PDA

View Full Version : Total Fustration - PLC vs DCS


jaichains
October 29th, 2008, 11:16 AM
Guys,

I am in a total fustration condition and i am looking for feedback from everyone to help me out.

We have a PLC based DCS architecture on our oil platform i.e. PLC's are doing the job of DCS like F&G, PSD, ESD, PCS. I know that PLC do a good job in controlling a process but lacks the capability of a true DCS (advance process control like Statistical process control, Maintanance, production optimisation, alarm handling/rationalisation, industrial network security etc etc).

PLC's were never made to take over as DCS, i appreciate that with time they are getting better and better with some advanced control features but are never a good choice to control an entire plant. I need to convince my management to move out of PLC based system into a more robust and capable DCS (Delta V, Honeywell Experion or Yokogawa)

Please can confirm that the above statement that i have made is valid and share your views on this. Thanks in advance.

Paully's5.0
October 29th, 2008, 03:16 PM
....... but are never a good choice to control an entire plant.... I need to convince my management to move out of PLC based system.....

Please can confirm that the above statement that i have made is valid and share your views on this. Thanks in advance.

First, tell us why you believe a PLC system is "never a good choice to control an entire plant". I personally think that is an erroneous statement, but of course you really do need give us your definition of "Plant Control" .

What are you trying to accomplish to justify the cost of swapping your PLC components to DCS components?

Now, if you are talking "building" control systems HVAC/Lighting...you've got a point, but controlling a manufacturing process....

Elaborate please.

CroCop
October 29th, 2008, 03:18 PM
Good luck.
If you can't summarize why, & you're already sold on it, how in the world do you expect anyone else to come to the same conclusion?

danw
October 29th, 2008, 04:55 PM
To me, the big distinction was always that a DCS is centralized and server based.

But my direct contact with DCS is marginal. I know that the sales departments any of the vendors you mention would be more than happy to assist you in developing cogent arguments for a DCS. They're always on the prowl for a 'champion' within an organization.

Give 'em a call and ask to talk to sales.

Dan

jstolaruk
October 29th, 2008, 05:38 PM
The correct statement is that PLCs are fine for controlling a plant and PCs are fine for DCS responsibilities. The two work as a team.

John Gaunt
October 29th, 2008, 06:47 PM
The problem with DCS's is that they are big, expensive, centralised and quickly become outdated. Then you are pretty much locked in to something that gradually becomes hard to maintain, hard to keep people who really want to work on something newer, and as it is then obsolete you don't want to develop it further.

I favour using PLC's and then keeping them fairly small (task related). Use a SCADA system to tie it all together and you really have a DCS equivalent. Eventually you will find you need to change brand of PLC and SCADA you use. Resisting change for a while is good as you don't want to be suporting too many different things but when there is good reason then make the change.

Bob K
October 29th, 2008, 07:02 PM
Would these processes be..
F&G = Fire and Gas Detection System?
PSD = Process Shutdown System? (a guess)
ESD = Emergency Shutdown System?
PCS = Process Control System? (another guess)

If so would any of these processes require SIL ratings or ratings required by your industry or by your company?

I am not familiar with capabilities of a DCS but could you get SIL ratings (or whatever ratings required) with DCS software / hardware???

If anything this would be a first step towards building your case.

Mike6158
October 29th, 2008, 08:57 PM
but are never a good choice to control an entire plant.

Never? Ever? I think you are getting sucked into the hype because the statement that I quoted above is male bovine fecal matter.

The line between DCS and PLC is blurred. The more you know about both systems the blurrier it gets. I know of numerous large plants that are controlled by PLC / WW systems. I can think of a few that are 200mmscfd plus turboexpander plants with full fractionation. The last one that I was involved in had a PLC on each engine, the turbo expander, and fired heaters as well as a "main" plc for process control and motor control all on an ethernet network with (6) Wonderware HMI's (not client server) all polling the 15 PLC's at once. No problems... not one. I built a C1/C2 controller for the DeC1 and it maintains bottom spec perfectly as long as the chromatograph does it's job. If the chromatograph fails or gets too far out of range the C1/C2 controller reverts to straight temperature control and alarms. I have routines built that calculate gas flow based on AGA8 and NX19. I have liquid flow routines that calculate the flow of liquid across an orifice that utilize flowing and 60 degree specific gravity. I built air fuel ratio control (curve fit) routines for some of the engines. I have a surge control routine for the expander that TTI approves of. I've done all of these things in PLC's and I've done them in DCS systems. I do not prefer one over the other. I make money no matter what the customer prefers. But... if I built my own gas plant I would install a PLC based system. It's more cost effective, more robust, and considerably more lean than a DCS system. By lean I mean FAST. IO update times at sub one second. DCS systems have SO MUCH OVERHEAD that you invariably find yourself slowing the scan rate on certain blocks down so that the system doesn't bog down.

The Delta V and Experian DCS systems are nice. However, they are essentially PLC IO with a very FAT top end (HMI). I love the way they program... I hate the way they work.

