Anyone have done this work on their machine before, kindly share your experience: any problem after upgrade, what needs to prepare before, how long outage needed for installation, advantage and disadvantage comparing <I> and HMI, any work required on gateway computer since. FYI, we were
using interface computer to communicate between ABB INFI90 DCS, which controls our boiler and Mark V for our Turbine & generator. I'm planning to upgrade our machine. Thanks in advance.

 

Refreshingly, this subject hasn't come up in a while (though it usually surfaces in post about locating ARCnet cards). This is an attempt to describe some of the issues which have been experienced by some people upgrading to GE Mark V HMIs, as well as some of the options which have been made known here on control.com. A search for 'ARCnet card' will produce many threads, and reading each one will provide contact information for some of the options.

Your operators will have to learn how to use it. For example, the target colors have been changed--and not to protect the innocent!--there are no red targets to indicate immediate action, rather, there are only green targets and grey (analog setpoint) targets. An operator has to click on a green target to see if it requires confirmation or is an immediate action target! (Actually, some very early GE Mark V HMIs DID have targets colored similarly to those on the <I>; why the change was made is not known....)

Further, the names of many of the displays have changed, and the locations of many of the targets have moved....

Generally, on the first usage, the 'Alarm Silence' button doesn't work, and the 'Alarm Lock' and 'Alarm Unlock' functions usually don't work, either.

On the GE Mark V <I>, when an alarm is acknowledged and the condition is resolved (when the status changes to a logic "0") the operator must use the 'Alarm Reset' function to remove the alarm text message from the Alarm Window. On the GE Mark V HMI, when an alarm condition is acknowledged and is resolved the HMI AUTOMATICALLY resets the alarm and removes the alarm text message from the display--which can be MADDENING for slow dithering alarm conditions, AND for operators who are unaccustomed to alarms automatically being reset. And, it CAN'T be changed so that alarms must be manually reset!

Newer GE Mark V HMIs use a "new" scheme for Diagnostic Alarms--they don't list the associated processor in the text message. It's CODED into the drop number, so the operator or the technician must have a "secret decoder ring" to deciper which processor the Diag. Alarm being annunciated is associated with! (Actually, they just need a calculator and the correct formula.!.!.!!!) And this isn't documented anywhere.!.!.!

Most Mark V panels will require a PROMset upgrade to allow communication with an HMI. If not performed properly by a knowledgeable individual, it can take several hours or several days to be able to re-start the unit. (To change PROMs, it is necessary to shut the unit down and power down the processors. To start the unit up, it is necessary to get each processor back to I/O Status A7, and, again, if the upgrade process isn't done correctly then I/O Status A7 can take some time to achieve.)

Many of the targets on the GE Mark V HMI do NOT have their function correctly tied to the proper signals to enable or disable, or to select or deselect, functions. Also, many times the length of time a target is active when clicked on and confirmed (if necessary--akin to the EXECUTE button on the <I>)--is set to 1/16th of a second, which is usually too fast for the Mark V, which generally executes sequencing at an 8 Hz rate, once every other 16th of a second, so it quite frequently misses the signal.

Both of these conditions can be very difficult to find, and, since many of the targets and buttons on a GE Mark V HMI are programmed and designed differently, it can be difficult to resolve because what works for one target won't always work for another. So, plan to spend some time testing EVERY target on EVERY display BEFORE the installation person leaves site (or you'll be sorry!).

There is a "mythical" limit to the number of GE Mark V HMI Servers (HMIs with ARCnet cards) per Mark V turbine control panel: two. (GE Mark V HMI Servers can be multi-unit, but only two servers can request information via the StageLink (ARCnet card) from any single Mark V turbine control panel.) Some users have reported they have three or more GE Mark V HMI Servers per Mark V turbine control panel, but this author is certain that in order to achieve this, they have optimized and "cleaned up" the CIMPLICITY projects (databases) originally provided by GE.

Many operator interfaces (none provided by GE, though) have the ability to hover the cursor over a data value and see the signal which is providing the data on the display, something which is SORELY lacking from GE Mark V operator interfaces.

