Unplanned Downtime

J

Thread Starter

Jeremy Pollard

How do yous guys and gals from an OEM perspective and/or a user perspective deal with reducing unplanned downtime? Is predictive and/or preventative maintenance the ticket?
Or is it RCM programs that do it for you?

For a column I am writing for Control Design!!! :) thanks all

Cheers from: Jeremy Pollard, CET The Caring Canuckian!
Crisis, necessity, change

Integrator, Educator, Consulting, Columnist - Control Design
 
B

Bob Peterson

As a system integrator about all we can really do is give people quality Equipment and provide whatever Diagnostics are available.

While there are some things you can do such as regular maintenance that may reduce unplanned downtime, the fact that it's unplanned means you really don't have any control over when it happens. Many times when things happen that are out of the ordinary, it takes time to figure out what happened and then to fix it. These days a lot of places don't even have spare parts on site so even once you figure out what's going on you may not have a spare part for a day or two.
 
I agree, "unplanned" means it's just that--unplanned. Historically, reducing unplanned downtime was attributed to planned maintenance and attention to detail. The latter took trained, experienced people to check and recognize potential problems. In today's world, those people are disappearing, and even when there are diagnostics people tend to ignore them--because the equipment is still running so why shut it down? And is it an unplanned outage if someone recognizes a potential problem or some diagnostic indicates an impending shutdown that can be avoided with an early (unplanned) outage???

In my field (power generation) the focus is on predictive maintenance and being able to extend the periods between planned outages based on data--hordes and hordes of data, which means lots and lots of sensors (and redundant sensors) and usually on off-site entity to monitor the data and plan the outages. BUT, when something breaks, it usually goes horribly wrong and the unplanned outages can last longer that the planned outages--and cost more in parts and lost production.

And, it's also becoming more and more of an issue that "stores" is its own profit and loss entity these days, with companies and plants believing they can reduce stock and thereby reduce expenditures and inventory--saving money. But, when you need a USD150.00 card and it's going to take two days to get it and the plant is shut down because of that (meaning lost generation/revenue) that never gets factored into the money-saving equation.

The longer I'm in this business the more I fervently believe we are all risk managers. Some are better at it than others, but, we're all gambling to a certain extent with the decisions we make about deciding to extend an outage or shut the machine down now, only to have it possibly have unanticipated problems shortly after the planned outage. People with good knowledge of the process and the systems usually make very good decisions; those with little knowledge of the process and/or systems don't usually make good decisions. And those that make the decisions based on numbers in cells on worksheets in spreadsheets--well, they often make the worst decisions (from an unplanned outage perspective). There are just intangibles ("things") which can't be quantified or anticipated with a formula in a cell (or multiple cells...).

Predictive maintenance can be a great thing--if you have enough sensors (and redundant sensors!) and they're properly maintained and calibrated (think training and experience), and you have the right "rules" for interpreting and massaging the data. I think it (predictive maintenance) is in its infancy, and has a ways to go.

Unplanned outages are going to happen--usually because someone made a decision not to repair or replace some part during an outage (planned or unplanned)--or the planned outage interval was extended too far in the case of predictive maintenance. And, not having spares readily available is going to exacerbate the problem--sometimes greatly.

Quality equipment, meaningful diagnostic messages, and attention to detail by trained, experienced technicians and -operators is, in my opinion, the best way to reduce the frequency of unplanned outages. And having spare parts on hand can reduce the duration of unplanned outages. Those formulas in the cells on the worksheets of the spreadsheets need to take into account the lost production/revenue and manpower costs associated with not having the spare which reduced the inventory and expenditure costs.

You probably weren't thinking about predictive maintenance, but it's the coming thing for large, rotating equipment (and you probably weren't really thinking about large rotating equipment, either). There are many ways to improve the quality of the equipment--and even the diagnostics: DFSS (Design for Sick Sigma); LEAN; etc. And, they're being used all the time to improve equipment, as well as new manufacturing processes (additive manufacturing). But, I believe it's going to be a LONG time before machines get smart enough and people actually believe and trust them (the smart machines with all their sensors (and redundant sensors)) to better plan outages to reduce unplanned outages.
 
Sir, do you have any books/article recommendations about this?

