From: Anonymous <nobody@replay.com>
Date: Wed, 2 Dec 1998 07:39:08 +0800
To: cypherpunks@cyberpass.net
Subject: Re: y2k+oil refineries
Message-ID: <199812012241.XAA19806@replay.com>
MIME-Version: 1.0
Content-Type: text/plain



North's "scrap the refinery, the chip broke" scenario was forwarded to
comp.software.year-2000 by Paul Milne, resident doomsayer, foul-mouthed
religious fanatic, economic Marxist, and all around kook.  It prompted a
very knowledgable reply which pointed out how silly the whole idea was,
from someone with relevant experience.  Maybe next time Detweiler could
check c.s.y2k to filter out North's nuttier ideas before forwarding
them here.


Author: SAG
Email: stephen.and.marilyn.goldstein@worldnet.att.net
Date: 1998/12/01

fedinfo@halifax.com wrote:
> 
> From GN:
> 
> Category:
>  Noncompliant_Chips
> Date:
>  1998-11-30 13:28:22
> Subject:
>  Fuel Production Plants
> 
> * * * * * * * *
> 
> I have one very, very reliable source within that industry who tells me that
> the oil refining industry can't cope with the task. I am told that the problem
> of embedded systems can NOT be fixed EVER, no matter how much time were
> allowed. WHY? Because the refineries themselves would have to be dismantled to
> uncover these embedded systems. Essentially, the refineries would have to be
> destroyed and rebuilt!!! These embedded systems are buried within enclosed
> systems. Identifying, testing and replacing these systems is too impractical.
> It is financially unsound to do so. The better option is to build new ones.
> There is no time to do this with less than 400 days left and the time to build
> a new refinery is 3 to 5 years.
> 
> So, what does this mean? My sources, especially the most reliable source tell
> me that it means that on January 1, 2000... the oil and gas refineries will
> cease operations. These facilities also have converted their inventory
> management control systems to small inventory levels so that inventory levels
> are running about 1 to 2 days capacity. A few years ago, the industry averaged
> a 1 to 6 month supply in storage tanks. Not so today.
> 
> The results: By 1/5/2000 there will be no gasoline, no diesel fuel, no
> natural gas, no heating oil, no fuel oil products at all. This means trains
> will have no fuel. Trucks will have no fuel. Cars will have no fuel. Electric
> Power Plants will have no oil for fuel, nor coal... because there will be no
> fuel to power the vehicles to get it to them. So, there will be no
> electricity. And you know the remaining domino schematic from that point.
> 
> ======
> 
> I have no way of knowing whether the embedded systems are just that, embedded
> in such a way that they can not be accessed. I imagine that some could be
> under fifty feet of concrete or otherwise inaccessable. If this is so, then
> there is not the slightest doubt at all, that it is indeed all over. Period.
> 
> I am not going to bandy about the issue of whether this source is reliable or
> not. It makes no difference. The only thing I am concerned with is whether
> these systems are in fact inaccessable. I know that not every system is
> inaccessable. The question is whether there are enough inaccessable systems to
> mean that the refineries etc. will not function.
> 
> Of course, I would love to hear someone counter this with evidence showing
> that what has been related above is not true, or it is an urban myth etc.
> But, I would like 'evidence'. Not conjecture or inuendo or suppositions.
> 
> If it were so, that there were indeed these inaccessable systems, would the
> companies involved let it be known publically? I think not.
> 
> If this report is accurate, the remediation is 100% moot as I have said all
> along, and you can kiss your butt good-bye.

I'm not sure if I want to play this game Mr. Milne -- you offer this
article which is, conjecture, innuendo and supposition but demand that
any counter story be documented by "evidence."  That's pretty tough and,
frankly, I'm not up to the task as my knowledge is with electric power
plants -- not oil refineries.

They do, of course, share some important characteristics and so I'll
give it a try.  The approach will be from two perspectives -- first,
I'll explain some power plant construction practices that contradict the
refinery story; second, I'll offer another review of the nature of the
"embedded" systems that might be at issue.

Power Plant (and refinery) construction . . . .

