Currently have 3 paper machines under DCS control. One of these items is getting past its use by date and requires updating.
Question is, can a PLC be capable of providing the same level of control options as a start of the art DCS? Is anybody actually using a PLC for this function?
Yes. Depending on your application the terms can be used synonymously. With your application.. paper machine control - I would say yes. If you are looking to upgrade, I recommend you consider a Ethernet PLC. WAGO has a very powerful cost effective ethernet controller which starts at around 395 List and offers a site liscense IEC 61131-3 site liscense programming tool for 600 list. This system is very flexible and modular and offers the complete range of I/O to choose from - digital, analog, special funtion modules. They also offer solutions for DeviceNet, Profibus, MODBUS, etc..
Check it out - once you give it an honest try you will probably never go back to expensive, large, clunky, traditional style PLCs or DCSs.
or call 1-800-346-7245 to speak with technical support engineers about your application for free!
PLCs are very capable of providing the controls for a paper machine -- but it is a very political battle with most companies. I got caught up in the debate at a mill in SC. Years ago the line between DCS and PLC was analog and discreet. DCS could do analog (PID loops) better than PLCs -- while PLCs could do discreet much less expensive than DCS. Nowdays the PLC market has caught up with DCS as far as analog capablities while the DCS is still very limited as to cost compared to I/O count. Add to that the much faster scan times and ease of finding programmers for PLCs it becomes a no-brainer -- unless you make a living working on DCS system. I haven't seen an application where DCS would be a better selection than a PLC. The paper and pharmacutical industries just seem to locked into DCS -- good sales people perhaps?
Perhaps if you understood the control of paper machines your opinion may be different. A PLC is just not capable of paper machine control. It takes more than a PID for controlling this dead-time dominant process with primary measurements that take 30 seconds or more (for cross-directional scanning.)
Richard H. Caro, CEO
2 Beth Circle, Acton, MA 01720
Tel: +1.978.635.9449 Mobile: +.978.764.4728
E-mail: RCaro@CMC.us <mailto:RCaro@CMC.us>
Buy my books: Automation Network Selection
Wireless Networks for Industrial Automation
This is not quite correct. We have been migrating DCS systems to PLC for years now on full paper and tissue machines.
Your statements as to the CD and MD control capabilities of PLCs is false. The new systems (ex. ControlLogix) are more than capable of doing CD and MD controls with deadtime compensation on the basis weight, caliper, color and humidity loops.
Time to look into the capabilities of newer generation of PLCs... the kids are all grown up now!
In my country I am aware of several paper machines controlled by PLC's, the recent trend being to Siemens PCS7. PCS7 may arguably be a DCS however at the end of the day it runs on hardware from the S7400 PLC range, and the difference in programming does not make it more or less capable of tackling
the process control task for the machine.
There is a difference to controlling a paper machine, and fully controlling a paper machine. PLCs are quite capable of controlling tank levels, flow control and drives. However as stated by Dick Caro the QCS (or CD and MD controls) are very much a specialist field that is dependant more on the software than the hardware.
I totally agree that PLCs can be used as part of the basic control system. But high level supervisory controls and QCS controls are still exclusively DCS.
I don't know for whom you work, but there are only 3 principal suppliers of basis weight, moisture, opacity, etc. control to the pulp and paper industry: Honeywell, ABB, and Valmet/Metso. None of them has migrated their control systems to a PLC. If you have done this work using ControlLogix, you have kept the secret rather well. I know that the ProcessLogix software running on the ControlLogix hardware has a number of powerful function blocks, but I would be hard-pressed to implement even a Smith-Predictor algorithm for simple deadtime dominant control on this system. Unless this is a trade secret, why don't you tell us for whom you work, and how you accomplished this without using low-level programming. I assume that your company then takes responsibility for support of the system, since I doubt any paper company would want this responsibility.
AIA Automation Inc, based out of Canada.
Yes there are only three primary suppliers of scanners (but not the only suppliers). Scanners do not necessarily mean QCS though.
Even if the scanner comes from the big three, the control algorithms can be implemented in the PLC. We have done this WITH a scanners from the big three and they do work quite well and can be much more cost effective than going the QCS route.
As for how we do this, you would actually have to pick up a manual and read about the function blocks available in the ControlLogix platform. We do not use some twisted low level code. We use standard FBs available in the controller.
And yes, we can and do sell this code library as part of our projects.
PLC based systems can and do provide any DCS functionality. The only difference is not having everything under one hood. Monolythic DCS suppliers vs. Integrated systems from many specialised sources.
As for machine suppliers not going this route, we don't really care. We provide our systems to the end user... the paper manufacturers.
Michel A. Levesque, eng.
