Hi. I'd would like to know if anyboy could help in my DCS migration. Today I have a TDC3000 and I'm intending to make a full subsitution of this system to ABB AC800 or Yokogawa CS3000. First, I'd like to know if does anybody have experienced this kind of migration. Second, if does anybody know if I can sell my currente system (as spare parts, for example). Is there any used DCS market anywhere (in Brazil if possible)? Thanks a lot for the help.
I've been involved in a couple of migration projects, including a TDC2000 to AC800 migration project. a) On the hardware side, the only option is to do re-wiring. Exactly how you do it is dependent on a number of factors, but after the migration project the wires must be connected directly from the field instrumentation to the new DCS. b) The actual change-over of loops from the old system to the new can be done either on-line or off-line with the process shut down and many factors would determine the best choice for each case. A combination is very often used; do the non-critical loops on-line with the plant running attempting to cover most loops like this and then do the critical loops off-line with the plant shut down. c) One must at least evaluate to do automatic software configuration. For the existing Honeywell system you can get text files ("EB"-files to use the Honeywell term) of the software configuration. Now you must write a custom computer programs to convert this to a format suitable for the new system. Although I have done something similar, I would be very sceptical and challenge this approach. (Let's just say I've learnt the hard way that it's not always wise to be too clever). The default approach that I would follow is to do the software configuration of the new system from scratch.
I have been involved in two migrations. The first was from a TDC2000/Modicon PLC system to a Foxboro I/A DCS and it involved installing new I/O cabinets and completely rewiring all the field devices. We performed the migration over a two year period and had both systems running side by side until the migration was completed. The upside was it was a slow transition for the operators, the downside was it took forever and rewiring everything was very expensive. The second migration was from a Westinghouse WDPF to a Foxboro I/A system using migration cards and utilizing the existing Westinghouse racks and field wiring. An 900 I/O point system was migrated in two days. The upside was it was quick and very cost effective, the downside was the operators had a very rapid transition to the new system, although that is quite manageable with good training. If you haven't seen the I/A migration route I'd definitely check it out as a comparison to your ABB or Yokogawa solution.
..of course the *downside* to retaining the existing racks and field wiring is that you are now stuck with multi-vendor support issues and when the Westinghouse bits fail at 12 midnight because a crucial speck of dust was disturbed during the upgrade and you've run out of spares and can no longer buy a replacement part (because Westinghouse are p*** that you didn't go with them), you are back where you started.
To avoid all this, you still need to progressively change over the racks and field wiring once the new system is up and running, and since you then need to replace your I/O cards (again!) - because migration cards don't fit in standard racks - the cost of the project skyrockets.
Although the side-by-side method is longer and more painful - at least you wind up with a fully-supported and up-to-date control system in the end (and happier operators and maintenance managers ;-)
Cameron, I am not sure what your experience was or what company it was with but this was my experience with Foxboro.
The only Westinghouse components that were retained were the racks, the field wiring assembly's, and a few power supplies. It has been my experience that the racks and the field wiring assemblies rarely have anything go wrong with them unless you have a catastrophic failure. If I need one, and I haven't since the original system was installed in
1989, I can get one from someone other then Westinghouse. Foxboro anticipated that most vendors won't want to support a system that was migrated to a competitors system and fully supports the completed installation including the non-Foxboro components of the new system. There is no need to replace I/O cards again after the project is done. The new migration cards are identical to the latest Foxboro I/O cards except they are in a form that will fit into a competitors frame, in my case it was a WDPF Q crate. This was a compelling reason for us to go the I/A route because the Westinghouse solution was to retain the existing 15 year old I/O which would have put us into a bind in the near future. If any I/O needs to added in the future you can either buy more migration cards if rack space is available or you can add a new enclosure with standard I/O modules. The standard I/O modules are cheaper and it makes more sense to go that route if you are adding a significant amaount of I/O. Both the migration I/O and the standard I/O co-exist on the same network and they are the same as far as the system is considered. We chose the side by side route for the first migration because the TDC/Modicon installation was a mess and the I/O was located in the control room and we wanted to distribute it in the field. The migration solution was chosen for the second project because the WDPF installation was pretty clean and we would have never been able to justify the capital expense of a complete replacement of the field wiring. Both routes have pluses and minuses. I don't work for Foxboro if that is what you are wondering. I can say that they are the most customer service oriented control system vendor I have worked with in the past 18 years.
I have to agree with Eric for three reasons; I like to work with Foxboro gear and software, there is a hidden cost to all control systems - documentation and time to make the changes. Most control systems are self-documenting to some degree but I am still looking for a field wiring system that is.
Speaking from personal experience, we have some control gear that is very quickly approaching extinction and there isn't much I can do about it because we lose so much revenue shut down. We have three different kinds of outage; a 5 day annual outage, a 10 day outage every so often and a 30 day outage every 4 or 5 years. We count the days from breaker open to breaker closed, those days will include plant shutdown and plant startup. That is the most I will ever get. So I would rather not change my field wiring (why should I?) and live with better software and operator interfaces.
