Concerns regarding PLC Analog Process Control

To date most of my programming experience has been with DCS systems and some limited amount with PLC ladder logic (digital only).

The attraction of using PLCs for analog control loops is mainly the lower hardware cost of PLCs when compared to DCS. However, I am concerned that it may be more difficult to program analog control loops on PLCs. I have searched the threads in previous forums, and some of the issues discussed were very interesting.

However with my limited knowledge of PLCs and SCADA systems, would sonebody like to advise on the following points:

1. For PID loops, is it possible to achieve a bumpless transfer from Auto to Manual, and viceversa ? Is this achieved in the HMI or a SCADA unit ?

2. Is it "easy" to program a Cascade Controller (ie with Remote Set Point), and having Remote/Local Auto/Manual transfers. Is this done with preconfigured function blocks ?

3. Can alarm set points on analog input signals be set in the PLC as well as in an HMI ?

4. Can Hi and Low analog signal selectors be programmed in the ladder logic program ?

5. Is short term trending (for loop tuning) done in the HMI?

6. How important is it to consider "loop update time" on a PID loop for control performance, for example on relatively slow processes like tank level control ? Also, what is meant by the term "PID trigger rate" ?

Thanks,
Bert

1. It is an option on quite a few PLCs.

2. No comment.

3. No problem for PLC/HMI. Just write to a variable.

4. Do you mean for scaling? Is so, yes. Just use a variable to scale the high/low limits off the HMI.

5. Can be. Not overly difficult.

6. You'll need to sample your process roughly 30x per ossicalition. Not a problem at all on slower changing processes like tank level. With quicker things, there are inherent signal delays associated with PLCs & scan times, but you'd need to be looking at well under 10 mSec to run into this on most PLCs.

I generally use Siemens S7 PLC with Citect SCADA.
For that system, the answers are:
1, This is possible and is a feature of the S7 PID block, not of the SCADA system.
2, It can be done, easy is a relative term.
3, Yes, although in that case they are generally only set in the PLC, with the SCADA system reading the value off the PLC for its own display and logging purpose.
4, I am not sure what you mean, but you can usually do most common analog functions within the PLC.
5, I generally do my short term trending within Citect, can do this down to 100ms, and citect has built in popups to allow loop control, adjustment and trending.
6, It is generally important, although some systems will work well within a large range of values and some won't. Generally, Loops use up a lot of resources and will slow down the scan time. This may be important, or maybe not, it depends on the process being controlled. You may need to employ strategies like spreading out the loops so that only a few are solved each scan, or adding interrupts, or fixing the scan time. The solution depends on the application.

4. Do you mean for scaling? Is so, yes. Just use a variable to scale the high/low limits off the HMI.

Click to expand...

Thanks for your input CroCop. Often on interacting control loops there is a need to select the higher (or lower) of two analog signals, such as on an "overide" control system. The analog output, of say the high selector, would then be sent to actuate a control valve. In such cases can the high signal selector be programmed in the ladder logic ?

Bert

Yes. Very easily. Run a compare between the 2, and move 1 low and the other high. Then have a binary status that chooses low or high. That's no big deal for a PLC (assuming you don't buy the absolute cheapest model on the market).

The general answer is yes, it can be done. Now whether it is worth the extra work to do it, and if it will perform well is a different question entirely.

DCS vs PLC

I have done quite a bit of programming on plc's and DCS systems.
Generally speaking anything that can be done with a DCS, can be done with most of the larger PLC systems. However, DCS systems come with all the bells and whistles built in, the PLC platforms require a lot of coding to accomplish the same task. In my opinion, the high hardware cost of the DCS is offset by the increased programming time need for the PLC.

If you already have and are familiar with the DCS system, for new projects I would expand the existing system. Depending on the application you could also use a PLC for small skid type equipment and network it into the DCS system, then the PLC would handle local control, and the DCS could perform alarming, historizing, itegration with the rest of the plant, etc...

If you are dealing with a new installation that does not have an existing DCS, then when making your selection, you must include application development as one of your cost factors.

If this is something you are going to do "in house", then your time is already paid for and using a PLC may be the best choice because of less hardware cost, but you will have to decide if you have enough "spare" time to program the PLC with all the DCS like functions.

If you have a DCS background it might be a good idea to pick a PLC which has a program editor that is aimed at process control. Function block programming is best for the type of control you mention. Siemens S7, Allen Bradley ControlLogix, Mitsubishi Q series (QxxPH version), all feature function block programming. The best though, is an old DOS program called APT which is used to program Simatic TI PLCs (message to Siemens: you let a good thing go to waste).

If you are dealing with a new installation that does not have an existing DCS, then when making your selection, you must include application development as one of your cost factors.

Click to expand...

Yes, the programming development cost is a concern, and also the ability to tune the PID loops, seems to favor the DCS. However, the fact of the project is that it is a new installation and there are relatively few analog I/O.

The IO count is approximately: 20 AI, 8 AO, 40 DI, and 25 DO.(Includes 8 PIDs) This is a relatively small count for a DCS, but my gut feeling is that the configuration time will be less costly using a DCS, especially when you consider integral features of the Operator interface, graphic design, and ease of configuring PIDs. I am thinking in terms of using a Delta V.

For this small size of installation, however, a PLC is probably more cost effective ... any opinions ?? But having to learn another programming language for the SCADA seems to be a disadvantage.

If you have a DCS background it might be a good idea to pick a PLC which has a program editor that is aimed at process control. Function block programming is best for the type of control you mention. Siemens S7, Allen Bradley ControlLogix, Mitsubishi Q series (QxxPH version), all feature function block programming.

Click to expand...

Yes, if function block programming is available on PLCs this would be very attractive. Thanks.

You should take a good long look at ControlLogix and its function block programming. All the features you've listed are there. I think you would be quite comfortable in that environment. The same processor can include ladder and SFC programs and all types can be edited in run mode.

For SCADA: RSView32 includes pre-built faceplates designed to work with the PID function blocks. And the developers (Dynapro) are fairly handy to you - although they've migrated from Kits to Marpole to Annacis Island.

Thanks Gerry, I'll certainly have a good look at ControlLogix and RSView32.