Diffrence in between PLC and DCS

Hi.



First of all, sorry for my lack of some basic knowledge.

Actually, I have worked with AB SLCs and never worked with DCS. That is the reason that I don’t know about the difference between SLC/PLC and DCS. Can any one explain me in detail or direct me through a useful link?

Thanks in advance.

http://en.wikipedia.org/wiki/Programmable_logic_controller

http://en.wikipedia.org/wiki/Distributed_control_system

PLC's are designed and optomized for systems with lots of discrete input and outputs and programed with boolean logic disguised as ladder logic. Not so good at analogue or commuications. Usually have to add HMI.

DCS's are designed and optomized for analogue input/output signals and PID control. Not so good at Discrete control. Very good at analogue and commuications. HMI usually built in. Control functions are precanned and the control scheme is put together by linking the control functions.

The larger the size of the system the more cost effective dcs's become.

The differences between PLC's and DCS's are becoming smaller every year.

Difference

PLC for "Programmable Logic Controller". Historically a PLC forte was in discrete control of manufacturing processes. Most of the inputs and outputs for discrete control are binary, meaning they have only two states: on and off. Like a switch. Little processing power is needed for computing on/off signals so PLC tended to be very fast and are used in machine tool and other industries. The terminology "PLC" is interesting in itself. When they started becoming popular in the 1970s they were often called "relay replacers" since the logic for on/off type operations was done with relays arranged in a digital logic format. It was thought that calling them a computer would turn off many electricians and would scare people away from buying them. Manufacturers of the most popular PLCs include Allen-Bradley, Siemens, Modicon and a bunch of others.

DCS stands for "Distributed Control System". DCS's were designed to control processes, not discrete operations. As such, a large number of the inputs and outputs are analog like a 4-20mA signal or 0-10V signal. These signals generally represent temperature, flow, pressure, pH, conductivity or some other process variable. Complex algorithms are programmed into the DCS to provide accurate control of thing like food and medicine manufacturing. Speed wasn't as critical as with the PLCs but accuracy and complexity was. Typical manufacturers of DCS systems include Moore Products (now Siemens), Foxboro, Triconex, Leeds & Northrup (defunct), and many others. The "distributed" part of the name is indicating that the control functions are spread out amoung many different hardware parts. Using one big mainframe computer to control a plant was attempted back in the '60s but it soon became clear that one bug, hang-up, or failure could cause the whole plant to shutdown. Distributed control gives a safety margin!

Of course, since the early 90's, microprocessors, memory, speed and everything else about computers has increased at a fantasitic rate. The distinct line between PLC and DCS is blurring a little and it is not usual to find a PLC performing simple to moderately hard process control functions. Analog cards are readily available and software is available also. DCS can also perform discrete manufacturing sequences but their higher cost and complexity make it improbable that someone would choose a DSC to control a lathe or grinder. But if just a couple discrete control steps are needed at a process plant (on/off control for a tank level) than it can be used.

if all you need are "basic ideas" then the link below might be useful ...

http://www.plctalk.net/qanda/showthread.php?postid=12395#post12395

this has helped quite a few people come to grips with the common PLC/DCS/SCADA/HMI alphabet soup problem ...

I'm not sure I agree totally with the 'lots of little controllers' concept of a DCS in the more modern systems. All the ones I work on have fewer actual controllers than a PLC system. They support more I/O, peer-to-peer and redundant (usually propriatory) comms.
They can generally do the work of 5 or 6 PLCs, given they are also about 10x the price!
The money does buy you redundancy though - redundant power, I/O comms, control comms, I/O cards and dual if not quad-redundant (with the new Quadlog stuff) controllers.

As Ron said, designed to be used where you absolutely CANNOT shut the process down.

Operator stations are usually pretty run of the mill Dell boxes, 2 of the 3 most popular modern systems don't run server/client architectures but just have the individual stations interface directly with the relevant controllers.

Another thing that isnt mentioned is how a DCS manages different operators views of the process. The software allows operators to have control over and see only what they need to. This makes managing BIG systems much easier than with a PLC/SCADA system.
One other note is that in a DCS system ALL the equipment is provided by the same manufacturer and purpose designed to be compatible. In a PLC/SCADA system there are usually some compromises or 'black boxes' in it to make everything talk.

I'm sure this post will cause some discussion but I spend half my time on DCS systems and would like folks to know what theyre getting into!

Our DCS system was designed with several primary/slave "little" controllers for redundancy with seamless failover. In addition, online changes can be performed to the backup processor while the process is still running, and added to the primary module without interruption. The entire plant comm loop between the nodes is redundant propietary cabling (adding to the "10x" the cost) In addition, the function blocks for PIDs and for the controllers are VERY powerful despite being decades old.

