Predicting the Unplanned Outage

In 1997, two Dilbert comics featured a dimwitted boss demanding advance notice of "unplanned outages" (here and here). The strips were funny for their outlandishness, yet purportedly based on real-life horror stories from the trenches of cubicle workers.

Today, the strips are outdated, no longer outlandish. Outages are, in fact, predictable… with the right technology in place, that is.

That technology is predictive asset maintenance analytics, which is being put to work in a number of industries. Automotive is one.

Two years ago, we learned that Volvo launched predictive asset maintenance for its products based on SAS Warranty Analytics and embedded analytics in its vehicles. The year before that, Toyota teamed up with to introduce Toyota Friend, a Salesforce Chatter-based asset management solution that interconnects you, your car, your dealership, and Toyota itself. Both Volvo and Toyota's solutions help give them, and their customers, a heads-up when something might be about to go wrong with one of their vehicles.

The oil and gas industry also is deriving benefits from predictive asset maintenance on the ocean floor, as we learned here on All Analytics. (See Shell Taps Big-Data From Way Down Deep.)

Most of what goes into predicting and diagnosing equipment failure and other asset management analytics is simply putting information already on hand to good use, suggests Anuj Marfatia, an IBM program director who gave a presentation on predictive asset maintenance analytics (and shared the Dilbert cartoons) at a recent conference. Still, many organizations ignore their data -- commonly deleting otherwise useful data after six to eight months, he said.

Part of the problem is that a pervasive culture of big data is overwhelming the enterprise mindset. The solution, therefore, is to break the big data down into "manageable data." "Where is your biggest problem?" Marfatia asked rhetorically. "What are the processes associated with it? …Who are the people associated with it?"

He wants specifics. He wants enterprises to understand their assets. He wants managers to be able to accurately measure the size of their issues. And he wants them to work it all out on a whiteboard.

"It's really being able to take bite-sized pieces. Managing all [of the data] at once has not been successful."

Marfatia related a case study of an automaker that had all of the data it needed to fix a serious problem in its manufacturing process -- but didn't analyze it. Once the company did so, it was able to predict when failure would occur. The data revealed that these failures were consistently happening during a certain time of day and that the equipment was at its hottest during that time.

With this analysis, a manager familiar with the factory layout instantly diagnosed the problem: sunlight. During the time of day of equipment failures, "light came in [through the windows] and heated up the engine blocks."

By accurately diagnosing what Marfatia called an otherwise "undetectable" problem and acting to move its essential processes appropriately, the automaker realized an 80% savings within a single year.

Indeed, big money is at stake when it comes to predicting "unplanned outages." Navigant Research recently reported that in the utility sector alone, asset management and anomaly prediction analytics are a $2 billion-plus industry; that figure will more than triple by 2023 as smart grids get smarter.

It's easy to see why. Israel Electric Corporation, for instance, once had but 30 minutes of notice before a system or equipment failure; since deploying predictive maintenance analytics, the company can determine failure 30 hours in advance.

The difference that predictive asset maintenance solutions can make is huge -- and the ramifications are obvious. "If something is failing," Marfatia said, "you need to fix that as soon as possible."

After all, it's not 1997 anymore.

Joe Stanganelli, Attorney & Marketer

Joe Stanganelli is founder and principal of Beacon Hill Law, a Boston-based general practice law firm.  His expertise on legal topics has been sought for several major publications, including U.S. News and World Report and Personal Real Estate Investor Magazine. 

Joe is also a communications consultant.  He has been working with social media for many years -- even in the days of local BBSs (one of which he served as Co-System Operator for), well before the term "social media" was invented.

From 2003 to 2005, Joe ran Grandpa George Productions, a New England entertainment and media production company. He has also worked as a professional actor, director, and producer.  Additionally, Joe is a produced playwright.

When he's not lawyering, marketing, or social-media-ing, Joe writes scripts, songs, and stories.

He also finds time to lose at bridge a couple of times a month.

Follow Joe on Twitter: @JoeStanganelli

Also, check out his blog .

Clean Your Data with Visualization and Algorithmic Tests

Speakers at Bio-IT World explore techniques for biotech researchers and others working with big data to identify the accurate data in their data files.

Data Sharing: A Matter of Life and Death

Cooperation among medical researchers -- done right -- very simply can mean lives saved, but the research community needs education on how to execute on that collaboration.

Re: Predicating the unplanned in many areas.
  • 7/31/2014 10:06:50 PM

@Maryam that or happen upon a recall notice posted at the grocery store!

Re: Balancing act
  • 7/31/2014 8:21:28 AM

@SaneIT. It sounds like you had your ducks in a row when it came to support. It's still sad that the vendors would have those tools as a selling point and then not ensure that there was trust in their own systems.

Re: Balancing act
  • 7/31/2014 8:11:53 AM

@Jamescon, Yes it is q bit of a catch 22, they want you to use their tools and the toolset is a selling point for them but then they are largely ignored until you get deep into the support structure.  At a previous employer I negotiated an arrangement that got me past the first two tiers of support for any server hardware issue because it added 30 minutes on average to a call and we demonstrated a track record of being right when we called for support and I agreed that when we called we would have the standard traps prepared so that we could immediately send them to a technician on their side.  I know not everyone can do this but when it comes to high availably in a big data center you have to push for something outside of the standard support model.

Re: Predicating the unplanned in many areas.
  • 7/30/2014 11:27:31 PM

I can't wait, it will truly change our lives on another level. Today Google is our only option for finding about troublesome repair issues with cars or appliances. The process to fix both is still clunky at best.

Re: Balancing act
  • 7/30/2014 11:06:06 AM

SaneIT, the under warranty scenario will naturally play out differently since no revenue is generated but potential unnecessary cost maybe incurred. Besides, by nature service departments tend to be reactivate operationally and don't like dealing in probabilities. That's to their detrement.

Re: Balancing act
  • 7/30/2014 10:57:18 AM

@SaneIT, OK, I see your point. It sounds like you've got a good, solid approach based on your experience.

Re: Balancing act
  • 7/30/2014 9:04:42 AM

@SaneIT. Sounds like a touch of hypocrisy on the part of vendors when they give you the tools to anticipate hardware failures but then wait until it's actually break/fix time. It sounds like the service group's mindset or marching orders haven't changed; it's still about keeping the customer quiet while keeping costs down.

Re: Balancing act
  • 7/30/2014 7:34:05 AM

@Beth, I can see why there might be some reluctance to just give a customer what they want when they call because not everyone who calls is going to have gone through the amount of troubleshooting that I make my team go through before picking up the phone.   I wouldn't want to tell support staff in a company that large to just ship parts when someone asks for them and I understand they can't always staff highly technical people for the first line of support.  What I learned to do over the years was to tell them what I had to back me up before I told them what was wrong.  That way I could escalate the call without going through the standard "turn off and turn it back on" game.

Re: Balancing act
  • 7/30/2014 7:09:40 AM

@Jamescon, what bothers me most about that is the fact that most hardware vendors give you the ability to do this type of monitoring but they don't trust their own solutions.  I have traps that send me alerts when write errors hit a specific threshold, but try telling a hardware vendor that you've been seeing high write errors on a specific disk for the past two days and you know that the drive is about to fail.  The answer usually "call us back when it fails"  not, send us the logs so we can see what is going on.

Re: Predicating the unplanned in many areas.
  • 7/29/2014 10:57:24 PM

@Maryam this is the future for the Internet of Things. Our appliances will self-monitor and send crash reports to somewhere... Manufacturers will monitor data to see which units fail most often and from what cause. The industrial Internet is coming and analytics will be its best friend.

Page 1 / 3   >   >>