BTW-

F&G- should be it's own stand alone system. PLC works fine. DCS does not.

PSD- Either works here since so many things trip other things it's best to leave it in one system

ESD- I prefer hard wired ESD systems. Typically I'll drop the power on a PLC output card with an ESD relay as well as "tell" the PLC that the ESD system has tripped so that it can drop the output out as well. Kind of a twofer that you can't really take credit for in a SIL analysis (unles the PLC is safety rated which is gross overkill in a gas plant).

PCS- PLC or DCS... doesn't matter. Each has it's strong suite. Just don't kid yourself... if you "upgrade" to a DCS you WILL NOT save money. It is an EXPENSE. It doesn't create money... too many people sell DCS projects as a way to make money. Mooo...

One more thing- DCS = Distributed Control System. Have you ever pondered what that really means. Here's the deal... way back in the time just after electricity was invented, electronic control was making it's way into the market. Processing power was LESS than what your car has in it (by a long way) so they "distributed" the load over numerous control processors. So, if you uyse multiple PLC processors in a logical manner, networked together to a common HMI or HMI's then guess what... you have a distributed control system

jaichains
October 30th, 2008, 03:49 AM
So, if you uyse multiple PLC processors in a logical manner, networked together to a common HMI or HMI's then guess what... you have a distributed control system

Mike,
This is exactly what we have got on our plant. PLC's based DCS. I also understand that DCS is a generic term meaning just distributed control which we can get using PLC's also. PLC for F&G,ESD is fine. But DCS is much more capable control than PLC. see below:

Ř
Production Scheduling/ Accounting
ŘExpert System/Statistical Process Control
ŘProcess modelling
ŘAlarm Handling/Rationalisation
ŘPredictive Maintenance

JesperMP
October 30th, 2008, 04:57 AM
You want a complete "package" that has all these things. That is understandable and quite reasonable for a large proces application like an oil platform.
It is all software (except the "industrial network security" .. see below), so there should be nothing that impedes that a PLC based system has this functionality. Siemens PCS7 for example uses standard S7 technology to achieve DCS functionality.
The question is if there exists a "canned" DCS solution for the PLC platform that you have on your oil platform.

May we know what PLC it is ?

About "industrial network security", do you mean redundant networking, or access sequrity ?
If you mean redundant networking, then some PLCs supports this. Frankly I am shocked if you have an oil platform without redundant networking.
If you mean access sequrity, then there also exists solutions for this, and as far as the PLCs are concerned, it is transparant to them.

jaichains
October 30th, 2008, 05:43 AM
Jasper,

GE Fanuc 90-70 and 90-30 PLC's are being used on our platform.

Totally agree with you regarding PCS7 which is basically S7-400 PLC based DCS. Thats the reason i never mentioned them in my list above. But s7-400 are very capable PLC's,no doubt about that.

Yes, We have redundant networking. What i mean by industrial network security is access control, Intrusion detection. I agree that this feature is IT related.

tomalbright
October 30th, 2008, 06:17 AM
We have a PLC based DCS architecture on our oil platform i.e. PLC's are doing the job of DCS like F&G, PSD, ESD, PCS. I know that PLC do a good job in controlling a process but lacks the capability of a true DCS (advance process control like Statistical process control, Maintanance, production optimisation, alarm handling/rationalisation, industrial network security etc etc).

This is something I have seen before. PLCs DO provide a stable, consistent form of control, with easy to access parts and materials. I work on oil platforms - I am assuming you are talking about production platforms? All of mine are fairly small, and have from one to five PLCs on them. I have seen (or heard) platforms that have Delta V or something similar on them. In my opinion, that's vast overkill for what can be gathered with a PLC and PC-based data gathering.

Why? The DCS is basically "top heavy" by design - all the controls are in one place, and all of the IO is distributed. This is good in some places, but on a platform you may need several PLCs that can either work seperately, or tegether, as needed. You don't want to take your platform offline for software changes, as this can cost money by loss of production.

"DCS" in the classic sense of a centralized controller isn't a good option for a high production facility. Don't give up on your PLCs!

jaichains
October 30th, 2008, 06:39 AM
Why? The DCS is basically "top heavy" by design - all the controls are in one place, and all of the IO is distributed. This is good in some places, but on a platform you may need several PLCs that can either work seperately, or tegether, as needed. You don't want to take your platform offline for software changes, as this can cost money by loss of production.

"DCS" in the classic sense of a centralized controller isn't a good option for a high production facility. Don't give up on your PLCs!

I disagree, DCS is not centralized control. Like PLC it is distributed but is more capable and comes with a complete packages which is benefical for optimisation, production management for large oil platforms.