DCS vs PLC

There're many differences between PLC and DCS and I would like to go deeper in what Herm wrote. It is true that the DCS are optimized to work with analogic signals such as 4-20 mA or 1-5 V, which is archevied by the complex algorithms within the controller and HMI. For example the DCS controller has a PID block with all the parameters you might need, and is attached to the HMI. If you want to do the same thing on a PLC you need the PID block, many digitals and analogics to the auto/man, modes (P, PI, PID, PD), simulation, constants and integrate those signals into the HMI, just to make the PLC work in the same way as a DCS works. All of this parameters is emmbedded in the PID algorithm on the DCS.
The main practical reason why a DCS works with analogics and PLC works with digitals is the execution schedule of the program. On the DCS the execution is deterministic, which means you set a constant time for the execution of the program. The cicle on the DCS goes much as on the PLC, first communications and diagnostics, read inputs, run logic, write outputs, but in the DCS this time is constant, on the PLC is not. For example if you set a execution time of 1 second, and all the program is done in 0.25 seconds you have 0.75 of NO OPERATIONS or IDLE time. This is important when you do analogic control, but you cant afford this in digital control when for example you are protecting machinery and if the lubrication goes off you dont what to wait 0.75 seconds to modify the process. So the execution time on the PLC is probabilistic so you want to work ASAP. Time execution on DCS are aprox 0.25, 1, 2 seconds, but if you have a PLC programs that goes on 100mS must be a very large and complex program. Remenber that if you are controlling a temperature loop 1 second executions is usually very fast.
Now lets talk about the differences in communications. On the DCS, comms are also deterministic or time stamped. For example if you want to pass a value from the controller to another controller or the HMI, each controller sends the value with a time mark on it. This is very important on flow totalizers and many analogic applications. Older sistems with token topologies take advantage of this and give every station on the network a constant time to transmit the values. But on some ocations some of the stations or nodes on the network can interrupt the flow of non critical data to transmit critical data.
This is why it took so long for many DCS to work with ETHERNET. Now the newer systems use time stamped communications and allow a huge bandwidth (100Mbps or 1Gbps).
This breach between this to systems are closing, since you can buy some PLCs that can work on deterministic execution. Also the boards are the same. For example the Process Logix (DCS) and Control Logix (PLC) shared many of the I/O boards. Also the ABB PLC is very similar to the DCS on hardware.

SLC_Integrator said:

As Ron said, designed to be used where you absolutely CANNOT shut the process down.

Click to expand...

That's my understanding, too, based on what was described as the primary distinguishing characteristic of a DCS as being its ability to accept a downloaded changed/updated program into whatever is considered the program memory and then being able switch from the old program to the new program in one smooth, contiguous (transparent) scan cycle, while maintaining the outputs from the previous cycle.

I think that is what jroachell is saying:

jroachell said:

"In addition, online changes can be performed to the backup processor while the process is still running, and added to the primary module without interruption.

Click to expand...

The recent class of hybrid controllers (rack stuff that looks like a PLC but is designed first for process, 2nd for discrete/sequence) cannot do that. On hybrids, a program change takes less than a minute, but requires putting the process in a safe state (manual mode), downloading the program change and then allowing the controller, to start-up with whatever initialization process is inherent in that controller's design. It's that less-than-a-minute period that to many continuous processes (refineries, steel mills, chemicals, paper) would require huge amounts of effort and time to put into safe mode and restart. These are the industries who use and DCS's.

SLCIntegrator and jroachell, with your hands-on experience, can you confirm whether I understood this correctly? Can a DCS, in fact, make next-cycle program changes without the delays of upload/initialization/etc?

As a corollary, do the high end PLCs, Controllogix, for example, those that are now 'competing' with the low end DCS, do they shut down for program changes or accept a program and switchover in a smooth manner?

Dan

A DCS will accept edits of pretty much anything without having to shut the process down. There are some caveats though. For example with a Honeywell system a loop has to be in manual to make any changes to the controller. I think this is just a safeguard to ensure the loop is in a known position while changes are being made. Also, some points such as logic blocks need to be 'inactive' to edit the actual logic blocks. This is only a problem for that particular logic block and the attached equipment. In defence of this system, the DCS is smart enough to 'back initialise' any controls attached to these inactive points and put them in a safe state so nothing untoward happens.


One big advantage a DCS has is the ability to actually upgrade the formware/software on the fly.
I've done several 'upgrades' on a live running system without any kind of process interruption. While it can get rather 'exciting' if anything doesnt go quite to plan it is a fully functional feature of any good DCS system and one that I have never seen in any PLC system.

Some PLCs such as the Honeywell PLC HC900 (which they claim to be HIBRID) can accept small changes online, but the scan time is longer during the change.

On the DCS there're usually a lot IDLE time during which the controller applies the changes. No need to disturb the process.

DCS & PlC are marged

Maybe a good time for a voting pole PLC vs DCS?

Wait make it interesting PLC vs DCS vs SoftPLC?