Wednesday, June 4, 2014

Process Wednesday: how to measure reactor volume? How about radar?

Another short snippet from Derek Walker's very informative The Management of Chemical Process Development in the Pharmaceutical Industry, this time about reactor volume measuring technology: 
There are also a great number of both simple and sophisticated level measurement devices...  
Ultrasonic level measurement depends on a transducer sending an ultrasonic pulse to the liquid surface which is reflected back to the transducer. Electronics convert the ultrasonic lag time into a distance corresponding to depth... Although ultrasonic level indicators require little maintenance and are unaffected by the nature of the liquid (acidity/basicity, dielectric constant, or specific gravity), they cannot be used in circumstances of excessive foaming or turbulence or in high-temperature situations where stratified vapor layers are present.  
On the other hand, radar detectors can be used in vessels exhibiting a wide variety of conditions: Radar energy passes through air or vapor space with imperceptible changes. The greatest limitation to the use of radar detectors is price, because of the complexity of the microwave and timing circuitry. In an advanced measurement unit, the radar level gauge sends out a continually swept microwave signal with varying frequency. The transmitted signal is compared with the signal sent back in a microwave mixer device, the difference in frequency being directly proportional to the distance. It will be obvious that high-quality signal processing is necessary since the liquid surface in a working reactor is never calm and unwanted echoes within the reactor need to be separated out.
I've never had the opportunity to use either ultrasonic or radar measurement of levels in the reactor; the reactors that I have worked with have had pre-measured landmarks that are quite accurate in diagnosing the volume of solvent in the reactor. That said, I can imagine instances (crystallization, perhaps) in which knowing the exact amount of solvent is necessary. (Also, if there's a good bit of solvent vapor in the reactor, it can be difficult to see!)

(Incidentally, a quick Google search indicates the use of radar to track volumes of cooling water in the nuclear power industry. Yeah, that makes sense.)


  1. Really the only time I can think ever want precise volume info is when intend to add more liquid and need to know if might exceed the vessel limits. In my experience most often charge done by weights or accurate metered pumps thus have good idea of what volumes should be plus hopefully have a detailed profile of the true reactor capacity but on occasion have come close to overflow values when pushing extractions. Perhaps if do have to monitor large waste tanks for routine reporting it could be helpful?

    1. Ahhhhhhhh, "true reactor capacity"! I love it. Very good point.

    2. Level indicators are actually very useful for determining the mass of an analyte (most commonly product). Volume, density, and w/w% assay allow you to determine mass of your analyte (V * D * w/w% assay = mass analyte), which is a great piece of data to have as you move through your process to track the mass balance beyond just analyte losses to waste streams. You can certainly do the same calculation using the mass of the solution in the reactor, but that usually means dropping the solution to drums to get a weight. I've heard some places have reactors on balance systems, but unfortunately have never worked with one.

    3. I have heard that reactors on balance systems are unreliable to the point that they don't do much good.

  2. In my little experience, we've found landmarks to be much more trustworthy than radar for things like determining the endpoint of a distillation. Radar is useful in order to tell you when to stop watching the radar and start looking at the landmarks. You could also theoretically use a mass flow meter to measure the mass while transferring from one tank to a second, but this also hasn't been reliably accurate in my (again, not a lot) experience.

    Also, if your process doesn't have a wide range on the level of the mixture which must be measured in the tank, then it's not exactly a very robust process.

  3. Awesome blog. I really enjoyed to read your useful blog content. do radar detectors work