Process Puzzler: Make Solid Progress

Readers suggest ways to improve crystallizer controls.

1 of 2 < 1 | 2 View on one page

We recover ammonium sulfate from a power plant scrubber and sell it as fertilizer. Temperature control of the liquor going to the crystallizer is critical; liquor is circulated through a shell-and-tube heater. Poor control causes nucleation, creating small crystals at the expense of large ones. The difference between the inlet and outlet stream temperatures must be ±1°F. Our customer originally insisted on using a controller temperature board in its central control room, which is more than 300 yards away, but now is demanding that we use K thermocouples. Will either idea work? What's the best way to get the accuracy needed?

As byproduct recovery is involved, minimum expenditure is important. In my view, the ±1°F will not promote crystallization in the mother liquor. For fertilizer use, crystal size should not be critical unless packaging of the final product is a factor.

One common processing technique is to concentrate the liquor to a specific gravity and add seed crystal plus air shock, in combination with slow cooling. This will promote the desired large crystal growth. At end-point temperature, the vessel is discharged to a centrifuge for separation of the product; the mother liquor is recycled. This process is a little more complicated if continuous instead of batch. Drying and sizing over screens, e.g., for a 20–40 mesh cut, will obtain a uniform free-flowing product of the proper size. Another possible alternative would be to market a concentrated ammonium sulfate solution for use directly as fertilizer.
Robert Drucker, consultant
East Northport, N.Y.

This is an instrumentation problem as opposed to a crystallization/process problem. As observed in the plant, variations of greater than ±1°F can result in fines due to excessive flashing and nucleation in the crystallizer circuit. This is observed in many continuous systems since they are designed for a specific supersaturation level to stay within the metastable zone and thereby avoid undesired fines formation due to primary nucleation.

Assuming that their current process involves sending an ohm signal to the board in the control room 300 yards away, the signal is degrading due to line losses and electrical flux interferences from motors and power lines. The result is that the signal will not be accurate enough to maintain the desired and necessary 1° range. This would be a very old system.

Regardless of the temperature measuring device in the field, be it RTD or thermocouple, the field sensor should send a 4-to-20 ma signal to the control room to achieve the required accuracy when transmitting over such a long distance of 300 yards. This type of signal should not degrade.

An RTD or thermocouple should work. If a thermocouple, J or K is acceptable, but Type J is my personal preference.
Dr. Wayne Genck, crystallization consultant
Genck International, Park Forest, Ill.

Although using K-thermocouples throughout the plant may make inventory easier it won't solve your temperature control problem. Three-hundred yards is way too far for cables without a signal booster.

The only instrument that I would propose would be a four-wire resistance temperature detector (RTD). A three-wire RTD won't do better than ±2–5°F, as crystallizer manufacturers have admitted to me. This puzzle poses a more difficult challenge.

Loop control is based on a differential of two temperatures. Anytime a measurement is compared to another you introduce three potential errors: the elements, the transmitters and the interactions between them. This is why material balances and pressure drop calculations made using different instruments are often wrong.

There are two solutions: 1) a single transmitter for each measurement; or 2) install a differential transmitter. In the first case, the distributed control system would use a calculation block — which could be another source of error. In the second case, the differential transmitter could avoid transmitter error by taking signals from two different elements. Another reason for the four-wire RTD is compensation for corrosion or poor connections. Liptak's "Process Measurement and Analysis," Vol. 1, p. 574, states: "A single ohm of difference between the legs is reflected as a 4.7°F error."

1 of 2 < 1 | 2 View on one page

Join the discussion

We welcome your thoughtful comments. Please comply with our Community rules.
All comments will display your user name.

Want to participate in the discussion?

Register for free

Log in for complete access.


No one has commented on this page yet.

RSS feed for comments on this page | RSS feed for all comments