Of course, I have also worked on complete upgrades in a previous life. All of them were machines or materials handling lines where we had the luxury of weekends where we could swing bits from one system to another.
That's two reasons, and have never worked with the Foxboro gear so I can't comment on the first one. I agree that most customers don't take change documentation and time into account - they just want it done, and done now! - but again I can't see how that makes a partial conversion less painful in the long run.
The systems we tend to work on (airport refuelling, power generation and the like) often cannot be shut down at all - changeovers must be done live with proper training and procedures in place. Breaker open to close must be minutes not days... all the more reason to have a complete working system right next to you that you can fall back to very quickly.
P'raps I got it wrong, but I took Eric's "field wiring" to mean connections from the signal conditioning nests to the field and not the field wiring itself, per se. The better systems out there (Yokogawa, etc.) all allow you to re-use the old nests and terminal blocks and literally unplug the old I/O and plug in the new.
If Foxboro have a _full_ range of I/O in "migration package" format to suit Westinghouse racks - that's great and if so, they would be one of the only ones. Manufacturing I/O cards to fit different manuafcturer's racks is extremely expensive and not very common in my experience. More common is to replace the old processor with a "Remote I/O Migration Module" allowing you to use an entire I/O rack as 'Remote I/O", connected to the new node in the cabinet next door using some form of high-speed comms.
With this sort of system, although it's great for testing and you can get your new platform up an going really fast, the truth is that once a system is "migrated", the old bits remaining are forgotten, just waiting for the next I/O card to fail...
Have you considered Honeywell's New Experion platform for your migration? They offer large discounts on hardware costs for customers who already have an installed legacy system like the TDC3000. I haven't had any experience with ABB or Yokogawa but I have with Rockwell's PLC5's and Controllogix and I'll take the robustness and reliability of Honeywell any day over Rockwell.
The Honeywell Experian system is a AB plc with a server connected to it and a WEB browser connect to the server. If you want IT to do process automation by sending the work to India then Experian is the system you want. This is from first hand experience.
I suggest you check out the new Allen-Bradley systems. They can do everything a DCS does for less than half the price, and there are many independent integrator close by.
I think you have it wrong. The experian system cannot do want a PLC can do. For one thing it cannot communicate with other systems. I know one place it was installed and to get an OEM's PLC to communicate with it required hard wire inputs and outputs to be wired from the PLC to the DCS.
Also, as far as cost, the AB I/O etc. is more expensive than the Honeywell. There are several places where the experian system was installed and Honeywell I/O was used because of cost.
In most processes you want to control temperature, pressure, etc and log data. The PLC systems can do all of this. The DCS is great for large integrate oil refineries especially when they spend the money for optimization routines, but most application a PLC that is compliant with the IEC 1131 programs can handle most applications
sits in the Rockwell chassis) can talk "peer to peer" to PLC's using any of the following protocols, which covers all the cases I have come across so far...
- Modbus RTU using the SIM module (Serial Interface Module) which covers lots of PLCs.
- Allen Bradley DH+ using the SIM module
- Allen-Bradley PCCC or CIP via Ethernet or ControlNet to PLC5/ ControlLogix/etc.
- Profibus DP
- hardwired I/O
And there are a few neat things that I like about the Experion Controllers which I have not seen done well (or in some cases at all) in PLC's yet, such as:
- Viewing the logic charts with live data in the operator stations without any additional configuration, just call up the tag detail in the station.
- Truly only a single database for tags. Build the logic and it is available without any import, translation or anything, just use the tag.
- Model Predictive Control in the Controller, which is a great improvement over PID for controlling systems with transport delays.
- On-process migration for firmware and software upgrades so there is no downtime for control of the process.
- Seamless controller redundancy which requires only configuration of a couple of modules and no code.
- Advanced continuous control function blocks which automatically support cascading loops, back initialisation, mode handling, safety interlocks, etc.
There are of course some tasks which will be more suited to a PLC , like lots of fast digital I/O or fast sequencing, high speed motion control, etc. The top end of Experion can talk to PLCs from many vendors, and then you can mix PLC's and C200's in the same system using each when they make sense, and providing all the data on a common HMI.
TheHoneywellExperionsystem can talk to just about any controller that was made to
Essentially what this means is you can link the Honeywell C200 platform to Rockwell ControlLogix PLCs (in parts of the the plant that need fast digital I/O processing) getting the best of both DCS and PLC functionality and at the same time reducing the spares count.
A win-win situation for large mixed-use plants (steel milling, etc.) and definitely the way of the future for all modern control systems.
I'm questioning your rationale on what you would like to change to [ABB AC800 or Yokogawa CS3000]. It may be worth considering upgrading to Honeywell Experion - There would more support from Honeywell.
The best resale would be a trade-in with Honeywell. There are re-sellers of PLC and DCS used/surplus equipment -but I don't think they would give you a better deal than Honeywell's trade-in.