Instrument Control Loop

W

Thread Starter

Waqar Ahmed

In a process control loop we have three parts,
1) Sensor
2)Controller(DCS, PLC etc)
3)Control Valve

I want to ask if this loop is on Auto mode and we want to attend hunting problem in this control loop then how we will check it?
Please also tell me about standard which we will use?

 
On any live process plant this procedure can be followed directly or be adapted slightly to suit your needs.

In some industries you will have two or more instruments doing the same measurement like an ESD and DCS or a control and trip transmitter and so on, but the basics stays the same in that you need to make the loop or instrument safe before you work on it. This means that your control room operator must be aware of what you plan to do, you must have the necessary permits and isolation certificates signed and in place. This means that the control room permit controller must have issued you with the permit to go and do the job. From his side he should check that the inhibits on the trip systems have been applied by the control room operator and might also allocate a dedicated field operator to you to control the process manually while you are working on your transmitter or control valve. If it is a control loop you will be working on then one of the first things you will do when on site will be to contact the control room and ask them to put your controller on manual. If it is a critical control valve application you might have a field process operator with you that will control a by-pass valve manually while you repair the control valve. If it is a trip transmitter you will be working on you have to ensure all the outputs from this transmitter is inhibited before you work on it.

This then basically answers your question on how to work on a control valve on a live plant. Short answer is to ensure that your control loop is on manual first and that you have the go ahead from the control room.

Some useful information could be that If you are unsure what the instrument does or how the control loop works find the loop drawings, P&ID and cause and effect drawings first. You might also need the original internal vessel drawing to have a look at the original calibration settings that was allocated by the process or design engineers in order to set your level transmitter up accordingly. Each job is an investigation by itself and it is expected of you to get all the correct information on the permit to work and the isolation certificates that you present to management for signature and approval. If not everything will get back to you as the ultimate responsible person so make sure you have the correct information first before you even apply for a permit to do the work. A Good source of information is the DCS itself as well as the control room operators so go and have look if the information you need is on the DCS already or talk to and get some info from the CCR operator and only then go and find the drawings if you need more info. Doing it like that speed the process up most of the time and after a while it becomes easy and is done in a matter of minutes.
 
If the process was previously running normal without hunting or instability, then something has changed or is causing a disruption in the control loop response. You will need to investigate to find the source of the problem. Sometimes the problem is caused by a valve that is not moving properly in response to controller output signal. The transmitter can be checked to see if it is suffering from any calibration, linearity or deadband issues or is simply faulty. Sometimes process can clog or buildup around areas where sensors are contacting the process which can slow detection of changes in the process. This may require pulling or cleaning the components or internals somehow. The effect of the slowed sensing of process changes will alter the tuning of the control loop, and can lead to instability.

As for the question of what standard to follow, I would suggest follow the SOP's for the particular plant/unit you are working in, as there is variation and no universal standard. But in general, as the other gentleman mentioned, you would start with communicating with the operations folks to review the problem and make some basic research and understanding of the issue. Then get a permit to do the work. Coordinate with operations before during and afterwards. Make sure you follow appropriate LOTO requirements, block and bleed, wear appropriate safety gear, use the right tools, etc. Be sure you are qualified to do the job or task. If you are not properly trained to do something, get someone who is.

Whenever you are working on live process it is significantly more hazardous to life and property.
 
H

Hans H. Eder

As posted by dave, something must have changed. Besides hardware (valve) problems, there could be a change in the process behavior or a disturbance variable has become active (e.g. pressure swings in a line, feed temperature fluctuations etc.).

In the first case we need to concentrate on the controller, re-tune it. In the second case we do not have to worry about the controller but need to locate and cure that disturbance.

To find out what the situation is we can:

- take the controller in manual mode. If the PV becomes quiet, then we need to work on the loop (controller, valve, ..), else if the situation gets worse, then the controller is doing a good job and we need to hunt down the disturbance.

- look at the auto-correlation of the PV (available in some tools as in our TOPAS). It gives a good indication if the controller is properly working or not.

Trouble-shooting can be very time consuming (and thus costly) and frustrating. Therefore we devote quite some time in our training courses to this topic to enable the participants to approach problems in a systematic, fast and safe way. For details see http://www.act-control.com/training.html
 
Top