Montreal office director
AIA Automation Inc.
I think it's more to do with the fact that DCS vendors sell a paper machine control system as a package. By doing this the engineering is reduced and therefore the costs. Only some modifications need to be made for the particulars of the machine.
I know that Siemens offer a package for paper machine drives control. But I don't know of any PLC vendors that provide this kind of control system package. The hardware isn't much different, but if you have process experts within the vendor who can help you with the high level controls that is where the big dollars are. It isn't much good if you save $$$ on the hardware is you need to spend $$$$$$ on the software to get the higher level controls you require to optimise your plant.
It would be interesting to know if companies like Kvaerner or Voith offer their process control software for any of the major PLC systems. I know they can do it for all the main players in the DCS field, Yokogawa, ABB, Honeywell, Metso.
I don't know about a paper machine, but we are running a metal lamination line, a metal slitting line, and three blown polyolefin films winders on DCS systems and I wouldn't put a PLC in to replace them. At least not one that doesn't at least have some DCS functionality. A lot of the newer PLC/DCS are becoming more alike so you should have a reasonable selection to look at. We are using Reliance Automax and A-B ControLogix. We have one cast film winder that is running A-B PLC 5, and it does run OK, but sometimes the control leaves a little to be desired. Hope this is helpful.
Does the DCS need to interface with a scanner or QCS? Do you want to integrate the DCS and QCS or is stand-alone alright. If you want to integrate then one of the big paper industry DCS providers will be the only option. Such as ABB, Honeywell or Metso.
We have done portions of many paper machines with PC control (i.e. a soft DCS/PLC). There are also a few complete machines running with server-based (PC) control.
The important decision is not the hardware you select, but the software you choose.
If you select good software, you'll need less hardware to accomplish the tasks necessary. The less hardware required to do the job, the better.
I would tend to stick to the DCS.
The reasons being:
Connectability to QCS.
Connectability to Cross direction controls.
Large dead times associated with paper machine control systems usually require some type of advanced control which is readily available on most DCS platforms.
Most DCS suppliers will offer you different types of proven advanced control solutions for controlling different area's of the process.
DCS's scanning times have improved significantly over the last decade and i can't think of any control loop on a paper machine which could not be run efficiently on this platform.
The main reasons industry uses DCS instead of PLC are the following, in my opinion.
1. DCS does not rely on a server p.c. so all Human Interface Stations can operate independently, increasing the mean time between failures dramtically.
2. All major petrochemical companies use only DCS as they are more capable controlling continuous processes. (Shell, Petronas, Chevron, etc.)
Some DCS suppliers guarantee continiuty of their products, IE all models will be able to connect to it's successor. With PLC, unless you buy the biggest and most powerful, you might have to replace CPUs etc when you expand.
I must say I disagree with the first and third statements. Almost all DCS vendors have moved away from their standalone operator stations that are connected to the process bus to a system almost identical to the standard
PLC with HMI setup. ABB's latest product Industrial IT Process Portal actually has three servers that are all required for just one client. They all run Windows 2000 and the DCS support staff need to be well versed in the
operation of an Microsoft Active Directory systems.
Also the last statement might be true for some vendors, but ABB for example have not made the transition from the old style Bailey Infi90 systems to the new systems easy. They haven't even kept the migration path standard, it
changes every year. Obviously this is a difficult situation because they bought a number of different automation companies and each had their own I/O, Processor, HMI and engineering tools.
If it comes down to PLCs or DCS on Windows, that's a no brainer.
I'd go with the PLCs everytime. And I'm a PC fan. Why make a huge
investment that will be deliberately obsoleted in a year or two?
You would have to replace the OS under the DCS which is more than
a little disruptive and opens lots of opportunities for headaches.
Just about the last thing I'd like to do with a DCS once you get it
all working is screw with it simply because MS declares your OS
obsolete. Or start rolling the dice with Service Packs, etc.
That's simply insane. Bad for your throwaway desktop, truly insane
for an investment that has to work reliably for a long time to recover
Control of a Paper Machine is a very Broad subject. There are at least three 2 x2 control Loops at the wet end alone.
1. Basis Weight and Moisture
2. Headbox Level and Flow
3. Retention aid and consistency
It is impossible to control any of these with a PLC or even DCS if you are looking for Adaptive predictive control. I have developed Adaptive Predictive Control for a Single input single out control configration for Bailey Net 90. I am in the process of developing a C based program for controlling the three controls mentioned above. Unfortunately the knowledge of tradespeople and engineers who have hijacked Control department have limited knowledge of Control not very much beyond PID. Modern control is based on Math and Model of the control Loops. Please let me know what you are after. and I will let you know what to do.