DCS vs PLC

K

Thread Starter

Krishna K. Naidu

Can anybody provide me with information on DCS vs PLC?
Our company has placed order with DCS vendor for revamping hardwired Gas turbine controls by the same vendor. Now management wants to cancel the order and go in for PLC based control by other vendor. I have to convince the management about their wrong decision. I have to convince them by providing advantages of DCS over PLC. Please help me.
 
hai,

i will give the basic difference between the PLC and DCS.

both are used in their own way of controlling action.

PLC is used in the Situations were the SPEED Of OPERATION is an important factor.

DCS is used to control the HUGE PLANT with certain Speed but it can handle more complex loops and handle large Inputs and Outputs. As the Pictorial Reprentation of the Entire plant is provided it is considered as the Advanced version of the Industrial Control System.

so the user has to decide weather to use DCS or the PLC Can do the Job.
 
You will not get any help from me (sorry), PLC are 10 times more reliable, faster, safer, etc. In my opinion your company has made the correct choice.

There are many people that can troubleshoot a PLC, your staff can view the I/O and troubleshoot the system... without waking you up at 3:00am.

I have several DCS/VB systems... and 100 PLCs (all types), I would stick with the PLC... Just hope that they are specifying a Allen Bradley or Siemens.

Regards,
 
C

Chris Jennings

Why do you need to convince them it is the wrong decision? If you don't even have enough reasons to state a case for keeping the DCS system then what makes you think they are better?

DCS, PLC whatever it doesn't matter. I would be more concerned about performance guarentees and vendor support and after sales service. That is what will make the most difference. Think about what you want from the project and then compare that to what each vendor is offering. You might think you know the answer but when you put the cold hard facts on the table you might be surprised.

If you use facts and figures to prove your point you will have a good case to put before your management. We can't do it for you, you have to.

Chris Jennings
 
It's not really clear what you're describing, but, if you are saying that your company's management wants to use a PLC for gas turbine control, there are many reasons not to do so.

DCS vs. PLC for turbine control? DCS. Why? More, better hardware for inputs and outputs, which leads to less complicated control systems, fewer imaginative and creative work-arounds, and a more robust and reliable control system. Many DCSs are used in petro-chemical and refinery applications for plant-wide control--PLCs aren't that commonly used in such plant-wide control applications, probably because of the I/O requirements (in addition to some redundancy limitations, as well).

Gas turbines generally have very "unique" inputs and outputs and protective requirements. Passive and/or active speed pick-ups; electro-hydraulic servo-valves; fuel valve and inlet guide vane position feedback (usually the OEMs use LVDTs, not 4-20 mA rotary devices); these make up some of the list of inputs which must be sampled at higher intervals than many PLCs are capable of, or require a bipolar milliamp output instead of unipolar milliamp- or voltage output, or require high-frequency (3 kHz) AC (7 VAC RMS) excitation and differential AC feedback comparators/converters which aren't typical for the majority of PLC applications. This author has seen some "interesting" work-arounds done by some very intelligent PLC application engineers and programmers--which make the PLC program very difficult to understand by most technicians and contribute to varying degress of control and stability based on the operating mode of the unit (start-up, acceleration, rated speed, and cooldown operation).

The protective schemes used by most turbine-generator OEMs don't adapt well to most PLCs and also require some "creative" wiring and interconnection--which, again, leads to difficulties understanding what the intent of the scheme is and how to troubleshoot it.

And, then there's the implementation of the control schemes and logic in the PLC. This author has also seen some incredibly obtuse PLC programming, which is barely functional at best, and which is unintelligible to everyone but the person(s) who wrote the code (and it was obvious that more than two or three people worked on trying to get the functions to work!). PLC programmers, while they may be very familiar with the commands and blocks and functions available in the PLC, don't GENERALLY have the turbine-generator operating and functional experience necessary to integrate algorithms and states seamlessly in the PLC to control and protect a turbine adequately.

For example, this author has recently been asked by a municipal utility which purchased some refurbished heavy-duty gas turbine-generators equipped with GE-Fanuc 90-70 PLC controls to assist with reliability and control issues. They have had several PLC programmers through their facility, all of whom have commented on the complexity of the application and have been unable to resolve all but one of the nine major issues plaguing the units. (There are many more minor issues, most of which appear while trying to understand one function or another!) They have sent their personnel to PLC classes, and they still have difficulty understanding what's happening and why.

