The facility is a Cogen Power Plant with 2 Alstom Gas Turbines, 2 HRSGs and 1 Alstom Steam Turbine. The new DCS would be used to control all turbines, HRSGS, switchgear, MCCs, Water Treatment Plant and all balance of plant equipment. Our DCS has been obsolete for years and needs to be replaced. I'm considering Emerson Ovation or GE Mark VIe. I am looking for positive and negative comments about these 2 systems from engineers, operators and technicians. Any comments you can share will be very valuable to me. I thank you in advance for your experiences and guidance in helping me make this decision.
Please also comment on the HMIs that go with either of these systems.
There is a GE NEXUS DCS that is much cheaper than Mark VIe system and very robust. If you could leave your office e-mail, i could send the technical datasheet and other information.
I don't know enough about the merits of either system to comment on whether one is better than the other for a specific application.
What I will say is that the quality of the integrator is probably far more important than what control system is selected.
You would never use a SIL rated safety system like the Mark VIe as a typical DCS. It is too expensive and complicated for general control.
We used a Fisher PROVOX DCS, since upgraded to Honeywell Experion, for regulatory control, monitoring and discrete control and Triconex Tricon TMR SIS systems for boiler BMS and turbine control.
I also believe that the Mark VIe does not have all the SCADA and OPC drivers that new more open DCS systems have.
I could be completely wrong but you will only know by evaluating the systems based on price, capability, expandability, redundancy and interoperability, and the cost for all the options.
We have Mark V Control systems on our turbines and have considered upgrading to all of the newer versions that GE offers. We have also considered other systems such as the Emerson Ovation system. The problem with GE's upgrade is that it is not an open control system. Many people we have spoken with who have the VIe do not like this feature. I was at a conference where GE said basically there are only a few people in GE who have access into all of the controls on the VIe. So if you get or have an issue with the black box logic it will take some time to get resolved.
We decided to shelf the control system upgrade for a couple of years.
CRSwanson, I have held off commenting since I like CSA have some strong opinions that I don't want to taint my comments. That being said here goes.
I have worked quite a lot on the Emerson Ovations systems, the older WDPF, GE MKVI, MKV and MKVE, GE Fanuc PLC's, Allen Bradley controllers, and some smaller ABB products. I had the privilege of being part of a build and commissioning of a combined cycle plant using an Ovation system for the DCS and GE MKVI for gas turbine control. Steam turbine control was ABB.
The Ovation platform is a spawn from Emerson's purchase of the older Westinghouse WDPF controls platform. The Ovation hardware is very good. It is not meant for use outside of areas that are not climate controlled. In these areas I typically see some brand of PLC's in use. The Ovation programming atmosphere is very simple windows based point, click, no brainer that my 12 year old could figure out. In my experience it is one of the cleanest and nicest looking for SAMA type logic. Emerson has lots of communication drivers that will talk to various other devices fairly easy. They also have SCADA software and modules that can talk DNP3, Modbus, Allen Bradley DF1 etc. Ovation graphics on the other hand are the worst I have ever worked with. The graphics language is I think a hand me down from the old WDPF system. Emerson does not provide a full library of objects like other vendors but instead only provides a custom set of macros and other objects for your project. Building or customizing anything is a nightmare, unless you enjoy programming in source code. Emerson is good at selling services. They have an "Evergreen" program that suggests upgrading workstations every 4-5 years, depending on hardware and how many revisions of Ovation have passed since your initial install. They sell other add-on packages for hardware support, anti-virus and windows patches, and Sureservice support. Their service has degraded exponentially since I started working with them in 2003. They have lost many good older engineers. Their Sureservice support group has been moved out of the country. If you call them with anything other than a simple request expect to be on hold for a long time, or be told they will call you back later when they have an answer. With all this would I still buy an Ovation system for a DCS-YES. Would I suggest it for a turbine control system-NO.
GE MKVIE is a migration from the older MKVI platform. In my understanding the MKVI was originally designed as a purpose built turbine control system. Controllers were arranged with multiple cores for redundancy and protection for overspeed and overtemp. Controllers designed to execute logic at 10-40 millisecond rates for fast control and protection. Hardware was purposely built with devices specifically designed to control hydraulically actuated fuel and steam valves, hardware and software for synchronization, and integration of vibration protection. The MKVI had limitations since its communication to remote hardware was based on older Ethernet communications using coax. The programming atmosphere is SAMA based like the Ovation, but not as nice looking. It is still point, click etc. and has a reasonable amount of macros and other objects for simple programming. GE uses Cimplicity for the graphic GUI system. The graphic objects are easier to use in my opinion compared to Emerson, but CSA might argue with me on this one. GE has some communication drivers, but not as many as Emerson. Cimplicity also has many communication drivers if you are only needing to communicate with devices and display information graphically but not really need to use this information in logic. GE does not do a great job at selling services in my opinion. I do not know if GE is offering any type of contract support like Emerson. They too seem to have a shortage of good engineers and field service staff like Emerson. Getting support is a challenge. They offer packages for anti-virus and windows support. Would I suggest the MKVIE for a DCS-Hmm I'd like to see one. Would I suggest for turbine control-absolutely.
