S7 400 Redundancy

culi

Member
Join Date
Sep 2007
Location
Tabilocan
Posts
8
All,
I'm involved with a project using Siemens S7 400 CPUs. I want to build a redundancy two PLCs, actually I have no exprience about it. Anyone has some advice me about Software, hardware. I did many projects using S7 300 CPUs.

Thanks,
Culi
 
To achieve true redundancy you need to use three PLC's (or similar devices) by three different manufacturers. Anything less and you are just fooling yourself.

If you want to improve reliability use the K.I.S.S. principle
"Keep it Simple, Stupid".

Use one PLC and keep onsite spares for EVERYTHING.
 
I posted this, really wanna get some good advices from expertise guys.Pls help me!
John Gaunt!******** comment. In case you don't know, pls shut up.Right?
 
OK culi, I will shut up.

But first I do have some experience with redundancy.

I work with S7-400 PLC's and have seen several attempts at using them as redundant pairs. On one system it was such a disaster, they had spent all their time (money) trying to get redundancy to work they had totally neglected the rest of the system and it was running totally manually.

A couple of years ago I worked on the software for a North Sea oil and gas system off the coast of Scotland using 4 computers for redundancy. The oil and gas is an industry where truly "money is no object" The system allways worked well using a single computer. It simply never failed. With two or four computers it never ran for 24 hours without errors. The whole project was so tied up with customer specifications that it simply could not be fixed without going outside those specs.

Ok, I'll shut up now. Good luck.
 
My first thought was that your anecdotes are silly, particularly the second one. I may be more convinced if you described the system where "money is no object" that "never ran for 24 hours without errors". A lot is left to the imagination when you provide zero technical details.

However, there's a lot to be said about your message. Redundancy necessarily adds complexity (read points of failure) to a system to provide additional reliability! In some cases it is worked out. Consider a RAID 1 (mirrored) volume. The increased complexity of the RAID controller versus disk controller is negligible compared to the added protection of having mirrored data. You can even have dual controllers. Adding hardware to go from a single hard drive to a RAID 1 array will provide a statistical advantage. But, behind the scenes, there's a complex driver and hardware that some group spent a lot of time engineering well - orders of magnitude more time than your PLC project. A crappy driver (redundancy guts) would make the whole system less reliable than just using a single drive.

The lesson behind the story - if there's a working redundancy product, go for it. I haven't heard of any PLC manufacturers that can come up with a generalized redundant scheme that's seamless to the PLC program. If they have, that's another matter.

If you're trying to set up your own redundancy scheme, you will likely fail. I've seen attempts like John's. OP, this is particularly true since you "have no experience at it".

I think that redundancy is so immature with PLCs because they don't fail often, and it's an incredibly difficult problem to solve in general. Ask anyone here how important "stability" is - they'll go on to no end. PLCs are pretty stable and having a "hot backup" is simple and effective. To do the job right, you're probably talking a ten-fold in cost for added minutes of uptime. This doesn't even take I/O and everything else into account.

John Gaunt said:
OK culi, I will shut up.

But first I do have some experience with redundancy.

I work with S7-400 PLC's and have seen several attempts at using them as redundant pairs. On one system it was such a disaster, they had spent all their time (money) trying to get redundancy to work they had totally neglected the rest of the system and it was running totally manually.

A couple of years ago I worked on the software for a North Sea oil and gas system off the coast of Scotland using 4 computers for redundancy. The oil and gas is an industry where truly "money is no object" The system allways worked well using a single computer. It simply never failed. With two or four computers it never ran for 24 hours without errors. The whole project was so tied up with customer specifications that it simply could not be fixed without going outside those specs.

Ok, I'll shut up now. Good luck.
 
Once again it would be nice to have some understanding of the system or process to understand the need for redundancy. As pointed out redundancy is a good thing BUT at a cost -- more stuff to break.

Some vital areas where redundancy is needed in my mind are
hospitals (redundant power supplies, backup generators etc) Might be nice to keep the lights on for the surgeon. Dang it is dark - I think I have the appendix.

