We are upgrading our control systems next year and are looking at 3 different total plant solutions. We would be going from GE MARK VI and Ovation 3.02 to our new choice.
First we have the MARK VIe option. I am not too excited about it because of the black box mentality associated with GE control systems. It would be a simple adjustment and I know it would work though.
The second option would be becoming a full Ovation plant. I like Ovation, but I know that tech support gets expensive, really quick.
The third option is a control system made by my company called ICON. They use Allen Bradley PLCs as the backbone of the system and it comes with free tech support for when my butt is getting kicked.
What do you guys think?
This topic comes up rather frequently on control.com.
I want to preface my remarks with this statement: The "closed," "black box" myths about the GE Mark* operating software is a lot of hooey. It's just different from what most people are accustomed to, but it's not closed (unless you have MBC (Model-Based Control)).
The thing that prevents me from recommending GE Mark VIe is this: You are going to get a LOT more than you think you're going to get. And, GE WILL NOT be able to tell you exactly what the differences is/are. You will only find out when you push the START button for the first time, and then there will be a slow learning curve that impacts reliability and availability.
Why? Because GE believes that as the unit OEM they are OBLIGATED (under threat of lawsuit) to provide the latest and greatest control schemes when the upgrade a turbine control system. And, the salespeople tout this very heavily--you will be getting the latest and greatest to help your turbine and auxiliaries run smoother, jump higher and last longer. BUT, ask them to tell you specifically what those improvements are, and how they will impact your starting times, and you will get the "deer in the headlights" look. And, NOTHING else. You will pay a premium for the equipment and the software, and will be happy to do so. And, you will "learn" about all the improvements when they cause Process Alarms or prevent a START or cause a shutdown or trip.
One example; they will add software to test the gas fuel valves for leaks (if you don't already have it). And, if your machine(s) is(are) older, it's likely the valves may leak--but if your current software doesn't have the leak testing scheme you won't know about it.
This leak testing scheme adds time to the start-up (they won't tell you about that in advance); and if it detects a leak, it will prevent a START. Then what are they going to do? Tell you to fix the valve(s)--which is the right thing to do. But, you need to get your machine started in the same amount of time as before, and you need it to start--reliably. Well, fix the valves, they will say. You will say, "It worked just fine before you put that Mark VIe in there!" And they will say, "It's functionally the same as before, but with all the new protective bells and whistles." And you will say, "Tell me what other "Easter eggs" we can expect?" And they will say, "Well, it's all the standard stuff we put on new units like yours." And another deer in the headlights look will follow. And nothing more.
The other two suppliers will take a look at the existing logic, and they will most likely duplicate (as best as possible given the hardware and software they have) the existing control logic and schemes, and give you just what currently have. Which is what you really want. They may offer some tweaks and upgrades--but they will most likely tell you about them in advance, and let you decide if you want them or not.
The problem with PLC-based control systems is that they aren't really "off-the-shelf" systems. There will always be a couple of special cards (high-speed speed sensing for sure), and likely cards for the servo outputs and LVDT inputs. And something for vibration sensing (though I hear A-B have some good off-the-shelf cards for that these days), and flame sensing. And, there will NOT be the focus on tripping reliability that the Mark* systems have (PTRs (Primary Trip Relays) and ETRs (Emergency Trip Relays)). It will be serviceable and people who know how to read A-B logic will do reasonably well--AS LONG AS THEY UNDERSTAND WHAT THE LOGIC IS SUPPOSED TO BE DOING AND HOW THE TURBINE SHOULD OPERATE.
The last bit is the thing that most people DON'T understand about GE-design heavy duty gas turbines--you can use any control system, or even train monkeys, to control the turbine. But you MUST understand how the turbine and auxiliaries should work to be able to do that--and to troubleshoot it when it doesn't seem to be working correctly. (And mis-perception is a LOT of the problem when it comes to GE-design heavy gas turbine operation--people thing it should do this, or should do that, or shouldn't do this or shouldn't do that, without ever giving it a moment's thought before they make that statement. They just think, "Well, it should [or shouldn't] do that!" And, that becomes reality. And, then they try to solve a problem with looking at the application code (sequencing; logic) once every two or three years, and start complaining that the system is "closed" or it's a "black box." (There aren't many GE field service people who can say how a turbine is or isn't supposed to operate these days (when you're only selling 20 units a year!) so getting help from them is difficult at best.)
The Mark* control systems are GOOD at what they do--SO GOOD that they rarely break down (with reasonable maintenance and attention to Process- and Diagnostic Alarms), but when they do most people have never worked on them or work on them so rarely that they forget what they learned or knew and start saying "It's f---in' CLOSED ARCHITECTURE!" or "It's a black box" (neither of which is even remotely true). It's just that they haven't had to look at things or study things or review things--because it just works, and works pretty damn well.
