Multiple PLC's in a line

Hey all,

We're doing some initial planning for an automotive assembly line and need to decide on the controls architecture. Allen Bradley has been decided on as the supplier.

The line will have 4 'zones' each with 2-3 robots, fixtures, turntables, buffer zones, etc. These are staged in parallel to a degree, if one of the zones goes down, there will be another one to continue production.

Our debate right now is how to manage the controls. Should we go with only one Control Logix and have it manage everything, running lots of remote I/O? Should we put in multiple PLC's to manage each zone independently?

On top of this an MES is running, and we need to maintain precise production data across the line, especially if the MES goes down, which could be difficult to manage with multiple PLC's.

Lots of challenges here, but nothing terribly difficult, just looking for the communities opinion.

What if the production PLC (or PLC's as I assume will be the case) go down?
What if there is maintenance or something required on a bit of kit and you now have to take down the entire production line?

I'm playing devil's advocate here as PLC's don't tend to fail as often as everything else, but it's a consideration.

Hey, and welcome to the forum. These engineers and programmers are the best there is and will always help you as long as you make an effort to solve your problem and not just ask for the easy button.
I worked programming a framing machine (made Pick-up truck frames) that eventually went to St. Thomas, ON and we used 4 standary PLC's and 4 Safety PLC's and everything was PointI/O. It worked very well and also interface easily with the Plant FIS. There were 5 quite complex lines, each having many cells and robots. Plus safety, HMI, material handling robots, turntables and a myriad of welding robots. You might consider something like this. A-B control Logix and Guard Logix.

I agree. Essentially the system is designed so that a zone going down won't stop the line, but designed for and actually having it work is a completely different can of worms.

If the zones are truly parallel, I'd vote for 4 controllers. Benefits:
- You can truly work on one without worrying about stopping another
- If they are safety PLC's, you can download to one without E-stopping the others (a download will cause Estops)
- Faulting a processor will only kill one line (hopefully you don't do that often, but...)
- Program scan times will be maximized. I don't know how fast things will move, but don't underestimate the potential headache of slow scan times in a large system.
- A communication or hardware issue (bad switch etc) should only stop one line
- You are far less likely to find your technicians accidentally troubleshooting the wrong program - wrong equipement.
- In short, I think 4 processors will pay for themselves in the long term ability to keep production happening

Disadvantages:
- Cost of extra controllers
- Might need to setup and manage some communications between controllers

kekrahulik said:

- Cost of extra controllers

Click to expand...

This isn't always true... it depends a lot on the family and additional ports required.

Processor cost will be negligible in a line of this size, how will the zones be built? Will they be delivered to the customer in phases or all at once. If all at once, (1) safety processor could do the job but if in phases, consider that you can't be debugging the second, third, or fourth phase while the first phase is being installed and debugged. My point being, the scheduling of the project will be factor on how many panels/processors there needs to be.

Anymore, say at Ford, GM, or FCA, even each station or zone gets its own processor. Over a period of years that the line is making parts, that cost disappears (to the customer). Tracability is very big now in carland so lots of RFID, 2D lasers, 2D readers, and interaction with a host computer are the norm so multiple processors actually get easier.

We often find that for commissioning purposes, multiple PLCs is way easier to handle than just one. It makes it way easier to have multiple fully productive engineers online and testing code at the same time. Also minimizes risk to the project, which sometimes is worth a little extra HW cost by itself.

Downside is managing the communications. Often it isn't a problem, but you definitely need to account for it in the planning stages.

More PLCs for sure.
Graduates have to learn the "don't put code in that faults the production controller" lesson the hard way, but having 25% of production staff staring at them is probably sufficient.

Just have a standard philosophy for managing Comms and Comms failures between PLCs and you'll be golden.

You have 4 zones, and then another set of 4 parallel zones? For sure you should have multiple PLCs. It sounds like your company is taking precautions to keep the line always running, so having multiple PLCs spread out would eliminate some down-time.

Do each of your zones have multiple machines with there own PLCs in them?

AustralIan said:

Graduates have to learn the "don't put code in that faults the production controller" lesson the hard way, but having 25% of production staff staring at them is probably sufficient.

Click to expand...