Aircraft again redundant power supplies and redundant hydraulics. Make a big mess when they crash. Are 4 engines really better than only two??

Submarines more redundant power supplies, 3 to 4 levels of hydraulic systems. Off the top of my head the only thing we did NOT have a backup for was the galley range. Well the good news is they are tidy when they sink -- right to the bottom out of sight and out of mind.


The other thing is who or what is going to do the controlling? What comes to mind is fire control where the computer cannot decide on which target to shoot (too many Tom Clancy novels ?).

How do you shift control from unit to the other?

Dan Bentler
 
I just want to mention that there are 2 variants for CPU redundancy with S7.

There are the S7-400 CPU's with an "H" in the typename.

And there is a software based runtime package for the S7-300. This variant has significantly longer changeover time as compared to the S7-400H/FH CPU's, but can be OK for some applications.

Apart from that, I think you should contact Siemens or an integrator that is proficient with redundant systems.
 
Dan - lots of submarines have "just" 1 core - but that hot rock does its job well ;-)

leitmotif said:
Submarines more redundant power supplies, 3 to 4 levels of hydraulic systems. Off the top of my head the only thing we did NOT have a backup for was the galley range. Well the good news is they are tidy when they sink -- right to the bottom out of sight and out of mind.
 
surferb said:
Dan - lots of submarines have "just" 1 core - but that hot rock does its job well ;-)

True - was a Nuc. Gotta admit anything Hyman G had a sayso in design was RELIABLE. Even the reactor had backups and if it scrammed we had the battery, EPM for propulsion, and diesel for power.

The only backup for the galley range was an oxyacetylene set, weld rod oven (boiled eggs at 2AM), and maybe hot steam pipes. Electricians were under a lot of pressure (wardroom hand wringing) when the range broke.

Dan
 
Dan's got the right idea with "RELIABILITY" - that's really what OP should be looking for and asking about. Redundant PLCs might come as the final steps, but I doubt it. I've worked with too many end users who demand "redundancy" for the sake of it, thinking that they can throw a couple thousand $$ at the problem with extra equipment and make it all better.

Dan - I could tell you were - that's why I was so targeted with my response ("hot rock", might as well have mentioned the little "steam molecule" traveling through the systems ;-) ). For everyone else Rickover, the first Naval Reactors director demanded that everything be reliable and redundant. I think Dan is litterally correct that the only thing without one or more backup systems is the gas range. That's why I was joking about the core being the other - of course, it has muliple primary and secondary loops and redundancy everywhere else with everything.

Very relevant to this post is that everything on a submarine was carefully considered. They didn't add hardware just to add hardware. A lot of the reliabilty was created by good design. Wish I could get into the basics of design safety with the rods - I don't think it's classified, but I'm not 100% sure. I'll take a look on the net and see if anyone explains it well. Dan - care to take a stab at it?

leitmotif said:
True - was a Nuc. Gotta admit anything Hyman G had a sayso in design was RELIABLE. Even the reactor had backups and if it scrammed we had the battery, EPM for propulsion, and diesel for power.

The only backup for the galley range was an oxyacetylene set, weld rod oven (boiled eggs at 2AM), and maybe hot steam pipes. Electricians were under a lot of pressure (wardroom hand wringing) when the range broke.

Dan
 
Last edited:

Similar Topics

Hi, I had a client upgraded their system to use a set of redundant PLC system with GE PLC CPE400, and a Citect 8.0 (2016) SCADA to get data with...
Replies
0
Views
1,448
how is redundancy accomplished using the above configuration? If the outputs are taken from both modules,wont they get added up?how does error...
Replies
4
Views
3,369
What is software redundancy, how to use, configure and test with plcsim ?????????????????????????????? In simatic Manager v5.4
Replies
1
Views
2,690
Give me how to make redundancy in siemens s7 400, which we can check with plcsim...I didnt used redundancy till now....Basically iam a plc...
Replies
4
Views
7,699
I am looking to understand how Siemens implements redundancy in their S7-400 series PLCs. Specifically, we provide a DCS that speaks modbusTCP...
Replies
1
Views
4,223
Back
Top Bottom