>You probably weren't thinking about predictive maintenance,
>but it's the coming thing for large, rotating equipment (and
>you probably weren't really thinking about large rotating
>equipment, either). There are many ways to improve the
>quality of the equipment--and even the diagnostics: DFSS
>(Design for Sick Sigma); LEAN; etc. And, they're being used
>all the time to improve equipment, as well as new
>manufacturing processes (additive manufacturing). But, I
>believe it's going to be a LONG time before machines get
>smart enough and people actually believe and trust them (the
>smart machines with all their sensors (and redundant
>sensors)) to better plan outages to reduce unplanned
>outages.
 
boosterhidrogen,

I do not. It's a fairly hot topic in the power generation industry; you can probably find some information and white papers on the Internet. I don't know what GE formally calls their efforts in the power generation industry but they are probably linking it to their Predix IIoT (Industrial Internet of Things) software--if not now, soon. They seem to be using Predix for just about everything (locomotives; jet engines; etc.). It does require a secure Internet connection, and probably a dedicated computer at the site to gather information from the turbines and generators.

Hope this helps!
 
From a user perspective, reducing delay rate/unplanned downtime is commonly done using preventive and predictive maintenance strategies, but another rarely mentioned method is to use continuous improvement. Continuous improvement methods can be used in the following way : For each unplanned downtime or production delay event, implement a countermeasure to prevent re-occurrence of the failure, by analyzing and searching for the root cause. By doing this, the failures/breakdowns can be incrementally resolved.
 
>How do yous guys and gals from an OEM perspective and/or a
>user perspective deal with reducing unplanned downtime?

As indicated by comments here, there are multiple ways, multiple circumstances. (no cookie cutter, one solution fits all.) To my surprise reading here, even some who say 'you can't, that is why it is unplanned'??. But take your question "deal with it" meaning reduce, limit risk. That is where my comments are will be derived from.

First as you asked points of views from different perspectives (OEM, end-user, etc.), I wanted to let you know my views are from yet another perspective, as founder of TDC (True Downtime Cost methodology) and author of "The True Cost of Downtime" (2nd edition), and spending over 2 decades focusing our industrial maintenance and engineering training around reduction of unplanned downtime, I offer the following insights ...

To best solve a problem, you need to identify and measure the problem. More importantly, it has to be done accurately. See TDC https://bin95.com/news/Whats_True_Downtime_Cost.htm

So then once all the metrics data has been collected, you need to analyze and prioritize. Get the that lowest hanging fruit first for immediate cost reduction, and to motivate continuous pursuit. Like me, I prefer TEEP of OEE, because of accuracy and a much greater savings/profit. OEE can, and is often manipulated to make one look positive, plus varies with scheduling. TEEP is much less likely for that to occur, more accurate, and looks at the big long term picture, (Asset utilization). REF: http://downtimecentral.com/OEE_TEEP.shtml

As for OEMs, I believe they should put more emphasis on the design of their machines, automation, and software towards reducing the downtime of their end-users. Currently the majority of the market is more focused towards how can OEMs do their job easier, not how can their end-users can work with the product easier in the long term. The OEM only has to 'deal' with the machine 1-3 years when it is new. the end-user on the other hand, has to deal with the machine for the entire life-cycle, next 10-30 years. So the true downtime cost (TDC) over the life cycle of the equipment can cost the end-user enough to have purchased an army of machines. ... but most don't measure it, so don't realize it.
Ref: http://aboutoem.com/

Hope this helps, gets some additional thinking going on.
Don
[email protected]
 
J

Jeremy Pollard

Thank you Don (bin95)!!!! Very cool!!!!

http://www.htmsensors.com/spotlight.aspx

what are your thoughts on this 'solution' as such...??? I asked if it was relevant and didn't get any takers:) TIA

Cheers from: Jeremy Pollard, CET The Caring Canuckian!
Crisis, necessity, change

Integrator, Educator, Consulting, Columnist - Control Design
 
>Thank you Don (bin95)!!!! Very cool!!!!

Yep, that is what I am talking about for the OEM side. Before you purchase a machine, the end user can ask the OEM, is Spotlight ready? MobiNet compatible? Is OEE AND TEEP already incorporated in the analysis software?

A few years ago I did a lame social media post, about a pokemon game, where you put on 3d glasses or or use cell phone to view machine, and spot the problem sensor in real-time, real life, as you walk up to it on the machine. With thousands of sensors, you don't have to look up where sensor is located in an electrical layout, you can see it live on machine. http://www.htmsensors.com/spotlight.aspx is a step in that direction.
 
J

Jeremy Pollard

Thank you Don.. appreciated:) BYW seems you the first in augmented reality!!:)

Cheers from: Jeremy Pollard, CET The Caring Canuckian!
Crisis, necessity, change

Integrator, Educator, Consulting, Columnist - Control Design
 
Thread starter Similar threads Forum Replies Date
M General Automation Chat 3
M Process Control 6
D Process Control 0
Top