My other thoughts based on some comments from other contributers and personal experience.
Control system cost-I constantly hear arguments for cost. Using a PLC for turbine control has been done and can be done well, but it can also be done very poorly. If you are basing purchasing a control system solely on cost, then I wish you luck.
Using a system like the MKVIE for DCS will cost you more. That is true in my experience. The "MARK" system, previously known as the "Speedtronic" platform was built to control a turbine. Hardware is expensive and will cost more than a PLC or a system designed to be a DCS.
Open control system. I hear this argument a lot, and LAUGH! Emerson and others will promise to replicate your existing control by sending your logic overseas to be reverse engineered. They will give you exactly what you have now that is totally open. If you have simple logic that has worked since the beginning of time, that is what you will have with your new system. But if you have any problems with logic, or have a desire to upgrade logic or parts of the turbine I would suggest that Alstom will probably be less than helpful if you have a Emerson or GE system controlling your Alstom turbine. If you are fortunate enough Emerson or the integrator will have some engineer they hired away from Alstom that might understand your logic and the proper operation of your turbine, or maybe not. The GE black box logic is another point a hear a lot. I am not trying to defend GE, but they built and designed the turbine, they designed the logic and spent time and money to write specific controls. Some of these include model based control, special frequency response algorithms etc. People seem to complain that this logic is not open and deemed proprietary. I don't understand this argument. If you are going to upgrade a control system and want to use the same logic you have in use today that is open, then make this part of the spec. to the supplier that your existing logic remain open and unlocked. If you are upgrading a turbine control system and want additional logic that brings new features like new combustion algorithms or frequency response then expect that some of this logic might be locked up. In my experience Emerson is great at copying other vendors logic. Ask them to write you new code for frequency response or model based control or whatever you want, and see what they say. I just personally don't think its fair to expect a company to invest money writing special logic that is open, so another vendor can copy it and offer it as their own.
In my perfect world a plant would be controlled by one central system. It would eliminate all the headache of getting different devices and systems to communicate. Minimize spare parts and personnel training. Hardware would be reliable and suited for extreme temperatures and vibration etc. The supplier would have knowledgeable staff that understands the system, and operation of your specific turbines. But I have not yet found this plant. I still don't think that one system exists that fits as a DCS, specific turbine control system, and controls field devices in high temperature and high vibration areas, or areas with specific requirements for atmosphere or otherwise.
I think I better step off the soapbox for now. I hope this helps.
Thank you for your comments. It is always good to hear more than one opinion.
I get the impression that the upgrade from Mark VI to VIe was to make the Mark control system more adaptable to total plant control, not just GT control. Is that true?
Do you know how long the Mark control system has been used for total plant control? Example: GT + HRSG + BPO, or Steam Turbine + Boiler + BOP?
Also, as the Mark control system went through improvements, Mark I, II, III, IV, V and VI how difficult was it to install the upgrades? Were there significant hardware changes that required I/O terminations to change? If it was difficult, do you know if GE has changed their upgrade philosophy?
Thank you and others for any response to these questions.
One of the main benefits of the upgrade from Mark VI to Mark VIe was to allow for dual redundancy configuration of controls for the HRSG and BOP, reserving the TMR controllers for the gas and steam turbines. (TMR could be used for the HRSG and BOP, but dual redundant controls were less expensive and nearly as reliable. The competing control systems from Emerson, ABB, Invensys Foxboro, etc. are all dual redundant.) Another benefit of the Mark VIe is the ability to use remote I/O.
The Mark VI was used for GT + HRSG + BOP beginning with the first H series GE gas turbines and was also used on some projects with 9001FA and 7001FA gas turbines. This was around 1998 to 2005. The Mark VIe came out around 2006 and replaced the Mark VI.
During the Mark V era of turbine controls, the HRSG and BOP control was mainly done with Foxboro IA and ABB/Bailey Infi 90 controls. The Mark V turbine controls were TMR architecture microprocessor based controls.
During the latter part of the Mark IV era GE provided (1985 to 1988) the DATATRONIC control system for HRSG and BOP control. The Mark IV turbine controls were triple redundant microprocessor based controllers (not TMR architecture). The BOP DATATRONIC controls were hot backup dual redundant architectures (but with a single I/O processor - nice single failure point!!).