Sorry; I get a little passionate about this subject (can you tell???!!!). But if you asked me today what I would recommend, I wouldn't recommend Mark*--for the reasons above. You don't know what you'll be getting. And, the execution of the upgrade leaves a LOT to be desired (HMI issues; lingering alarms; unknown "upgrades" and "enhancements"; etc.). If you could get GE to just duplicate what you got, and offer enhancements you could pick and choose--that would be great. But they WON'T do that. You'll get what they give you, and pay handsomely for it, and be very glad they decided to sell it to you. Your machine will run faster, jump farther and last longer--but they won't be able to tell you how or why or how it will impact your operations. Until it does, and then--well, it's the latest and greatest, and functionally, it's identical to what you used to have.
Hope this helps! Let us know what you decide!
I really appreciate what you put out there, it is a good perspective.
For our Mark VI, I can't do certain things because doing them is password protected through varying levels of permissions. I have the password for a few levels, but after that I don't have access. This limits me from adding new equipment into the control software. That is why I call the Mark VI a black box software. If these passwords are publicly available, than I completely take that back!
Both of the vendors we are talking to have in-depth experience and knowledge on GE turbines, so that should help with the needing to know how the turbine is supposed to operate. The ICON system has not been used for many 7EAs, which makes up a bit more hesitant. Emerson has installed their control system in several of the same frames and have the most experience.
I will be sure to keep an eye out on some of these problems as we go through the upgrade process. Thank you again!
Password protection is not closed architecture or black box. It's possible your site has the "EPIC" code from GE that prevented site personnel from making any changes to code. GE sold that as an enhancement to prevent people from "inadvertently" making improper changes, but it was really to have to call them out to make any changes and keep them servicing "their" equipment. And, you can request GE to permanently disable this EPIC code, making all but the highest level of password protection available.
With the two lowest levels of passwords you should be able to make Control Constant changes, I/O Configuration Constant changes (scaling; inversion masking; etc.), and you should be able to force logic and make most simple application code (logic) changes. If you believe you need to change the "Big Blocks", well, then, you are most likely mistaken (blocks like FSRNV1, or TTXMV1, etc.). There were some glitches in some of the sequencing blocks, but those are easily fixed (not prettily, but easily).
If you can't "see" into Macro blocks (many of which DO NOT have Block or Item Help descriptions), all you have to do is click on the expansion icon next to the block in the Tree View pane, and the block will "open" to expose all of the primitives (ADD, SUBtract, MULTiply, DIVide, COMPare, MINimimum SELect, MEDIAN SELect, etc.) that make up the Macro. You will probably have to hand sketch the various primitives in the Macro and connect them to figure out how they work, but it's doable (and informative in the process!).
GE does not really hide what's going on in the application code (sequencing); they might make it difficult to understand, or time-consuming, and they definitely do a POOR job of documenting what happens in the blocks and Macros so that a detailed review is not necessary. But, let me tell you about Woodward's GAP (Graphical Application Program). When you're looking at functions (blocks) all you see are some signal names. Inputs may be on the left, or the right, or the bottom, or the top. Outputs may be on the right, or the left, or the bottom, or the top. And you CAN NOT see what's going on inside the block. PERIOD. It's a rectangle--that's all. You have to find a manual to find a description of what are the inputs and what are the outputs and what happens inside the block. And, you better hope the manual you are consulting corresponds to the revision of the GAP software you are looking at. I have personally called and emailed Woodward on several occasions complaining that the block isn't doing what is described in the manual, only to be told either the manual doesn't match the revision of GAP software used in the control system, or there was an updated issued to fix problems found in that revision of the GAP software that block was used in. Granted, most of the descriptions are pretty good, BUT, one still has to go find a manual and find the block (function) and then try to figure out inputs and outputs and what happens inside the block.
Not so with GE Mark* software. For all but proprietary blocks like Model-Based Control (ARES) blocks you can pretty much see exactly what's happening in the blocks. How do you think Emerson and other control system integrators figure out how to implement the blocks in their control systems (DCS; PLCs)? They may have some GE-trained people to help with figuring out the blocks and Macros, but for the most part they probably have some people with some turbine experience and some drive to figure out what's happening and how to make that happen in their control system.
I have heard of A-B control system integrators locking down their sequencing and preventing site personnel from making any changes--including changing setpoints/parameters and forcing logic--without having a control system integrator employee on site. These control system integrators insisted their code was proprietary, and they were within their legal rights to do so. (Several sites have scrapped the A-B hardware within two or three years because of problems.)
Anyway, best of luck! Let us know how you proceed, and how you fare.
Good comments from CSA. I would also add on a few more things. Note: I am big advocate of Mark VIe. I think there is so much capability in the system, and I blame the requisition teams not fully capitalizing on all that Mark VIe has to offer.
