Opinions, PLC vs PID controller

allscott

Member
Join Date
Jul 2004
Posts
1,332
I am looking at replacing an old Eurotherm EM2 temperature control system. The EM2 controls 38 temperature loops a pressure control loop and a water level loop.

The system works fine but the controls are from about 1991 and parts and technical support are starting to become an issue. As I see it I have 3 options.

1. Upgrade to an EM3.

2. Install a bank of individual controllers (we still use Eurotherm extensively) and send SP and read PV from the main PLC 5 via devicenet to them.

3. Install a new PLC to do the control and interface it with the existing PLC.

I would prefer to install individual controllers for 2 main reasons. First a single controller failure probably wouldn't cause the line to be shut down. When this line goes down for any reason it takes about 2 days to start it up again and the startup costs are huge. Secondly, I think that the individual controllers are better suited for temperature control as that is what they are designed to do.

I am being pushed to go the PLC route. People higher up than me would prefer a PLC solution quite frankly because that is all they are used too.

This is not my area of expertise so I will be doing a lot of research in the next few weeks. I am much more comfortable tuning a temperature controller than trying to tune PID's in a PLC, especially this many.

Right now I'm just curious to hear other peoples thoughts on the matter.
 
38 Temperature loops.
Heat Only?
Cool Only?
Heat/Cool?

Pressure control is usually easy with a PLC PID.

You had mentioned that a stand-alone controller is attractive, as the loss of one zone does not take down they system. A PLC (PAC) is determinisitic and robust, We trust them.

If econimics are an issue, I would think a single controller would more cost effective than 40 stand-alone.
 
38 Temperature loops.
Heat Only?
Cool Only?
Heat/Cool?

Pressure control is usually easy with a PLC PID.

You had mentioned that a stand-alone controller is attractive, as the loss of one zone does not take down they system. A PLC (PAC) is determinisitic and robust, We trust them.

If econimics are an issue, I would think a single controller would more cost effective than 40 stand-alone.

All of the temperature loops are heat/cool. Some are air heat, some are water heat. All of them are air cooled.

The pressure loop is the one that has me most concerned, it is fast acting where the temperature loops are not. It has three valves to control pressure, two inlet and one exhaust. I am considering getting rid of the larger inlet valve and just using the trim valve as time to pressure really isn't an issue.

I guess I should trust the PLC too as there are 3 of them controlling the line...
 
A single loop stand alone temperature controler designed for HEAT/COOL applications most likely will out perform PLC(PAC) out of the box, unless you invest the learning curve to fit your selected PLC (PAC) solution.

A PLC(PAC) single PID may not be a good solulution as a single set of PID constants probably are not a good fit, as the heating dynamics (thermal lag and slope) are probably different than cooling, unless you plan on having different PID variables when heating vs. cooling.

UNLESS, Can you experiment and play with similar temperature control loops with real time equipment that is not in production, to obtain "trends", and data, and tune PLC PAC based PID control.
 
2 key questions:
Do the Eurotherms talk Devicenet?
Does the PLC 5 do heat/cool with separate PIDs?

- The I/O wiring is a break-even.
- Cutting and mounting 40 single loopers is more work than mounting a rack.
- The Devicenet wiring is additional work.
- Does Eurotherm have a rack-based PAC that can handle heat/cool because it's designed to, like Honeywell's HC-900?
- the SP recipe changes, however implemented, are common regardless of platform.

If PLC 5 does not handle heat/cool without a lot of work, I'd do multiple single loopers.

If the PLC 5 does handle heat/cool, then it's probably more work to do the single loopers, unless having 40 temperature indicators is a plus.

The dual inlet pressure valve should run like a batch controller - the large valve runs on/off, full on until ~90-95% of SP, then the big one closes, and the small inlet valve is enabled to 'bleed' up to SP.
 
For the heat / cool functions I would advice you to have a look at Omron Celciux EJ1G Modular Temperature Controller, very easy to setup, works fine with DeviceNet.
We have these units installed on seal bars, and control these within a few degree of set point, where a PLC5 PID was hard to handle it within 10 degree of setpoint.
 
I would strongly consider the dedicated temperature controllers. You can probably find some that are designed to handle heat and cool. I like the suggestion above about using a field bus for communications so the analog wiring can be minimized. The temperature controllers probably have a auto tuning that the PLC does not.

The pressure control is different. I WOULD NOT USE A TEMPERATURE CONTROLLER FOR THIS!!!!! As you noted, a temperature controller is designed for slower systems. Also a pressure control system is an integrating system, temperature control is not an integrating process. I would do the pressure control in the PLC. What kind of pressure control?
 
2 key questions:
Do the Eurotherms talk Devicenet?
Does the PLC 5 do heat/cool with separate PIDs?

- The I/O wiring is a break-even.
- Cutting and mounting 40 single loopers is more work than mounting a rack.
- The Devicenet wiring is additional work.
- Does Eurotherm have a rack-based PAC that can handle heat/cool because it's designed to, like Honeywell's HC-900?
- the SP recipe changes, however implemented, are common regardless of platform.

If PLC 5 does not handle heat/cool without a lot of work, I'd do multiple single loopers.

If the PLC 5 does handle heat/cool, then it's probably more work to do the single loopers, unless having 40 temperature indicators is a plus.

The dual inlet pressure valve should run like a batch controller - the large valve runs on/off, full on until ~90-95% of SP, then the big one closes, and the small inlet valve is enabled to 'bleed' up to SP.

The Eurotherms do talk Devicenet, we have several dozen of them out here.

On a similar line we have a PLC5 with 1771 TCM cards doing the temperature control.

http://www.ab.com/en/epub/catalogs/12762/2181376/2416247/1239760/1551233/tab6.html

I do not see a similar card for a CLX system.

Eurotherm has their EM3 system which I believe is similar to the HC-900. It has been ruled out as an option though.
 
For the heat / cool functions I would advice you to have a look at Omron Celciux EJ1G Modular Temperature Controller, very easy to setup, works fine with DeviceNet.
We have these units installed on seal bars, and control these within a few degree of set point, where a PLC5 PID was hard to handle it within 10 degree of setpoint.

Thank you for the input but if we do decide to use individual controllers they will be Eurotherm 2216e. They are our plant standard for temp controllers and I am familiar with their operation.

http://www.eurotherm.com/products/controllers/single-loop/2200/2216/
 
Peter we are controlling Nitrogen pressure at about 8bar in a tube 18" in diamater and approximately 200' long.
 
Peter we are controlling Nitrogen pressure at about 8bar in a tube 18" in diamater and approximately 200' long.
The pressure will not increase or decrease very quickly. You could try a temperature controller. I would still feel more comfortable with a PLC because then I can add extra code to do things like add a bias or feed forward.
 

Similar Topics

With the Automation Directs Open Source Arduino Compatible PLC and the Open plc Project, plus all the Raspberry Pi HMI's, what are your opinions...
Replies
12
Views
4,263
I'm hoping I can get some input from you folks here. I have been working on and off in Automation for over a decade, mostly with Siemens PLCs. I...
Replies
12
Views
5,327
I am currently looking into standardizing my companies control options; given the simplicity and redundancy of our systems i feel like a PLC/HMI...
Replies
7
Views
2,965
When communicating between different PLCs using the messaging instruction, is it better to have PLC 'A' read from PLC 'B', or have PLC 'B' write...
Replies
2
Views
2,183
Page 356 in the current Grainger catalog caught my eye. They're selling this Dayton PLC that seems to be rather cost competitive for a low-end...
Replies
10
Views
5,924
Back
Top Bottom