On the Rise: Long-distance Repair

G'day gals, guys & gurus,

I'm retired but remote diagnosis has been around or at least the dream of if for almost as long as I can remember. It's definitely a goal that should be developed and it does have the potential to save a significant amount of time, effort and costs, particularly when you are talking about equipment that is located in remote regions of the world that are difficult to get to.

However, bitter experience has revealed that while it's a good idea when implement it is rarely as beneficial as was hoped for. One of the reasons for this is that to effectively remotely diagnose a problem you need to have seen the problem before and faults are tricky little buggers that seem to be capable of reinventing themselves in ways the design/repair engineers had previously thought of.

However, remote diagnosis if coupled with long term trend monitoring combined with very experienced field/service/design engineers has the potential to predict faults before they occur and therefore prevent them from happening.

Unfortunately repairing faults before they occur or even remote fault diagnosis is hard to quantify the success of and therefore extremely difficult to quantify management that insist on seeing a financial return on investment.

Until some way of quantifying the success of remote diagnosis and pre-fault repair on a financial basis then there is little hope of such forward thinking concepts getting much past the good idea stage.

Regards, masu

Having Stuxnet and other hacking possibilities around and attaching themselves to vital machinery controls would surely predicate the use of quarantine machines between the machinery controls and any net over which a diagnosis and treatment is given.

Even then there would always be the possibility of infecting the quarantine machine.

Otherwise I go along with No.2 which I can't view as I post this.

G'day duikerbok,

That's a very good point, however, many of the problems with security for control systems can be overcome by using better and more secure and dare I say proprietary operating systems rather than open source code or cut down Windows.

Open source operating systems are fine where you are primarily just manipulating files and data but using them to drive complex machinery is fraught with danger not only from hackers and malicious code but on the reliability front as well.

On the other hand if you used a specially designed and proprietary operating systems you have far more control over who, when and why access to your systems is gained. You can still use the internet for handling the communications, but properly written code with decent encryption of the communications should be able to pretty much stymie any hacker.

Remember, Windows was originally designed as an operating system for personal computers not for controlling things like nuclear reactors or power stations so using it for such critical applications is just asking for problems.

Just my opinion.

Regards, masu

But most of the scientific world operates some of the most complicated equipment humanity has ever built using various forms of Linux. The entire EPICS data collection system is all open source. Now if you think that operating the little particle accelerators at Argonne, Brookhaven National Labs, CERN, and many other research facilities as simple machines then you maybe correct.

Oh and as far as security, open source is by far much more secure. There are no hidden proprietary sections of code that things fly in and out of with the guarantee of a traveling salesman that all is secure.

Another complication to the entire idea of any and all repairing of any system is that fatigue and failure rates of many systems have become very well understood today. Very often today when any of the critical parts of any system fail after the infant mortality period most of the remaining critical parts were soon to fail, too.

The plan has been around for a while, and as Masu hinted at it is less that perfect for diagnosing a breakdown. But for monitoring performance and preventing breakdowns the system seems to work well from what I saw. I spent some time working for a company over 15 years ago that was tasked with keeping a fleet of trucks and other mining equipment running 24 hours a day. I mean the big trucks. They were located almost 2,000 miles away from the construction site but had systems that monitored everything on these monsters from oil temperature all the way down to tire pressure. They could schedule maintenance and get parts to the site before the truck had to be shut down or broke down. I was told downtime for one of these things is measured in thousands of dollars per hour so keeping them running 24 hours a day was paramount.

They told me the next thing they were planing to replace with automatic monitors was the driver.

Great Point!

Monitoring systems that detect a partial failure before a complete failure have saved many corporations a lot of money. People forget that one of the first uses of the Fast Fourier Transform analysis (FFT) came from the vibration reading microphones attached to the continuous printing presses of the newspaper industry. By listening for a partially failed bearing with this type of an unbiased electronic ear, catastrophic failures with days to weeks of down time for a machine were replaced with half hour scheduled maintenance.

Then again, all your diagnostics and monitoring is totally useless it you can't get the spare parts. One of the companies I worked for was originally managed by engineers that had come up through the ranks, however, over time as they retired they were replaced with accountant type managers that didn't have the first clue about engineering.

One of the things these gits instigated was a just in time policy for spares where they cut back the spare parts stock by about 90%. The idea was that instead of keeping them close to where they were needed they would keep them in a central sore and thereby reduce the number of parts required. They would then be shipped to where they were needed supposedly arriving just in time.

Guess what? It was more like a never in time than a just in time system and it got just about every customer off side when they had a system down for two days instead of two hours.

I realize that detailed monitoring should be able to predict failures before they happen but failures have a nasty way of sneaking up on you and then coming in bursts with similar failures happening in multiple machines and locations at around the same time. When you only have one spare that happens to be on the other side of the continent and there are two machines down that both need the same part you are going to have at least one very upset customer.

Proper spare part stores are crucial and unfortunately they have been an area of neglect for many companies over the last few of decades.