Remember, power plants and refineries are:
o Capital intensive
o Make increased use of computerized process control systems
o Operate a continuous manufacturing/conversion process employing high
temperatures and pressures.

Typical design life of a coal-fired power plant is 40 years.  During
that period, any number of components *will* fail and *will have to be
replaced.*  Wouldn't make much sense to put a $500million investment at
risk 'cause a $20.00 part failed in some inaccessible location.  Thus
EVERYTHING is accessible.

You want evidence and I'm not about to take you on a power plant tour
but even if you've never been inside a power plant the construction
practice is similar to that found in many other industrial settings and,
if you've ever been aboard a naval vessel or taken a cruise, ships. 
These settings are nothing like your home (or even an office) with all
the infrastructure components nicely hidden behind walls -- pipes are
exposed and clearly labeled -- power and control cables are laid in
exposed cable trays and risers.  Electronic components are often mounted
in racks with easy access from front and back -- probably not at all
like your computer or home theatre configuration.  All this is to PERMIT
MAINTENANCE, REPAIR and MODIFICATION.

So you need to understand that the problem isn't just maintenance, in
general, or for y2k.  Process plants need to be modified frequently --
market conditions demand a change in inputs or outputs.  Can't get
enough sweet crude?  Exxon doesn't close the refinery -- they modify it
to handle the new feedstock (I got the grand tour from a fellow I knew
who worked at Exxon's Baton Rouge refinery, years ago).

Want some pictures?  Here are a few from several "continuous process"
industries: http://www.powerprocesspiping.com/,
http://www.bdmechanical.com/ppiping.htm,
http://silverweb.nf.ca/m&m/cb-p&p.htm, and
http://www.shambaugh.com/process.htm.

Now consider the nature of these embedded components.  This has been the
subject of much discussion in csy2k.  I'd like to suggest that there are
three broad categories of components in this context:

o The control system(s) -- SCADA -- one or more intelligent nodes build
on traditional minicomputer, or more recently, microprocessor-based
server technology.  These are very accessible -- they're in, adjacent
to, or very near, the control room.  In an even more modern "distributed
control system," some of the capability will be located in different
areas of the plant.  Nevertheless, they still have to be very accessible
-- stuff breaks -- memories get hard, multibit errors, disk drives fail,
etc.

o Remote devices -- the control systems can't run without data and so
you have dozens to thousands of remotes -- devices that measure process
parameters (temperature, pressure, mass flow, volumetric flow,
acceleration and mechanical position (valve position or tank levels))
and a smaller number of remotes that can act on SCADA commands (start a
motor, open a valve, etc.)  Though most of these devices are pretty
dumb, some may have a bit of silicon-based intelligence and therefore
susceptible to y2k problems.  BUT, they all have to be accessible
because 1) they malfunction or need to be recalibrated and 2) plant
modifications often require removal/replacement to accommodate process
changes.

o Embedded systems -- I think that the consensus is that the real
embedded systems are those systems that are not obviously run by a
computer -- the power plants water chemistry analyzer or the refineries
gas chromatograph.  These are on the site but they're not in the "plant"
-- they're in the lab -- very accessible.

Now "accessible" doesn't always mean you can walk right up and touch
it!  Guess where the flue gas monitoring remotes (temperature, opacity)
are located -- top of the boiler or in the stack.  But here's the
scenario -- opacity indicator increases (flue gas has more smoke in it)
but the stack gas "looks" clear.  Do you shut down the plan on
environment limits or do you send a maintenance man to replace the
remote?

Now is it possible that conditions would be different at an oil
refinery.  Sure, like I wrote at the beginning, I'm not really familar
with these.  But, really, we'd be looking at some obtuse examples --
imagine that rather than using a tank level indicator to see how much
gasoline is in a tank, someone came up with the bright idea of putting a
strain gauge *under* the tank and determing the amount of product in the
tank but computing the weight on the gauge and using that to compute the
amount of product.  The device fails and, we'll all agree, it is not
practical to replace it.

Does the refinery stop using the tank?  I wouldn't think so.  They
retrofit a traditional tank level indicator (float) to provide the
needed information. 

HTH,

SAG