Looking Back - 2019

Blackstart Reliability in 2019

In April of last year, I branched out on my own to start Blackstart Reliability. I count myself fortunate to be able to take on this challenge, and I couldn't do this without my family, friends, and wonderful colleagues and connections here on LinkedIn. I owe many of you a great deal of thanks. As I found myself looking back on what I've attempted to do in the last year, I noticed some consistent themes. (Photos are linked to my posts over the last year.)

Theme 1 - Bad Actor Assets

How we deal with our losses matters. Bad Actor management and resolution processes was a consistent theme.

Understanding how we can break the cycle of repair and replace, month-in and month-out requires a plan.

One tried & true method for understanding how to deal with a problem is to figure out how to describe the problem effectively. For problems in reliability, sometimes that means understanding what reliability looks like from a failure data perspective.

When we understand our reliability data using above, we can begin to understand the trends of our bad actor equipment and in our organization. Are we running our equipment consistently to the ragged edge (leading to wear out), or are we missing consistency in our start-up procedures (leading to infant mortality).

As we work to understand failure, we naturally come across many failure modes. In October, we celebrated International Bathtub Day, where we took the opportunity to acknowledge the Bathtub curve and how and why it is still relevant. This curve is particularly interesting when dealing with equipment with multiple failure mechanisms of a similar nature.

In 2019, Blackstart Reliability applied some of these tools as part of it's service "DataStart for Bad Actors" on a chemical metering pump to reveal not 1, not 2 but 9 different failure mechanisms for a pump (and importantly how to properly characterize and cost-effectively mitigate the failures).

As "DataStart for Bad Actors" took shape, we began to learn that sometimes reliability issues for a "Bad Actor" may be overshadowed by overarching system issues. This led to an increased awareness that Bad Actor resolution many times requires a multi-specialty approach (leaning on the expertise of multiple parties).

The more I dug into Bad Actors, the more I came to understand why "DataStart for Bad Actors" was important. Whether we were trying to justify installing a VFD or to create a new PM, understanding our failure patterns was going to be essential.

If you'd like to learn more about "DataStart for Bad Actors", send me a message.

Theme 2 - Condition Monitoring

One of my original focal points for Blackstart Reliability was to understand current disruption taking place within predictive maintenance, <ahem> condition monitoring. What started for me as an interesting data problem looking at a typical question encountered by maintenance ("Will my motor last until the shutdown or do I need to make a 'proactive' repair now?"), turned into "DataStart for Predictive Maintenance". In 2019, we applied "DataStart for PdM" to condition monitoring programs such as Ultrasonic, Lubrication, and Vibration analysis.

As I started to talk to more and more people about their programs, I wanted to understand their purpose at a basic level - Why do you CbM?

Some of us used condition monitoring to attempt to eliminate unexpected failures. Others aimed more moderately (C. 90% of potential failures addressed before failure?)

What we showed with DataStart was that preventing failures with condition monitoring requires us to understand the P-F interval (time from a detected potential failure to an actual failure in the field).

Of course, understanding the "P-F Interval" is considered by many to be the holy grail of condition monitoring. Just how hard it is to understand is made clear by many texts, but none so clearly highlighted the problems as John Moubray's RCMII text where he resorted to descriptions of "several weeks to months" to describe P-F Intervals for most methods.

So, how did Blackstart Reliability make progress in this area? We started by combining data typically found in different sources, along with what we know about how defects progress through different stages and we ended up "DataStart for Predictive Maintenance".

By understanding the nature of our raw data (and it's limitations), DataStart is able to reconstruct an estimate of the risk of running that motor to the next outage, and importantly also the losses we would incur if we repaired it too soon. I refer to finding this "Just-Right" zone for maintenance as the "Goldilocks Principle of PdM".

Armed with this new data, we can ask ourselves this question (as I posed it last Halloween!). Is our condition monitoring program reactive or passive?

As we started looking closer at some condition monitoring programs, we found several areas for improvement - such as optimizing inspection frequencies, optimizing our maintenance planning and ultimate intervention/repairs. Lastly, the analysis forces us to take a good luck at the outputs of the program (is it doing what we need at a fundamental level?). Answering this fundamental question (with data) can really help us to modify how our program is structured, how and who collects the data or performs the analysis, and what standards we are using.

All the while, we started to see how the approach with DataStart (based on reliability fundamentals) differed from what we were seeing in the marketplace.

Rather than push us to be more reactive by getting mobile alerts (at home of all places), this 4th stage of the industrial revolution was finally going to give us something back of real value: TIME. How we spend this gift of more time is up to us. The post below was an example asking how we could use it.

Overall, I was very happy with the progress made in this area in 2019. I am extremely excited about what 2020 will bring and some of the new partnerships on the horizon.

If you'd like to see what "DataStart for Predictive Maintenance" can do for your program, send me a message! (PDA = Predictability, Detectability, and Actionability)

Alejandro Erives

Owner & Reliability Liaison - Blackstart Reliability LLC

Blackstart Reliability's Unique Services:

1. Bad Actor Asset Management & Analysis: "DataStart for Bad Actors"
2. Condition Monitoring Program & Maintenance Optimization: "DataStart for Predictive Maintenance"
3. OEM Services: "DataStart for OEM's"

In Blackstart Reliability's DataStart model, success is dependent on your input whether you are the end user, an OEM, or a product or application SME.

DataStart is a trademark of Blackstart Reliability LLC.

Copyright (c) 2019-2020 Blackstart Reliability LLC