Dear

I'm working at petro-chemical process. We monitor reactor temperature using thermocouple 'k' type.

It monitor the temperature of reactor inside which are normally 400 degC. Under normal operation, the movement of temperature signal range is very narrow.

But when it rains, temperature values are increase around 2 degC. When rain stopped, the value is return (decrease) to its normal temperature slowly.

There are 20 thermocouples at reactor.
But, above strange phenomenons are occurred randomly.
(Anyway, temperature increase always at some T/C when it rains)

For example, at one rainy days T/C A,C,F,L showed temperature increasing, but other rainy days T/C A,G,R,V showed temperature increasing.

I've checked wiring and connection already to find out no problem.

What can cause this strange thing ?

Hope your sharing ideas.

 

I’d be surprised if the thermocouple was not grounded.
Here is my Columbo style logic...

There are very few things that will make a thermocouple read high. Nuclear radiation is the only thing I have seen that is able to cause a high reading.

You say the problem is intermittent and occurs only when it rains. The problem goes unpredictably from one sensor to another. Electricity is unpredictable and can go through strange paths when there is water present. I know this because I never have problems with my land line telephone at home except when it rains. So, let’s assume your problem is electrical noise. The TC circuit would not pick up extraneous common mode, crosstalk or static noise if you used ungrounded thermocouples or isolated transmitters. A second problem may be that you are getting moisture into the head, conduit or somewhere else along the circuit path that could cause the previously mentioned interferences oooorrrr...

Lastly, you could have a ground voltage differential that is passing through the wire shield (screen for you English guys) if you happen to be using shielded extension wire. The rule of thumb is to ground the shield at the thermocouple if the thermocouple is grounded. Ground it at the instrument if the thermocouple is ungrounded.

In either case, if the ungrounded end of a drain wire, or a portion of a bare shield is close to another metal away from the intended ground, water will cause an effective ground and viola, you have a ground path loop.

All that being said, if you are using grounded TC’s send them to me and we’ll unground them for you.

sensors [at] jms-se.com

 

A couple of things to check out next time it rains.

It is possible that the T/C's are actually seeing a rise in temperature in the reactor and what you are seeing is a change in pressure in the vessel due to the sudden cooling of the skin from the rain. This would be something that your pressure control loop has never been tuned for and as such does not respond to quickly, allowing the pressure in the vessel to drop and the temperature to rise. Depending on where the various T/Cs are located and where you are in the reactor process may explain the apparent randomness.

This very phenomenon triggered a flare event to atmosphere and caused a fire at a major petro=chem facility.

Follow the rule that if it doesn't make sense we don't understand it yet

 

Just check the compensation cable connection throughout the circuit. Connections may be reversed at one point.