In order for a Mark V signal value/state to be viewed on a CIMPLICITY display the signal must be present in the CIMPLICITY project--and NOT all signals are in the CIMPLICITY project. AND, adding signals to a CIMPLICITY project is NOT documented by GE in any manual and varies depending on the version of CIMPLICITY being used on the HMI--so what works on another site with four year-old HMIs will probably NOT work on new HMI installation (meaning, you can't get detailed instructions from another site which will always work!).

If your site is accustomed to using the Synonym feature on your <I>(s), it's NOT available on the GE Mark V HMI....

Changing engineering units on an HMI (from metric to english or vice versa) is not an easy task on a GE Mark V HMI (not as easy as it is on the <I>).

Even setting up the dot matrix Alarm Printer (the "Logger") proves very difficult for most people (including GE personnel), as has been proven by a couple of posts here on control.com

There really is only one compelling reason to upgrade from an <I> to a GE Mark V HMI: the inability to purchase ISA (Industry Standard Architecture) bus (sometimes called "legacy bus") ARCnet cards. And this is a VERY compelling reason, since IDOS isn't capable of communicating to a Mark V turbine control panel using a PCI-bus ARCnet card--and even GE isn't selling ISA-bus ARCnet cards!

There are some features on GE Mark V HMIs which are not available on the <I>s, like OPC, and being able to run multiple VIEW tools in separate command prompt windows (which is a very nice feature--but the VIEW tools are STILL command-line executables, with ASCII text output!). The micro-processors used in the newer operator interfaces are also faster--which is also nicer. But MS Windows and CIMPLICITY require very fast microprocessors to even approach the speed of display updates available on later <I>s (supplied with Pentium III microprocessors).

Using a GE Mark V HMI Server, one doesn't need a <G> processor to communicate using GSM (GE Drive Systems Standard Messaging protocol) with a DCS; it can be done directly through a GE Mark V HMI Server.

Multiple display windows can be open on a GE Mark V HMI (though they must be re-sized to view them, as with any windowed operating system).

There have been several posts to control.com about alternative Mark V operator interfaces, and a couple about alternative ARCnet cards which are alleged to be compatible with those provided by GE (how compatibility can be proven since there is no published standard for ARCnet cards used on the GE Mark V StageLink (the ARCnet-style LAN used to connect operator interfaces to the Mark V turbine control panels, AND to EX2000 exciter regulators employing Status S communications on the StageLink) is difficult to understand...).

There is also one contributor to control.com which has indicated his/her company can provide Pentium III-class CPUs and ISA-bus ARCnet cards to upgrade an IDOS-based operator interface to.

Lastly, there is one contributor to control.com who has indicated that there is a solution available which does NOT require a PROM upgrade, uses the same PCI-bus ARCnet cards provided by GE with Mark V HMIs, can duplicate the existing <I> displays (for minimal operator re-training and familiarization), and is MS Windows-based.

But, for most people, the choice comes down to purchasing a solution from the original equipment manufacturer (GE) or from some other entity. Most will opt to go with the OEM solution--but, they are NOT the only game in town. And, the upgrade is not usually without some "growing pain(s)" and more than one call for help or assistance.

Hope this helps!

markvguy

 

Relevant to the above, we are just in the process of retrofitting CSE Engineering's ITC replacement for the MKV <I>. This should be done around June on one machine and I will report back on it's performance to the forum. The price is very similiar to that of an OEM HMI and the features are better.

 

Thanks guys for the reply. I already got approval for ugrading work. I also try get alternative system but their price is double than GE so decide to go back to OEM. Our upgrading work is plan end of May, will feedback later.

TQ

 

Can you provide more detail of the better features?

 

We have a Fr 5 too with <I> and BOI. It's an excellent machine, to say the least. Since the issue of ARCnet was broached here, I have a query. Is it possible to buy a BNC to RJ45 converter fm the market and plug in a normal LAN Ethernet Card in a standard IBM compatible PC, connect it with C-core, install IDOS and get going?

