Detect Fail

Detect fail systems.

Normally the automation systems , they need have sensors to operation or detect problems.

In machines sequences, use:

1) Watch-dog Timer: Timer to determine the maximum time of an action (time-out)
2)Conditional logic: Is not exact, depends on the moment, can be
valid in a space of time. (or not possible existence)

It is not perfect detect, but of common use.

Question: Exists which tecniques to identify a fails or stopped?

The simple answer to your question is...how much money do you want to spend and how important is it to detect the fail condition.

For example: On a fan supplying air to a duct you might:

1. A run confirm feedback from the motor controller.

2. A motor current sensor to detect power in the motor.

3. A rotation sensor on the motor shaft.

4. A pressure sensor in the air supply duct.



Each of these methods will detect the a failure of the fan. The last method is the best, but it is the most expensive.


The first method is very cheap, but it does not detect if the motor has for example lost it's fan blades, or it's inlet is blocked.

Which type of method you choose depends on how critical the failure is and on how much money is you have to spend.


This is only an example, but I think it is generally true for most failure detection problems.

Latch some BIT to keep them "alive" for a longer time so you can "READ" them

Good alternatives, Pierre. But it is not simple, as understood

But the question, if possible:

Restrition: In The level logic, without additional hardware

Question: (this one question, writing differetiate)

I Eat defined, mount a condition to detect an abnormality of operation.

OR

As it evaluate a situation, that does not possess elements of detection fails.

OR

As it compose a condition, that sinalize a problem, indeterminate, of unknown cause.

Coments: Simple write the program to perform the operation, but difficult defined, the events that results in you fail..

The question is especifc, but of answer expand

[My thoughts are about, events in lines of manufatures]

Woooo! You will get an A on Effort but your english is worst than mine.

Whenever you want to detect something you must have conditions set for the detection to take place.

For instance, whyle a motor IS RUNNING you can get a pulse generated by one of the conditions keeping the running rung latched.

Since this rung will unlatch only on the next scan, the BIT causing this unlatching of the rung can be used with the rung output to SET another BIT. This would then give you a latched BIT indicating the source of the fault.

If you monitor this or use a SCADA or HMI to get an alarm message out you will not get it because one scan will not be long enough to READ the BIT condition.

You can even detect a failure whyle the electrical power to your PLC goes out.

You have to SET conditions. Its that simple.

You need no hardware to do this. Of course if you have nothing connected to the PLC then you have nothing to detect.