Now, they're hiring an experienced turbine professional and are willing to pay that person to learn PLC programming in order to solve their control issues. (The units have been in peaking service for some time, but have always been unreliable and problematic. A change in their operating status and contract requires a high degree of availability and reliability, and the units are being operated more than they have been in the past. The control system integrator who supplied the packages hasn't been able to provide much help over the years, and seems to have had a high personnel turnover....)

Because the speed measuring cards used in the PLC can't see passive frequency inputs below a certain level, the system is programmed to switch between active and passive pick-ups at a certain speed. (The active pick-ups require an excitation source, and mean there must be two types of spare pick-ups in the warehouse.) This causes problems during acceleration and deceleration at times.

The LVDTs are excited and have their feedback converted to a 4-20 mA signal by separate modules, which must be powered seperately from the PLC and which require the LVDTs to be re-wired because of the low excitation voltage of the modules. So, the feedback from the LVDTs is low quality, and that signal is converted to a 4-20 mA signal which isn't of very high quality either.

The electro-hydraulic servo-valves used on the fuel valve actuators and the inlet guide vane actuator require a bipolar milliamp source, and the PLC isn't capable of providing this, so a special card which has a bipolar DC voltage output had to be designed (non-standard, proprietary design).

And, the logic in the PLC...to print it requires roughly 500 sheets of paper! Droop Speed Control doesn't really work properly, and if the ambient is below approximately 70 deg F the unit will not reach CPD-biased exhaust temperature control (it will hit the Droop speed setpoint of 5% which is just below the exhaust temp limit) and so it's power output is artificially limited. The units could make more power, but the control scheme doesn't allow for gas turbine operational characteristics!

And the loading rate adjustment..., because of the way the ladder diagram blocks are sequenced, well, a 2 MW/minute loading rate is more like 5.5 MW/minute. And to achieve any kind of known, verifiable loading rate requires adjustment and testing, more adjustment and more testing, and more adjustment and still more testing, and on, and on, and on.... We got close to what they wanted, and then the scale's maximum limit prevented any further adjustment. We had to change the block and signal type, compile the program, download, and reboot the PLC--none of which could be done until the unit was cool enough to take off hydraulic ratchet operation because when the PLC is cycled all the pumps stop for about one minute, and then nearly every pump and fan starts and runs for about 20 seconds once the PLC's outputs are re-enabled until the program determines what state the turbine is in (or the operator selects Cooldown ON or initiates a STOP/OFF command sequence).

But the BIGGEST problem with the whole scheme is: Nobody can understand the PLC program, portions of which closely resemble the relay ladder diagram logic used by the OEM, and portions of which are just completely baffling and take hours to sort out how they work or why they don't work. And, while the original control company shall remain nameless in this post, they have been around for many years and have done a number of heavy-duty turbine control system retrofits in that time using PLCs (they have been named in other turbine control system-related threads on control.com). It just seems at that time (or at this time) that they lacked (or lack) the turbine knowledge and experience necessary to make these common heavy-duty gas turbine-generators operate reliably and stably using the PLC.

Now, your company may get "lucky" and the control company proposing the PLC may have a very turbine-savvy individual (or individuals) who can duplicate logic and control schemes so that the unit operates almost exactly like it did with the OEM control system. But, the chances of that are not great, since there just aren't that many (heavy-duty) gas turbines controlled by PLCs. And, the limitations of most PLCs with regard to hardware will also contribute to some "creative" and "interesting" schemes, which, unless they are will documented, will be challenging for most to understand and troubleshoot.

This author isn't even a great fan of using DCSs to control turbine-generators, but there are a couple of manufacturers who have the hardware and the application expertise to build a control system that is well-suited to turbine control applications and easy to understand and troubleshoot (Triconex comes immediately to mind).