Regards

 

Contact me at yahoo id: rahi172

 

The I to HMI upgrade can really be something interesting especially for the operators. As for the upgrade at my plant, the HMI is a flat panel computer and the interface is more user friendly as it's on Windows platform. All the best you guys for the upgrade...

 

Dear Mark V experts,

I have a team of GT guys who are capable in O&M for Siemens GTs and we'd like to take another step to take on Mark V and Mark VI. We are based in Thailand. Can you suggest where we can get training?

Please contact me ASAP!

geeteeguy [at] gmail.com

 

If there's anyway you can afford to do it, I would skip the headaches of the MARK V HMI phase completely and upgrade directly to the MARK Ve and its HMI. Your going to end up doing it sooner or later so you might as well do it now.

The fact is that MARK V's themselves are so old and obsolete that they will reach a point where GE will no longer be able to buy the chips. This is what happened with the INFI90 BRC100's. Motorola sold the business to FreeScale who quit making the old processors so ABB was forced to upgrade to the BRC300's and 400's. You can't buy a new BRC100 from ABB anymore. Technology keeps marching on.

At first I hated the MARK Ve because we had just gone from I's to MARK V HMIs and now we had a radically different system to learn with no training planned for the forseeable future. To me, the MARK V HMI is essentially a Windows implementation of the old I's. Ve is more 21st century.

We had a reliability issue in the beginning but GE figured out what the problem was and fixed it. The Ve HMI comes with Toolbox that is years ahead of its predecessors. Gone are the TC2K, ALARM.DAT, FORCE, and all those other files. All data is incorporated into Toolbox in a tabbed format. It even includes point descriptors in the logic and in many other places where it's appropriate. What a concept!

The LVDT calibration includes a high speed trend that can record the demand vs. position feedback for permanent storage so you can see performance changes over time. And there's are lot of things I haven't figured out yet because I haven't had the time.

All of the MARK V I/O remains the same. You just change out R,S,T & C cores which are now called PACKS. Why?? Why tweak our brains by changing the name? ANS: Because GE can do whatever GE wants...
The Ve processors are a lot faster and have way more memory so you can run programs such as continuous tuning and casing management. It's good system that will be improved upon over time.

 

It is true that an upgrade to a Mark Ve and the latest OEM HMI would pull you out of the obsolescence path, for a while, but at what comparative cost? GE states they will support their Mark V components and provide service through at least 2014. Also, there are a number of third party vendors and that provide Mark V spare parts and, by read of this forum, a number of excellent experts that know those systems that can help out.

The weak link is often the <I> and HMI and they are far less money to replace. In addition to GE, there are other companies that make replacements specifically designed to interface with the Mark V, share the data over Ethernet, track more tags at higher scan rates due to managing Stagelink traffic more efficiently, etc. CSE Engineering is one of those companies that has a product and there are at least 1-2 others around the world. You are always trying to stay ahead of obsolescence curve. You might consider pushing out a major controls upgrade as much as you can in this case.

 

I'm surprised that the MARK V's will be supported that long.

We're heavily supported by GE and have a good relationship with them so we tend to limit ourselves to their products. Their HMI software packages may or may not be the best solutions available, but we haven't tried any third party software to compare them to so I really can't comment on that. We're not completely unaware of their competition either. After all, our favorite joke about GE is, "You may find better, but you won't pay more." Another good one is, "Imagine it works." (I got that one from a GE technician.)

Being a utility operating four 7FA's in a combined cycle plant, our dispatch, which can control us remotely, runs them all over the map for anything from spinning reserve to full load plus duct burners, to var control, voltage control, load control--whatever dispatch needs them for. They also have an excellent heat rate, can go from cold start to full load quickly--much faster than any other plant in the system with the exception of our simple cycle units--and they'll swing load very rapidly. Dispatch loves us, and we hate dispatch.(I never thought about it this way but it's truly an industrial love-hate relationship.)

