Redundant Loop Controller Wiring

M

Thread Starter

Mark

Our manager wants a new PLC system for 40+ PID loops. In addition to the PLC, they are requesting backup stand alone loop controllers for all 40+ loops. The concern is if the PLC fails, there is no way to run the plant.

I have questions on the following topics:
1) Wiring of the analog outputs (analog inputs should be OK)
2) Preventing windup of the backup controller
3) Switching between the PLC and single loop controller
4) How to deal with Cascaded Loops
5) Any other thoughts you guys may have to consider?
 
I don't think it is a wise idea to have PLC and single loop controllers as a back up. If you want high availability consider the following options:

1. Use redundant controllers, redundant input and output cards.

2. You may use high availability fault tolerant controller in redundant configuration also.

2. Considering different plant areas may be use two different redundant CPUs and associated I/O cards for two different plant areas.

3. Take redundant power supply units.

4. Configure control valves in such a way that it remains in stay put position in case of positioner input failure.

5. Conduct an availability analysis in consultation with PLC manufacturer to see how you can achieve highest availability. Obviously cost will increase.

If you really want to use single loop controllers, I will advise use a manual switch over unit and keep single loop controllers in manual mode normal time.
 
Considering the magnitude of the cost of resolving the problems you stated for using back-up single loop controllers (engineering, additional components, wiring, additional failure points), you might well consider redundancy in your primary controller.

Whereas, at one time, only DCS's offered redundancy, redundancy is now common with current high end PLC or hybrid controller technology that is capable of 40 PID loops at reasonable throughput rates.

All appear to offer redundant processors & power supplies and the switchover module. Some systems use redundant I/O, others hot swap I/O.

We use Honeywell's HC-900 hybrid controller,
http://hpsweb.honeywell.com/Cultures/en-US/Products/Instrumentation/hybrid/hc900/default.htm
but there are other high end PLCs that offer redundancy, A-B's ControlLogix for one.

If you decide to continue down the single loop back-up road, Honeywell's single loop UDC controllers have 2 means of dealing with
switchover:

1) digital input function STOP I, which inhibits integral reset action. (You asked how to inhibit reset windup . . .)

2) digital input fucntion OUTPUT TRACKS INPUT 2 — Contact closure allows Output to track Input 2. While the switch is open, the control output is in accordance with its predefined functionality. When the switch is closed, the control output value (in percent) will track the Input 2 percent of range value. When the switch is reopened, the output will start at this last output value and normal PID action will then take over control. The transfer is bumpless.

The primary (PLC) controller's output is wired in series with the UDC input 2.

In normal operation the discrete signal to the UDC DI that signals switchover-to-UDC control is closed so the UDC's control output tracks the primary controller's output, which feeds the final control element.

At switchover, the discrete signal opens, the output starts at its last setting and the UDC's PID takes over for bumpless transfer.

If I recall properly, the engineer who implemented TRACK wanted to switch via relay between PLC output and UDC output to the valve,
(rather than let the UDC drive the final control element with the PLCs output value). He found that he had couldn't leave the UDC 4-20 current output open in normal PLC mode, because the controller detected the open output and went into a some sort of failure mode, so he had to allow the UDC current output to flow through a resistor when it wasn't connected to the valve.

Windup didn't happen with any reasonable degree of primary PID control because the PV was responding to control output. He did not use the inhibit reset action with the tracking action.

Some time ago I used Moore 352 process controllers (now Siemens, and now 353) but I can't recall exactly the function blocks used, and I lost my notes and manuals when I got 'rightsized'.

Bud
 
Top