Unfortunately, one of the most well-known heavy-duty turbine-generator manufacturers is having some difficulty in performing turbine control system retrofits in a timely and cost-effective manner, and they've "locked up" the sale of another governor control system manufacturer's panels which could be applied to their turbines. So, that leaves many owners with little choice if they don't want to use the OEM turbine control systems. There is a company in Florida, USA, Turbine Diagnostic Systems (www.turbinedoctor.com), which has designed and produced a dual-redundant turbine control system (the TurboNet DASH1), and has at least two of them in service as of this writing. So, there are other options which are coming out slowly.

Another reason for opting for a non PLC-based control system is that most every one of the PLC-based turbine control systems this author has ever seen have been "one-off" implementations, meaning that the turbine-generator owner/operator has to spend a good deal of money getting someone trained on the system, and they're not easily replaceable. While the same could be said for a DCS applied to a turbine control application (there aren't a lot of them), there are more people familiar with DCS applications of all types (usually power- or petro-chemical plants or refineries) than there are PLC people who are familiar with turbine-generator control.

In the final analysis, one wouldn't use a turbine control system in a cardboard box manufacturing plant, or a carpet-making factory, or a waste-water treatment plant, or a paint manufacturing facility, or an automobile manufacturing plant; why would one use a PLC for a turbine-generator control system?

markvguy
 
M

Michael Griffin

In reply to markvguy - I don't have any comments on the main thrust of your reply, but the problems with the LVDT signal conditioners caught my attention.

A general purpose LVDT signal conditioner should have an adjustable excitation level set via jumpers and/or potentiometers. Typically the signal conditioners are shipped from the manufacturer with the excitation set to a very low level. Upon installation, someone needs to read the data sheets for the LVDT and signal conditioner and set the appropriate excitation level using a voltmeter. This is a completely separate issue from calibration (but must be done before calibration).

If this isn't done, then it is very unlikely that the signal conditioner will be set up properly for that model of LVDT (unless the two were sold as a matched set - and even then caveat emptor). You'll still get an output, but the accuracy won't meet specification ("poor quality signal"). You typically won't be able to determine where the problem is without putting a voltmeter into the right points in the circuit. This isn't an unusual occurance in custom ("one-off") control systems projects, so I would not be at all surprised if that was the case here.

I said at the beginning of this that I wouldn't comment on the main thrust of your message, but I will add this point. It sounds as if the real question for Mr. Naidu should not be DCS versus PLC, but rather which of the two vendors in question (if either) has enough knowledge and experience to do a proper turbine control retro-fit. If either has prior experience, they should have an existing pre-engineered "package" which can be evaluated and also customer references which can be reviewed. If neither of them has any prior experience in this field, then this project will be most probably be a disaster regardless of what hardware they use.
 
D

DAVE FERGUSON

I agree... and have sat by listening to this conversation long enough.

Get out of 1990 and wake up. There is little difference from the hardware platforms these days. I have DeltaV and Controllogix systems and they are virtually the same as far as hardware costs and functions are concerned. They both have big plusses and minuses. For discrete control the Clogix is over 50 times faster, for analog the DeltaV has advanced algorithms that blow away the Clogix, for motion control, hands down Clogix, for plain old PID, wash...

For people who can work on it, Clogix all the way and on and on. The HMI side is also a wash with OPC now a days. Emerson is an old intelution IFix HMI (although now they have bought out the source sode fo their offering), RSView, Wonderware etc... just a paint set.

Networking, both, DeltaV seams more integrated but it is really just another "IT" system. I have all the AB stuff networked together etc.

I know a bunch of you will throw out all the reasons why this is wrong, but from experience not true. The PLC came from "lowly" discrete world controlled by Electricians and the DCS came from "upper board room", Engineering/IT world and there are still these old politically incorrect viewings of them by people who throw weight around with little substance behind their reasons (sort of like the tone of the original post).

When in reality, they have met in the middle and are both viable solutions depending on what you are doing with them. About half of the "DCS" systems I have seen are just doing plain old PID etc and could just as easily been "PLC". But someone sold a bill of goods that the "DCS" was better. I enclose those terms in quotes because they are old world terms and some new term should be created as it is a wash today more or less.

Welcome to 2006.