Naturally, our turbines and HRSGs take a lot of abuse from all that, plus they have to endure numerous starts. In fact, when it was left up to dispatch, they'd start them twice a day, which turned out to be too much for them (cross fire tubes started migrating, and other scary things began happening). Site management finally had to say no.

Unfortunately, these machines (DLN 2.6's) are only tuned a few times a year remotely from Atlanta, and not as often as they really should be (it's not cheap) so they become susceptible to lean blowouts, especially as the weather swings around, and the combustor dynamics and criticals during startups are pretty harsh so things tend to rattle apart, become migrate, and so on.

The MARK V's don't have the capacity to run the continuous self tuning software GE is now marketing, but the Ve's do. A substantial investment by itself, the CST software may well be worth it to us or others like us from the standpoint of efficiency gains, fewer trips, better startup characteristics, and so on. Who knows?

There's also GE's "casing management system" which we run on our Ve controlled CT (the V couldn't run it) which actually does improve efficiency. If dispatch would quit swinging it around so much (it must be at a fairly constant high load for the CMS to become active) it might eventually pay for itself.

The Ve can also run a their more fault tolerant exhaust thermocouple algorithm as well (which they say will not run on the V). That should reduce exhaust TC trips and run backs.

So in hind sight, upgrading to a Ve may not be worth it to you if you don't need the additional processing power.

 

Dear GA,

I have not had the pleasure to read any posts from you previously, I look forward to more comments from another GE user in the trenches. To poke a little fun should your statement not have read " We heavily support GE"?

We too commonly use the statment "You may find better, but you won't pay more." I have not heard "Imagine it works." but I like it.

My real question relates to the CMS for the 7FA. Can you briefly elaborate on its function? Hopefully it is ok to ask on this post, rather than starting a new thread! We operate 2 7ea's, in load following mode so I understand "swing", although thankfully our department manager so far has been able to convince our power control to try and balance the swing between our plant and some hydro's, so we don't take all the abuse!

 

We were one of the Beta sites for the CMS and the MARK Ve, from what I understand. I don't know a lot about the CMS because maintenance had absolutely no involvement in the project. What I "know" is really opinions. Some in operations claim that it runs occasionally and others swear it never runs, and nobody is really interested in it any more because the guy who spearheaded the project left the company.

What I think its function is to cool the casing to reduce the compressor blade tip clearance. It consists of a blower, duct work, a cooling jacket, some AOV's, a pressure transducer, throttle valve, thermocouples and, of course, the additional control logic. In theory, it will pay for itself due to the efficiency gains but to see the payback you, of course, have to run it.

From a maintenance tech's standpoint, it's just more useless junk in your way of getting to the things that really matter. But it's not as bad as climbing around all the clutter of steam power augmentation piping we don't use. (Why don't we use power aug? Because the HRSGs aren't big enough so the steam that gets diverted to power augmentation reduces the steam flow to the steam turbine, reducing its output by roughly the same output increase you gain on the combustion turbine. Plus, power augmentation accelerates the wear and tear on the turbine section, so your maintenance costs go up.) My personal preference is if you're going to install two systems you're not going to use, go for the CMS because it's cheaper and takes up less space and far less down time during hot gas path inspections!

But I digress. I'd be more enthusiastic about CMS if it ran more often than the moon turns blue. As I mentioned earlier, dispatch runs it all over the map so it's not very often that CMS runs. So a lot of money was spent on something we rarely use, and yet they say that they can't afford good training on the stuff we do operate and maintain?

Hah! Maybe if they quit blowing money on expensive gee-wiz-look-at-me-boss projects, we'd have some money for training. (Yeah, I know, I'm a dreamer.)

The bottom line with CMS is that for it to be worth the investment, you've got to run the turbine at a fairly high, steady load, which we seldom do (and never did before it was installed, so go figure). GE was up front about this requirement, in fact they initially told us we'd have to be at base load, +/- very little for it to be of any benefit. Over time, however, GE found you could run CMS at a lower load and a broader range so is probably worth the investment for folks who run within it's design envelope.