Mike6158
October 30th, 2008, 06:49 AM
Mike,
This is exactly what we have got on our plant. PLC's based DCS. I also understand that DCS is a generic term meaning just distributed control which we can get using PLC's also. PLC for F&G,ESD is fine. But DCS is much more capable control than PLC. see below:

Ř
Production Scheduling/ Accounting
ŘExpert System/Statistical Process Control
ŘProcess modelling
ŘAlarm Handling/Rationalisation
ŘPredictive Maintenance




All of the things that you list are doable in a PLC based system with the right PLC/HMI package without all of the overhead clutter and expense that a DCS brings to the table.

Structured Text language allows the programmer almost unlimited calculation power. The S7 allows the programmer to compile the code into reusable function blocks. Allen Bradley offers similar features with their Control Logix line. Unfortunately you can't compile the code into an FB but there are ways to call the routine interatively to do multiple calculations so that's not a hinderance. CLX has numrous function blocks built into the system with more coming all of the time.

manmeetvirdi
October 30th, 2008, 07:02 AM
Questions:
1) What DCS can do and PLC based system can't?
a) DCS have distributed I/O's but PLC's also have them
b) DCS can log various data, but PLC system with SCADA can also do this. That means all the points mentioned in post#9


Have never came across DCS so not in condition to elaborate much on this. But have desire to know what DCS can do which a PLC based system can't do.

surferb
October 30th, 2008, 07:02 AM
I agree with Mike and Jesper here.

The key is to get your terminology straight. Different industries and professionals regard "DCS", "PLCs", and "SCADA" as different things. The lines are converging with Internet technologies. Do TCP/IP enabled PLCs on a geographically separated regions constitute a DCS system? What DCS system are you referring to that's "more capable" than what PLC system?

BTW, your list represents software, not hardware, which is what PLCs and DCS are.

Production Scheduling/ Accounting
ŘExpert System/Statistical Process Control
ŘProcess modelling
ŘAlarm Handling/Rationalisation
ŘPredictive Maintenance

cornbread
October 30th, 2008, 09:08 AM
Biggest advantage of our Foxboro DCS is the redundancy it provides in processors and communications. Basically you need two failures to have an impact on the plant.

manmeetvirdi
October 30th, 2008, 10:10 AM
Biggest advantage of our Foxboro DCS is the redundancy it provides in processors and communications. Basically you need two failures to have an impact on the plant.

Processor and communication redundancy is perfectly achievable in PLC based system too. Example S7 414-H.

harryting
October 30th, 2008, 10:50 AM
Don't we get this type of discussion every year?

I came from a DCS background.

To me, the best analogy is that a DCS is like using Windows and PLC is like doing assembly language. PLC can do everything a DCS can do but do require more planning to have a structure that's easy to support.

It's a lot easier to get into messy logic in PLC while DCS already has a built-in structure in keep everything clean.

yes, DCS is a lot more $$$ but it's worth it in most continuous process environment like a refinery.

Just my 2 cents.

jaichains
October 30th, 2008, 11:48 AM
Don't we get this type of discussion every year?

I came from a DCS background.

To me, the best analogy is that a DCS is like using Windows and PLC is like doing assembly language. PLC can do everything a DCS can do but do require more planning to have a structure that's easy to support.

It's a lot easier to get into messy logic in PLC while DCS already has a built-in structure in keep everything clean.

yes, DCS is a lot more $$$ but it's worth it in most continuous process environment like a refinery.

Just my 2 cents.

Guys,
Your comments have helped me a lot. Thank you.
Ours is a big oil platform on GE Fanuc 90-70 and 90-30 PLC based DCS. Thease PLC's were built in 1990. Looking into the future, if we have to upgrade then it would be better to go for a DCS rather than to another type of PLC (RX7i, RX3i). Please share your views.

manmeetvirdi
November 26th, 2008, 08:52 AM
Attached filed sheds some light on DCS and PLC control altough emphasis is on DCS system.

Paully's5.0
November 26th, 2008, 11:38 AM
Guys,
.........Thease PLC's were built in 1990. Looking into the future, if we have to upgrade then it would be better to go for a DCS rather than to another type of PLC (RX7i, RX3i). Please share your views.

Financially it would not be the better move. IF all you need to do is upgrade the HARDWARE the 9070/9030 programs can be ported to the Rx7i/Rx3i counterparts with little hassle. The PLC Programming software (Proficiy Machine Edition) will convert the existing 9070/9030 project from one platform to another providing you a detailed list of potential issues that may need to be addressed. 95% of the 9070/9030 IO modules can be re-used in the Rx7i/Rx3i systems so you minimize the hardware costs, and you have very minimal programming costs (to correct any conversion issues). You also minimize downtime as swapping the PLC backplanes and modules is a snap, and if there is a problem, put the old PLC back in place and your can live to fight another day.

I'm currently in the process of doing this as the plant cannot be shut down for an exented period for a complete retrofit, and they do not want the cost of a complete retrofit.

If you NEED to re-write code to better optimize your process, if you have an endless budget, and if the system can be down for days/weeks for install/testing of a new DCS system, then hey it may make sense to completely strip out the old PLCs.

But it sounds like your heart is set on a DCS system.