Re: [foxboro] Can't figure out a way to justify it.

  • From: tom.vandewater@xxxxxxxxxxxxxx
  • To: foxboro@xxxxxxxxxxxxx
  • Date: Wed, 12 Jan 2005 19:57:02 -0500

        This discussion has been very interesting and I can't keep out of it
any longer.  Alex knows 1st hand that we have been trying to decide how to
keep our control system up to date.  We have had many discussions and
meetings internally, with Foxboro, and with external sources that
participate on this list.
        I have come to a few personal conclusions:
1. All electronic modules will fail and will have to be replaced at some
time.  Why spend new money to replace old technology?
2. Our FBM's aren't redundant and will eventually fail.  A small portion of
those failures will cause downtime.  For us the cost of that downtime is
very significant.
3. As replacement parts get more scarce the cost to replace failed modules
will increase until the supply is exhausted.  Newer modules will cost less.
4. Foxboro can, and probably will, build some kind of form fit replacement
for 100 series I/O that can be mounted in the old enclosures and utilize the
existing nosecones and field wiring but their capablities will not be
anymore than what they are today.  Why spend new money to replace old
technology?
5. Existing IE-32 and 1x8, enclosures are expensive and are only designed
for mounting the legacy IA controllers and FBM's.  New control equipment is
designed to mount in 19" racks that can handle any vendors equipment, are
extremely flexible, and cost significantly less than proprietary DCS vendor
racks.
6. The Carrierband LAN has served us well and has allowed us to coordinate
our control throughout our site and maintain our system from any point of
access, but the 802.4  5 MB/S tokenbus is no match for todays communication
needs and we have had to constantly monitor CBLAN loading.  The 802.3 10MB/S
nodebus is also past it's prime. The MESH network will provide better
reliability, will be built with non-proprietary, (cheaper), networking
equipment, and will allow the CBLAN to temporarily coexist during our
migration efforts which we project will take 10 years to complete.
7. Ask your managers how many of them are using the same PC they were using
back in the late 80's when the legacy IA system equipment was developed?
How many are using the same operating system?  How many are using the same
applications to do their work?  Ask them if they think there has been any
valuable improvements in control technology since then.  When they
eventually had to replace their PC's did they try to buy one that had
exactly the same capability?  If not, why?

        The migration will never end as technology continues to improve.
New capabilities and software applications will entice users to advance if
they are proven to be reliable and cost effective.  User's that choose to
hang on to their legacy systems with no yearly upgrade plan will eventually
be faced with the big bang cost of total system replacement or will follow
the path of the US Steel industry in the 60's and 70's, losing out to more
efficient competitors.  If Foxboro isn't the best buy available on the
market at the time of the big bang, then a more competitive vendor might be
chosen as a replacement DCS!! 
        We have been continually upgrading our system since we 1st installed
it.  From WP-20's to WP-30's to WP-51B-D, AP-20's to AW51B-D, CP-10's to
30's to 40's to 60's.  We have about half as many AW51's as we had AP-20's
because they are much more powerful and capable but the overall cost is
probably pretty close to the same.  We have about a third as many WP51's as
WP-30's because they are more powerful and capable and we have heavily
utilized remote DM's. The WP51's are more cost effective than buying that
many WP-30's.  We have replaced up to 8 CP-10's with 1 CP-60 and the cost
wasn't much more than we payed for one CP-10 back in 1989.  The CP's overall
cost has dropped the most.  Contrary to many beliefs, the cost to buy a DCS
system has decreased over the years because of technology advances and
fierce competition within the industry.  DCS systems are a commodity and
vendors will face increasing pressure from users that are ever more
discriminating.  That's my prediction and I'm sticking to it!

Tom VandeWater
Control Systems Developer/Analyst
Dow Corning Corporation
Carrollton, KY   USA

P.S.  Included below are a compilation of interesting comments made by many
of the list members about this subject.

Alex Johnson (Foxboro) wrote:
I wish I had the key to offering something that would drive replacements.
So, what would justify an upgrade in the minds of you folks - short of the
"rip it out because we have a new system and won't support our existing one"
that some vendors use.  I'd really appreciate your thoughts on what would
drive the brownfield sites to upgrade.
  
Stan Brown wrote:
OK, I'll take a crack at that.

