Have you ever connected dual element Thermocouple in parallel in the measurement loop for redundancy?

Can we connect duel element thermocouples in parallel so as not to trigger fault upon one failure ? Normally dual element thermocouples we connect the loop with one element and other just keep it as stand-by , upon failure of the connected sensor we take up an urgent maintenance activity to connect the redundant good sensor to bring back the measurement loop back ! Wondering if we can connect both in parallel upfront ?

1603259268654.png



Thermocouples are widely used in the temperature measurement of power plants, refineries, petrochemical and other oil & gas applications. There are umpteen number of temperature measurement loops which are very critical for the process and safety systems.

Say for instance if you use Thermocouple in a critical PID control loop the failure of Thermoscope could lead to control loop failure (I know that systems come with 2 out of 3 Thermocouple elements these days ; assume you have one dual element Thermocouple) or in the case of protection (safety) system the controller trigger protective action as a result of measurement failure.

Its well known that systems comes with fault tolerant measurement loops such as 2OO3 measurement loops in the modern instrumentation systems, even in those cases we have to simulate the other loops while doing the live maintenance on the faulty loops so as not to trigger one more loop faulty. Even with the 2OO3 redundant measurement loop, Process engineers feel uneasy when you have one measurement failure; unless you rectify the issue immediately your reliability dependence goes back to only one transmitter (NO I don’t think anyone have ever implemented 1 out of 3 logic). There are certain process systems whereby its not very safe to work on the live process with high temperature, high pressure, high speed, humid and super noisy conditions.

I was wondering if we can connect dual element Thermocouple in parallel so that if one fails other keep the measurement healthy by that we can avoid the costly shutdown, emergency maintenance at unsafe conditions and odd hour maintenance?

I am sure there were enough amount of R&D and thoughts went through on this topic but at the end user level there is no application whereby we use parallel thermocouple for redundancy purpose. I really wonder what stops this parallel connection in process plants? is it something to do with reliability? or is it something to do with impedance of measurement loop ? or is it something to do with high current with the 2 thermocouple ?

Any idea how the exhaust thermocouple (EGT) connected in the jet engines? I suppose one Thermocouple failure cannot trigger the emergency landing.

Appreciate if you can put in your thoughts, experience, and ideas!
 
If you have critical processes, redundancy helps by having another
resource to duplicate the first.
Connecting 2 thermocouple in parallel is not dual redundant as your
measuring amplifier becomes a single element.

I've never connected or seen 2 thermocouple in parallel. I would imagine
given the signal level and impedances in circuit would give rise to serious
calibration errors.

For 2oo3 voting, I've used 3 thermocouple and triple mv(millivolt) amplifiers
to suit, with cabling taking separate routes back to the safety system.
 
If you have critical processes, redundancy helps by having another
resource to duplicate the first.
Connecting 2 thermocouple in parallel is not dual redundant as your
measuring amplifier becomes a single element.

I've never connected or seen 2 thermocouple in parallel. I would imagine
given the signal level and impedances in circuit would give rise to serious
calibration errors.

For 2oo3 voting, I've used 3 thermocouple and triple mv(millivolt) amplifiers
to suit, with cabling taking separate routes back to the safety system.
Hi oneye14 ,

Thank you for the insights. You are correct that still we have few components in the loop such as cable , amplifier and Analog input card or Temperature transmitter etc which are solo in nature without any redundancy.
I have personally checked N type and K type duel thermocouples, in which both the duel elements are with the same resistance and generated voltages. Need to further study and try to experiment with a very non critical system and check. Thank you
 
Thermometry references will state that paralleled thermocouples will report a value of the average of the number of thermcouples.
A problem in practice with paralleling thermocouples is that since both thermcouples are presumably measuring the same location or point, both thermocouples are exposed to the same process conditions. The process conditions, heat, medium, and atmosphere, that can cause the failure of one thermocouple is common to both so the process conditions effect both thermocouple elements. A failure of one thermocouple is highly likely to be followed by the failure of the 2nd thermocouple because of common exposure.

Burnout failure of a thermcouple in one sense is actually a 'good' situation because all modern analog inputs can detect a failed thermocouple open circuit and then signal a fault condition.

The more serious problem with thermocouples is drift because most analog inputs can not detect thermocouple drift. A drifting thermocouple continues to output a mV signal, but the mV output no longer matches the standardized output of a particular type thermocouple because its junction is polluted; the junction is no longer a pure junction of the 2 dissimilar metals only, so the resulting indicated temperature is incorrect. But the analog input doesn't know that, it just sees a mV signal, converts it to a temperature reading and continues to report a drifted thermocouple temperature value as though it were normal. In a paralleled situation, the shift in average would be decreased by the averaging effect.
 
Top