Dave
 
Valid points, every one. If the originator of this thread was talking about using a DCS vs. a PLC for plant-wide control of water treatment systems, well-water systems, heat-recovery steam generator (HRSG) drum level control, liquid fuel forwarding and treatment system control, cooling water system control, condenser vacuum control, etc., the line is very blurred these days between DCSs and PLCs.

If this author remembers correctly, one of the original "benefits" of DCS (Distributed Control Systems is what the acronum stands for, it is believed) was that I/O racks, and in some cases CPUs (Central Processing Unit modules) could be remotely located around a plant and connected using a "bus" or even redundant buses to control operation of many inputs and outputs and processes. At that time, PLC (Programmable LOGIC Controllers) weren't so sophisticated, with limited or no analog control capabilites, and no distributed I/O or CPU capabilities.

Nowadays, 2006, many PLCs have remote I/O rack capabilities and much more sophisticated analog control capabilities, including PID and much more, and are much more capable of controlling large plants than they previously were.

But, if the originator of this thread was asking about using a DCS vs. a PLC for control AND PROTECTION of a combustion turbine-generator and its auxiliaries, the hardware capabilities (input- and output handling) must be considered ALONG with the experience and knowledge of application engineers of the company proposing to use one or the other for control AND PROTECTION of a combustion turbine-generator and its auxiliaries.

In this case in the hardware capability arena, DCSs have been handling more varied types of I/O for longer periods of time for control AND PROTECTION of power generation processes (boilers, condensers, feedwater pumps, etc.) than PLCs.

Siemens, for example, is doing their best to use their S7 platform for turbine control applications, developing specialized modules and cards and CPUs just for this application. One could argue that the GE Mk VI SpeedTronic (and its new "enhanced" version) are glorified PLCs (built on the VME platform and using special I/O cards).

But the fact remains, that two of the largest manufacturers of turbine-generators and PLCs are NOT relying on PLC hardware alone for this application--but are using "hybrid" PLCs with specialized I/O cards and, in some cases, CPUs, to control their power generation equipment.

Now the above addresses ONLY THE HARDWARE CAPABILITIES of a control system. As a couple of the contributors have properly pointed out, the turbine-generator application experience is a MAJOR consideration which must be taken into account, also, and this consideration should probably have the majority of the scrutiny and weight. An experienced application engineer or team could be very successful in applying either type of control system to the control AND PROTECTION of a gas turbine-generator and its auxiliaries and provide a solution which is easy to understand, troubleshoot, maintain, and operate.

This is a great discussion, and it would be most helpful if the originator would keep us informed of the status of the project and could provide some more information, as well. He certainly has had some good points--in favor of both control systems--to make a case or to even re-evaluate his own position, as the case may be. Again, we would really like to know how things move forward and end!

markvguy
 
D

DAVE FERGUSON

I agree completely.

The tools used are only a paint set, what one person does with a PAINT set
is paint a house, while another paints the Mona Lisa.

As with ALL programming, it is the painter and his skills.

My point was the lines have really blurred.

Dave
 
Thanks, Mr. Griffin, for the information. The LVDT excitation comes from the same device that scales the feedback into a 4-20 mA signal. Unfortunately, I didn't get the details on the devices on my first visit, but I'm returning later this week and will get more info.

markvguy
 
M

Michael Griffin

This reply may be too late to reach you before you arrive at the customer site, but you might also wish to have a look at what brand the signal conditioner is. If it is a brand I am familiar with, I may be able to point out a few other common areas besides excitation which can go wrong. I have probably seen more of this class of instrumentation installed incorrectly than correctly.
 
Mr. Griffin, the devices are manufactured by Lucas Scientific, which I'm having trouble finding a website for. The excitation voltage magnitude is not adjustable (via pot or berg jumper).

The Customer is using Kavlico LVDTs on the fuel valves.

markvguy
 
V

Valentino Krus

What are you talking about? You are complaining against the whole control society. Keep in mind that you are an engineer not a teen.

Thank u.

regards, V
 
I am in agreement and as a proponent of Allen Bradley systems, the difference is in the planning. all projects come down to the planning of the I/O counts and the processes, timing and configuration of communications and optimizing the data feed back to the the operator control stations.

Plain an simple, a better plan upfront with a defined plan for expansion, redundancy and maintenance procedures will be the difference in success and failure.
 
Top