Oil & y2k

greenspun.com : LUSENET : TimeBomb 2000 (Y2000) : One Thread

http://www.usagold.com/cpmforum/

ORO (12/15/99; 8:57:17MDT - Msg ID:21058) Cavan Man - chips in the old block This is far from my area. Process control was always the weak point for me. At least the upside down and transformed math of it. Strategies are clear enough.

What I was reading shows the following:

1. Embedded chips are a problem if they are installed in the typical distributed control systems. 1 in 10 chips in industry is or has an RTC.

2. These systems vary greatly in their structure and not all of them have RTCs with date tracking. Of those that do, the date does not necessarilly play any role in their logic function. Say 1 in 10 to 1 in 20 do.

3. The RTCs get their dates from the central control computer whenever anything is reset, which is quite often - sometimes a few times per hour or even every few mins. So - the important piece here is to know that all the distributed RTCs get an update immediately after the date switch. Controllers that don't get the change from the central computer could blow. So you have some danger there, but the date does not matter for routine operations that are hourly, only those that are one or more days apart - perhaps 5% of systems.

4. The control modules I worry about are those in the maintenance functions. These sit arround for days and even weeks without resets. Their failure is the long drawn out kind - when you think everything is running smoothly, the reserve pump is supposed to kick in when the oil was to be changed in the primary pump, but the pumps are controlled by a bad program and as it checked today's date, 1,15,00 - it was well before the 4 week period from the 12,10,99 maintenance (because 1.15.00 -12.10.99 is less than 4 weeks). The pump won't have its oil replaced and will eventually blow, or the maintenance people visiting the local oil reservoir to refill (if this is still done at all) will notice.

So we have 0.5% to 1% of systems being in danger. In a Refinery that is 150-500, in a Rig some 50 -100 - only we don't know which they are till there is a problem. So, everyone at the controls will be on edge, waiting for something to go wrong and will shut the whole plant down at the first suspicion of trouble - i.e. trigger happy. So by the end of 1 Jan 2000 we should not have any chemical plant or refinery in the US functioning at full capacity. There will be restarts galore and most will work. Later there would be a few problems with low frequency operations, and "brownout" type problems will recurr. We won't loose all operations all of the time. Just a significant portion every once in a while. Reprograming will be the daily drudge of the consulting techs. Lots of money - they cost the plant 100-150 per hour on regular work, what will they be charged when there is nobody to be found?

The central computers at the control stations will probably run ok in the beginning, but will eventually go bonkers here and there. Every couple of days, then every week, then once a month. Then it will settle down.

I would not be able to assess the depth of damage to production, I would guess that operations will come down 25% of the time, and overall output would fall 30%-50% at one point during Jan, and revive quickly thereafter. It sounds bad, but it isn't - the repairs will usually not involve replacing components but reprogramming from the central console. Maybe 10 or 20 systems would need to be replaced in the Rig, perhaps 50 -100 at the plant.

Distribution and scheduling in JIT systems is what bugs me. Add sprinkles of spot shortages of energy materials and raw materials. The problems are going to be not very sharp but sporadic and persistent, the inventory cushion is completely gone from all mechanical operations, and much reduced in the chemicals businesses. Like all protracted supply shocks, expect price inflaiton, reduced output, bankruptcies, and a lot of rich electronics technicians.

Take this with a lb of kosher salt.

That's been my thinking to date on this subject. Mind that I have not worked in a plant in over 3 years, and have not worked in a refinery since 1985 (e.g. programmed display layouts for Honeywell computer consoles in refinery)



-- Bruce (Bruce.dague@excite.com), December 15, 1999

Answers

What do you replace the embeddeds with, where are the replacements manufactured, and are there a supply that would be available for replacement? How old are the embeddeds and is anyone around that understands how to reprogram? Because you have not worked in a refinery since 1985 Bruce, that is also scary. Because programmers come and go and equipment has been replaced, it would be like searching for a needle in a haystack. Any disruption in manufacturing is going to be felt all the way down the line.

-- ~~~~~ (~~~@~~~.xcom), December 15, 1999.

Can we please get to the Bottom of these imbeds?

3%-5% to .01% - .05% some will fail-some won't. Some will fail if they shut down before 1/1/00 and then start up again. C'mon if there is this much confusion about chips in plants it does not bode well IMHO. Sounds to me like one thing is for sure--and you can take this to the bank (oops!)

NOBODY KNOWS WHAT THEY ARE TALKING ABOUT-FROM AN ABSOLUTE STANDPOINT-- -

-- d----- (dciinc@aol.com), December 15, 1999.


On the face of it your scenarios are encouraging. But consider this story, which actually happened about twenty years ago: a man driving down a freeway in CA realized he needed something in his glovebox, which was locked. So he pulled the keys out of the ignition to unlock the glovebox. Bad idea. The steering locked up, he went onto the median and crashed. Follow this metaphor? I can guarantee you, if that many processes in our juggernaut industrial economy need to be temporarily shut down all in the same time frame, we are going to crash. This particular guy survived to laugh about his stupidity. I don't think we will be so lucky.

-- StanTheMan (heidrich@presys.com), December 15, 1999.

So we have 0.5% to 1% of systems being in danger. In a Refinery that is 150-500, in a Rig some 50 -100 - only we don't know which they are till there is a problem.

===============================================

Somebody emailed me and asked me to look at this thread, which is the usual way I end up posting over here nowadays. Anyhow, usagold just claimed that an oil rig has (1/.005 *50) or 10,000 to (1/.005 * 100) or 20,000 separate systems on board. Or half that number, if you take the 1%.

Common, land based oil rigs aren't even remotely close to that figure, and I am counting pure electromechanical and hydraulic systems as well as electronic systems. Same for the figures on refineries.

Somebody needs to look up the definition of system. It DOESN'T mean 'a single part'.

-- Paul Davis (davisp1953@yahoo.com), December 16, 1999.


Moderation questions? read the FAQ