结构化文本与传统编程语言

Question

I have started Programming Logic Controller (PLC) using Structured Text. My Background is from IT or computer science. So I am used to traditional Programming languages like Java, C# etc... I just realize that PLC requires a different Mind set. Understanding scan cycle... Can someone explain to me more clearer the following paragraph I read from

One important difference between ST and traditional programming languages is the nature of program flow control. A ST program will be run from beginning to end many times each second. A traditional program should not reach the end until it is completely finished.

Answer 1

A PLC program is different from what you already know in that it typically has a main task, which is executed cyclically, ie once per scan cycle. The main task then calls a list of subtasks or -programs and when this is done, the main task is run again.

The scan cycle can be set to run at intervals, for example each 10ms. The PLC will then update the IO, run the program and then do nothing while waiting for the next trigger. If you have a fixed scan time, you can use this to make very precisely synchronized programs, but it is important that your program (main task and its subtasks) always execute within this time.

Some PLCs can be setup to run the scan cycles one after another, as fast as possible. The scan time is then variable, but you still have to monitor the scan times to make sure the program is running fairly consistent cycle times.

One of the most important lessons I have learned in the transition from other languages to PLC language, is to keep the program flow as sequential as possible, since any loops and deviations influence the execution time which must always be less than the scan time.

Answer 2

This means eg: during 1 sec your PLC program is executed 1sec/cycletime. So if you have a cycletime of eg 1ms, your program is executed 1000 times in a second. Thats the main difference. It is deterministically. This is what the paragraph means. And it is not finished when it reaches the end. it restarts with an updated process image.

In, lets say, non-PLC programs (like on your pc) you have something like a

 while(!_end)
 {
      /* your code */
 }

In principle, the PLC works similar to this. But on your PC there is no need to execute programs cyclically in hard time frames like eg 500µs. Most of the applications for your PC are ,lets call it, 'time uncritical'. The big difference is, that the PLC runs the full program and before it starts the next run it refreshs its process image (writing outputs / reading inputs). The program execution is triggered by a so called 'task'. This task is responsible for keeping your control in realtime. The process image, which I mentioned before, is the sum of all process values which are present for the PLC (like digital I/Os, motor driver values ,...). So you see, it is a kind of loop, but not the same like you know it from your c#/java-programs.

(btw, i couldnt open your pdf)

Answer 3

pboedker and ChrisKo answers are both good. A couple of extra warnings are needed.

A scan (or the scan time) can be shorter than your code's execution time. In most PLC's there is a "watchdog" to detect this, and warn you that it is happening. You often need to set this "watchdog" up, and set up a alarm/event handler for it.

Know how your PLC executes I/O. Some (Such as ControlLogix) are Asynchronous, I/O gets read into your controllers memory based on the RPI (request packet interval), and gets written out (same RPI) when you make a change to the I/O point with code. Others (Such as AutomationDirect's Productivity 2000 series) only write the outputs upon a COMPLETE scan of your code. I'm sure you can picture the pro's and cons of each scenario, especially if your controller is not completing your code before restarting a scan.

PLC's that allow you to have different scan rates for different pieces of code give you powerful tools for I/O and program flow management. Slow processes can be scanned slowly, allowing you more PLC time for other things.

Answer 4

I'm a PLC programmer. and I want to use PC to control PLC, PLC as a dumb IO. PLC programming is always cycling and scanning.And each scan time is several mili-sec. You can understand it as while statement which ChrisKo said.

while(1=1)
{
    // code
}