1. Offer us something we _want_ to justify (that's spelled UNIX!).
2. Offer increased reliabilty (Longer MTF, lower MTR).
3. Enhance self diagnostics.
4. Offer better easier to understand control functionality.
5. Make configuration easier.
6. Make configuration documentation easier.
7. Provide enhance connectivity to open databases.

John Kinsley wrote:
We are not planning to replace or upgrade all CP's and GW's.
What we did was look at the MTBF of these modules and our
observed failure rates and have concluded that the existing
modules should last for some time.  Our plan is to replace a
number of existing CP's and keep the ones we take out as
spares.  We're looking at replacing around 20% of the
modules which should give us enough spares to run for a
while.  We will monitor inventory levels of these spares and
failure rates.  We always have the option of replacing more
in the future when we see the inventory levels declining.
I'm curious if others are using this same approach.

Frits Shouten wrote:
About point 7.
Just about all process data that plant managers want reports from.
Hourly logs springs to mind. Maintenance reporting can follow from this.
Production data is another one. OAJ, Historian, Control database, FoxCAE,
etc.....  I've installed MySQL on a host(Blade) and just about everything is
managed from that.

Kevin FitzGerrell wrote:
1)  Switched networks allow for 8 segments on nodebus.  For sites that
already have 3 segment nodebus, this allows for easy extension of the
existing system to new plant areas without a CLAN.
2)  Modbus/Profibus fbms on CP60 are much more attractive than Integrator 30
solutions.
3)  B/B1 boxes are experiencing increasing incidence of component failure
(ram,NVRAM, floppys, HD, CD, Power Supplies) and don't run current version
of FoxView.
At plants with the 20 year old system, everything may well work about as
well today as when it went in, although maintenance and production costs
related to component failure might be starting to creep up.  While the
system does what it was designed to it is likely to have limited
connectivity with other systems, is probably not easily expandible, and it's
nearly impossible to implement optimizing/supervisory control.

Ken Haywood wrote:
where is the return? The justification comes when you walk into your boss'
office and say you want to spend $2.3 million to replace the existing
control system. The boss will say "Show me the money." Are you making
production targets? Yes? Will the $2.3 million be paid back in 12 months?
Maybe? How much more money can we make with this upgrade? Dunno? I have lots
of customers still running control systems vastly older than I/A who are
still waiting for the justification to rip it all out.

Alan Armour wrote:
Another issue you can ask is how comfortable people are trusting their
assets to 20 year old (or whatever) electronics.

Brad Wilson wrote:
My two cents ... our corporate policy (as I read it) seems to be toward a
common platform in order to take advantage of shared services.  As long as
we're repairing, replacing or upgrading individual pieces, the cost pretty
much falls "under the radar".  I expect that as soon as I put in a project
with a price tag that triggers a higher level review, they will push VERY
hard to replace our I/A with H******** (which has a huge majority in the
installed base corporate-wide).

Chad Airhart wrote:
According to the data we could gather on our system our FBM's are our most
critical equipment reliability-wise since there can be an economic impact
from a failure.
Our calculations showed we can expect an average of about one plant outage
per two years, one day in duration.  This failure rate would not justify an
upgrade and none of our parts are out of lifetime yet.

Nick Maher wrote:
I'm quite sure Foxboro are addressing these issues with the new equipment of
on-line upgrades, but I just hope it is seamless and on-line in the future,
otherwise the decision to migrate to alternative platforms may be enhanced.

Tom Lemieux wrote:
Here is another side to the Microsoft dilemma they know that PC's are
getting cheaper/faster all the time, the control system is looking more like
a PC but not getting any cheaper. Couple that with the fact that the most
reliable control system we will ever have is the one we've already got makes
it a hard sell. 

Andre Marchal wrote:
In our plant, the benefits of MESH, connectivity and other technological
prowess is a hard sell.
 
 
_______________________________________________________________________
This mailing list is neither sponsored nor endorsed by Invensys Process
Systems (formerly The Foxboro Company). Use the info you obtain here at
your own risks. Read http://www.thecassandraproject.org/disclaimer.html
 
foxboro mailing list:             //www.freelists.org/list/foxboro
to subscribe:         mailto:foxboro-request@xxxxxxxxxxxxx?subject=join
to unsubscribe:      mailto:foxboro-request@xxxxxxxxxxxxx?subject=leave
 

Other related posts: