samples of how test that. Igor will check to see if his Legal Dept. will let him share examples.