There was no MARK III SPEEDTRONIC gas turbine control - that was to be a minicomputer based system that was obsolete before it ever got into production. There was a Mark III steam turbine control which was made by the Steam Turbine Department of GE (not Drive Systems in Salem, Virginia).
Mark II SPEEDTRONIC gas turbine controls used some relay logic, some TTL logic, and low level integrated circuit analog controls. Later versions of Mark II included limited microprocessor based exhaust temperature control and combustion monitors. Mark II began around 1971 and lasted until around 1982.
Mark I SPEEDTRONIC gas turbine controls used relay logic sequencing and discrete (transistor) solid state analog controls.
I don't know if there were any upgrades from Mark I to Mark II. I don't think it would have made much sense to do that. GE did provide Mark IV Simplex upgrades for the Mark I and Mark II controls.
As far as I/O terminations changing, it would depend on the original configuration of the Mark I or II. The control panel had to be replaced. If the control panel was in a PECC supplied by GE, with plug connectors to the gas turbine, the Mark IV replacement would have been built with mating connectors to minimize the field I/O wiring changes. The interconnections to the Generator controls might have required new terminations.
CRSwanson, I think Otised did a fantastic job of answering most of your questions but here a couple more comments.
"I get the impression that the upgrade from Mark VI to VIe was to make the Mark control system more adaptable to total plant control, not just GT control. Is that true?"
Yes I think this was the reason, but I was never a GE employee so I can't tell you this was their thought process. But the MKVI was based on a European VME chassis standard, that did not really work well as a "Distributed" system. The MKVIE being Ethernet based is more adaptable to being "Distributed". The movement of all newer controllers and hardware from different OEM's seems to be shifting to Ethernet based communication. Previous communication media like Devicenet, Profinet, etc. seem to be shifting to Ethernet.
"Do you know how long the Mark control system has been used
for total plant control? Example: GT + HRSG + BPO, or Steam
Turbine + Boiler + BOP?"
I think Otised covered this perfectly. One other comment here is about using the Mark system for anything other than turbine control. Turbine control is what GE did, and what the Mark system was designed for since GE built lots of gas and steam turbines. Controlling BOP and boiler systems was not their primary focus back then.
"Also, as the Mark control system went through improvements, Mark I, II, III, IV, V and VI how difficult was it to install the upgrades? Were there significant hardware changes that required I/O terminations to change? If it was difficult, do you know if GE has changed their upgrade philosophy?"
This could open up a whole different discussion so I will try to keep it short. I really can't speak to anything older than the MKIV. Since introduction of the "E" platform GE has been working to offer different upgrade options for older systems. I think that this kind of started mainly with the MKV system. It has a large installed base with lots 7FA turbines that were installed prior to the bubble burst of 2000. GE has been offering lots of upgrades for the 7FA platform that maxed out the available I/O and computing power of the MKV. For this reason I think(just my understanding) they offered a MKVE or MKV Life extension that kept original terminal boards in place and installed new processor boards. The original MKV cabinets were very compact and difficult to work on so many do not like this kind of upgrade option. But it provides an upgrade to hardware and software in a quick turn-around and it cheaper than a full rip and replace. GE has this same option available for the MKIV and MKVI now. It comes back to time and money. I think most people given the choice would opt for a full rip and replace for an older MKIV or MKV with perfect drawings, termination lists and cleanly landed wires in cabinets. But bosses and managers look at the bottom line and time offline and opt for the quick and cheap, and I am not saying that's wrong, its just what I see quite often. So does GE have options to upgrade-yes. The "E" platform I think should be around for a while. GE is still learning some things on the robustness of the hardware at this point based on temperature and vibration limits. It seems like they are working more with GE Fanuc to design things so maybe they can learn some from mistakes of each.
Thirty-three thumbs up for MIKEVI!!! He said it MUCH better than the response I have been working on for two days.
Open system means many things to many people. And we do live. in a predominantly capitalist world and people, companies and corporations are allowed to profit and protect their innovations.
GE does themselves a disservice by not documenting their blocks better. If people would take the time to learn to "read"/understand the graphics in the blocks and combine that with any available Block Help information they generally find the application code (logic; sequencing) very easy to understand (it's pretty basic, actually). But by not clearly documenting blocks and turbine- and auxiliary operation they add to the myth of the system being closed (and I'm not referring to EPIC Mark* systems or MBC (Model-Based Control) Mark* systems). It would even help their field service personnel to learn, commission and troubleshoot units. But that costs money with no measurable monetary return in the present quarter (three-month period).
