Correlation flowmeters are typically applied to fluids in the turbulent flow regime and other fluid flows with coherent disturbances, such as slurries.

Pressure Sensor Array Technique
A multiple pressure sensor array attached to the outside of the pipe is used to measure pressure gradients caused by fluid vortices that travel down the pipe as a natural consequence of the fluid flow. These vortices are passively sensed at each pressure sensor. Computational techniques derived from sonar technology are used to determine the velocity of the fluid. This technology can also be used to determine the speed of sound of the fluid in the pipe.

Ultrasonic Technique
An ultrasonic transmitter/receiver pair is located at a fixed distance downstream of another ultrasonic transmitter/receiver pair. Their respective ultrasonic beams are used to actively sense turbulent eddies in the pipe. Cross-correlation computational techniques are used to determine the velocity of the fluid.

Excerpted from The Consumer Guide to Ultrasonic and Correlation Flowmeters.

When auditing, one should investigate as far back in the construction and design process as possible in as much detail as possible.  The flowmeter in question was insulated and located approximately 5 meters above grade so the ability to obtain nameplate information was limited.  Information about the flow element was limited to a memo with a certified flow calculation attached plus an instruction manual with a hand-written serial number written on its cover. 

The manufacturer's representative and the writer of the memo were contacted for information.  After much digging by the manufacturer, two certified calculations plus another subsequent calculation were found.  Interestingly, the inside diameters were different on each calculation.  Put differently, the information describing the physical attributes of the installed flowmeter was unreliable. 

The mystery was solved with a simple phone call.  After some discussion, the technician onsite confirmed that he had personally hand-written the serial number on the instruction manual directly from the nameplate prior to installation when the flowmeter was on the ground.  This is as close to solid information as could be found and enabled the manufacturer to determine which "certified" calculation was correct.  It can take a bit of digging to get information --- but sometimes you just get lucky.

This article originally appeared in Flow Control magazine. 

1.   1 percent
2.   2 percent
3.   5 percent
4. 10 percent
5. None of the above

The 10 inch error at full scale flow is 10 percent of the full scale differential pressure.  This error represents an approximate 5 percent flow error at the full scale flow rate.  However, the process pressure will compress the bubble that is initially present at startup so Answer C and Answer D are not correct. 

At operating pressure, the non-condensable gas will be compressed by a factor of approximately [(4 bar +1 bar)/1 bar] or 5.  Therefore, the height of the bubble will be approximately 10/5 inches, or 2 inches.  A 2 inch error is 2 percent of the full scale differential pressure and represents approximately 1 percent flow error at the full scale flow rate.  Therefore, Answer B is not correct. 

The approximate 2 inch water column error associated with the bubble is independent of the operating flow rate.  By way of example, the error associated with the bubble is 2 inches of water column when the flow is the 10 percent or 100 percent of full scale flow.  At 100 percent of full scale flow, the flow error is approximately 1 percent.  However, at 10 percent of flow, the flow error attributable to the bubble (2 inches of water column) is larger than the differential pressure measurement (1 inch of water column).  In other words, the flow measurement error varies with the operating flow rate so Answer A is not correct.  Answer E is correct (except at 100 percent of full scale flow).

Additional Complicating Factors

The above analysis presumes that the bubble height is a function of starting up the impulse tubing from an empty condition.  In operation, non-condensable gas can be bled from the system and additional non-condensable gas can accumulate in the bubble.  Both of these mechanisms can present additional measurement error.  

This article originally appeared in Flow Control magazine.