* I think it is important to try to have the whole control system on one common platform. It makes life a lot easier.
* You probably have password locks on the GT side. The DCS is usually configured differently. Although the software and system is the same, GE puts locks in places so that customers cannot modify the GT logic. When implementing the DCS, no Passwords are used. it is all open.
* There are some smaller integration shops that work well with Mark VIe and know it better than GE. There prices will be cheaper and they are quality of work is better...They still use Mark VIe. I am happy to introduce you to some of those business.
Happy to offer more advice if needed.
www.woodward.com has two publications on the latest Woodward GAP 3 (Graphical Application Programmer). Publication 03377 and 03306. In the main search field you will find them for download.
Under the Support menu for software, GAP and all the block help associated with each version is available for download. If you need a complete set of GAP help for all versions search for "GAP help". There is a standalone installer for the help files. Newer GAP versions contain any blocks that were added and updates to the help for existing blocks.
Higher level blocks are included in the base set, such as a redundancy manager block for dual or triple redundant sensors.
In the latest GAP version, user blocks are now available for creating custom blocks in a library from the included set of base logic blocks.
The functionality underneath can be viewed by drilling in directly on the block, from the navigation tree, or from the user block library. Logic flow is from block inputs to outputs. Before a block can execute all upstream blocks that have inputs to this block is processed first.
GAP 1.X was the initial release in 1995. GAP 2.X released in 2000 that created live monitoring of logic.
It was re-designed with a new look, new navigation, new searching, and new programing tools. It was released in 2010 as GAP 3.X. Visually the "square" blocks were replaced with standard shapes and many blocks have a graphic that show functionality. Explorer style navigation logic tree and browser like tabs are added. Modules were added to group logic in functional sub sections.
GAP 3 has recently created function scripts through the API interface called GoToGAP that allow jumping from an HMI tag or alarm point directly to the logic to monitor, view, and troubleshoot functionality.
Up until recently only Ladder Logic and tunable type settings, gains and curves for example, could be made on line. Woodward released this month full authority on line changes.
What about written explanations of functionality, and is it possible to "see" what happens inside a block without reading a functional description? In other words, is there graphical representation of what happens inside a block (even if it's not animated), along with a written description (one supplied with the software, accessible from the software, that matches the version of the software)?
It's great that other improvements have been made, but what about the core user/programmer/troubleshooter experience that was so lacking in earlier versions?
That's what enquiring minds want to know--that and can GAP 3.0 be used on any Woodward platform for GE-design heavy duty gas turbines (other than Frame 5s)?
We are going with the ICON system. That is the Rockwell PLC solution. It is made by another division of our company, and they have stated that they will not lock us out of the software. I'll post here with an update once we put it in. Thank you everyone for your input. It really made me appreciate the Mark software a bit more. It is clunky and hard to use at times, but I have up to level 3 of 4 access. This leaves me with a lot of options that I can do, but as far as I can tell I cannot add controls for new equipment or edit HMI pages to better suite our needs. The ICON system is supposed to allow that, but once again we will see.
Congratulations on your decision!
I would like to add that unless you have EPIC software you can add I/O and create new logic or modify existing logic with password level 2. And, if you have EPIC software, you can request GE to unlock it--it's YOUR software. Your company would have to formally make the request, but others have and it's doable.
Anyway, best of luck! Let us know how it goes for you! Also, as a matter of curiosity, will you be going with a SIMPLEX system (a single processor), or a dual redundant system (two processors), or the A-B TMR option (three redundant processors)?
Thanks for the feedback, also!!!
Even with EPIC you will have recurring issues with locked up logic. For some reason GE seems to think that GE is responsible for "safety" forever, and that to change logic schemes for things like turbine compartment ventilation trips is forever only doable by hiring GE to do it for you, which of course they will not.
Additionally, in most cases it is not difficult for individuals with patience and understanding of the system to get around the passwords - it just adds many man-hours and risk to things that would otherwise be pretty simple to do.
If you get into your contract that you, the owner, will have access to change and modify all logic and then proactively review all logic during project delivery, point out all problem areas and are willing and able to get the attention of executive level management within your company such that the issue is raised at sufficiently high levels within GE, you will be generally successful in getting access to everything but ARES and stuff like that.
I am in general agreement with most of the things CSA says in this forum, including that the MK-6e is the best "purpose built" turbine control platform on the market. However, the application engineering for any "non-standard" features and also FAT's, commissioning and after market support are horrendous. Even for standard turbine centerline logic and algorithms, you must be prepared and willing to make significant changes to much of the tripping schemes if you are to avoid numerous unnecessary trips and protective actions. Based on my experience with MK-6e projects over the past 8-10 years, I would not consider a MK-6e in a DCS application, simply because of the horrendous application engineering and commissioning support.