The single thing that causes most technicians and plant managers to get upset about so-called "closed systems" is that when a "problem" arises they have ill-formed opinions (I started to write preconceptions--but that presumes some previous consideration of the situation and that usually doesn't occur) about what THEY THINK the control system should be doing or not doing that is not based on any understanding of the turbine and its systems. And when they are aware of what the application code that has been running, sometimes for decades, was doing they insist, "Where is that written?" as if some written documents in a manual should supersede the long-running application code. Or, "It shouldn't do that! Should it?" Or, my favorite,"It never did THAT before you put that Mark VIe in there!"
That's all perception--and nothing more. It has nothing to do with open-ness or closed-ness, just perception And lack of knowledge and experience and training and personal ambition and drive and motivation. Even though the units have graphics they aren't video games--which oftentimes are not intuitive especially when one is first starting gaming. But, if one is motivated and ambitious and willing to put some effort into the game it--and many others--all of a sudden become more intuitive. Like everything. But it has nothing to do with being open or closed. And, again, that means many different things to many different people. Often not the same thing to two people discussing it.
Purpose-built, in this discussion, means built for the purpose of controlling, monitoring and protecting turbines using the types of I/O used for controlling, monitoring and protecting turbines. No special converters or interfaces that add to the wiring and complexity. And built-in methods for implementing the designers control and protection philosophies and redundancy schemes--all to improve reliability and availability. Does that cost some money? Yes. Can the same--or, rather, similar things be done with PLCs/PACs? Maybe--but usually not without proprietary cards (I refer to Solar gas turbine TurboTronic control systems which use A-B/Rockwell PLC platforms with proprietary cards developed by or for Solar to interface turbine control and protection devices to their turbines).
Everyone's ideal plant control system is a single system for all equipment in the plant. That can be done with Mark VIe and Mark VIe $IL components. It's designed specifically for power plants and auxiliaries. The problem will be that each gas turbine and steam turbine will have its own control system, and the balance of plant will have its own "(DCS") control system. And the gas turbines will be programmed by one group with their own nomenclature and methods, and the steam turbine by another group with its own nomenclature and methosds, and the DCS by another group with its own nomenclature and methods. And the integration will probably be less than ideal.
Now, all the turbines and the balance of plant could be done by a single Mark VIe and Mark VIe $IL components if desired or necessary, but that's so rarely done by GE that every one is a one-off, and the different turbines and balance of plant bits would still be programmed by different groups with different nomenclature and methods.
But, you're going to find the same from any control system supplier that claims to be able to supply a single control system for the entire plant--it's NOT peculiar to GE. The world is just too specialized today, and every group thinks their way is the right way, no one really wants to risk changing to conform to a single philosophy and nomenclature and methodology.
Now, that's not to say it hasn't been tried or "done"--but in my experience the three places I've know of where they've used a single control system for multiple turbines and generators and balance of plant, it has taken years to get working smoothly and correctly. And the control system supplier/integrator has usually either had to walk away from an incomplete project or was asked not to return, with the bulk of the remaining integration being done by the plant technicians. After a princely sum was paid for the control system and engineering to the control system integrator/supplier and years after the project was supposed to be complete. I know of one large multi-steam turbine coal-fired power plant that has been working for eight years to get all the plant equipment to run in automatic after finding serious deficiencies in the control schemes. And it's still not fully automatic.
The choice of any control system for any application of this magnitude should be based primarily on the knowledge and experience of the control system integrator's application engineers, their proven track record verified by talking to references (plants where they supplied the same or similar equipment) and their after-warranty service and support. The commissioning record for previous similar plants should also be considered heavily. Cost is important, and "ease of use" is important--but just about any programmable control today can be made to do just about anything--if the integrator/supplier understands the application and has knowledgeable, experienced people to complete the project. Full stop. Period. End of statement.
It's all about the knowledge and experience of the people configuring, programming and commissioning whatever control system is being used. If it works--it will seem "open" and intuitive and simple, even if it has idiosyncrasies and some warts regardless of the control system being used. But if it doesn't work and it's found to be a hodge-podge of different programming philosophies and methodologies then it's going to seem closed and foreign and difficult.
Technicians at the power plants that have "tried" a single control system for all the equipment will all say they like it better than multiple different control systems, but it is/was hell getting to whatever point in the project they are at. No question--a single platform is ideal, but it can be implemented in many not so integrated ways--no matter what control system is used.
The choice should be primarily based on the knowledge, experience and proven track record of the control system integrator and secondarily on the control system hardware/platform. Get and check references from any supplier/integrator you will be seriously considering. It's not easy, but it's better to learn about past execution records before a project starts than after. And it seems more people find out after than before when a couple of emails and some phone calls would have been very informative.
And that's that.
MIKEVI provided some excellent and informative information about certain systems/platforms, as well as some great thoughts on "open systems." Thanks again, MIKEVI!
I'm climbing down from my soapbox now and signing off this thread.