measuring viscosity in process

Why measure viscosity? Is there another parameter that could be measured more easily, to indicate the progress of a reaction? Temperature? Colour?

Is there an easier way of doing the mixing, perhaps in a way that viscosity didn't need to be measured - a static mixer with the reagents' inlet flows being controlled, for example?

I've considered temperature, however, ambient temperature varies greatly due to living in Texas. I would have to find a way to compensate for ambient temperature which I fear could be complicated. Color change might be an option. However, viscosity, I think, is the best indicator.

The reaction is a neutralization reaction. I haven't thought about using a Ph metering device for feedback...Hmmm...you have got me thinking...

pH is far easier to measure than viscosity, and most equipment has temperature compensation built-in.

With a solution of high viscosity, would I still get reliable results with A Ph probe? I would think that a pocket of caustic that hasn't yet mixed in thoroughly might give erroneous readings and trigger the next event too early. Do you know of any good reading on the subject of process controls?

It's difficult to state, reading so little about the nature of the reaction to date other than one of the reagents is Caustic Soda.

All right...all right...I'll fess up...I'm manufacturing soap. I'm making it by hand right now. If I automate, I can increase production. The reaction (if you're not already familiar) is nuetralization of a triglyceride with NaOH. I'd like for my automation process to be adjustable to accomidate different mixtures of oils. My primary oil is olive. I'm going to experiment with some other oils. Different types of oils react at different rates.

When processing by hand, the determining factor of reaction progress is viscosity...beacuse it's easy to see.

I've read a lot about industrial scale soap production, but I'm only looking to supply a locality rather than than the entire nation (at this point).

Hi,

if you mix two different liquids (?) you may get a completely differnt behaviour of the product.

This is most often described as viscosity being dependent on shear rate.

For pure liquids this is a linear function depending a lot on temperature!

(Mineral oil has a factor of 10 less viscosity at 100°C up!)

Your idea with the motor and some paddle is a very good one.

If you switch the paddle to a disk or cone (deviating from a flat disk by 3 to 10 degrees and let this run at a very defined gap at stabilised temperature you are near a good measurement.

If you measure your motors bearing torque at slow velocity and if this is much down of your measured torque that you translate to vicosity then you are still better.

That is why good viscosimeter use airbearings: not to introduce measurement errors by bearing torque.

If you use a PMDC motor then you can directly translate current to torque.

Maybe vibrating (in plane or perpendicular) plate is a good idea. This was once tried for vacuum measurement where torque is proportional to pressure (if molecular flow exists).

RHABE

I like the idea. It sounds great for my shoe-string budget and the junk I have laying around. One question however, is some clarification. What did you mean by,

"a very defined gap."

I think I have the idea, but I'd like to make sure...

Hi,

if you have a wide gap you will measure turbulent flow and this is not at all what you want to have.

If you choose a small gap - maximum Reynolds number should be 40 - you will measure laminar flow. Only with laminar flow torque is proportional to viscosity.

As torque is dependent also on the gap width it is necessary to have a conszant gap.

And a constant velocity of the rotating cone.

And a constant temperature.

And a constant torque in the motors bearings, seals etc.

That is why I recommend a separate small PMDC motor - omly these are suited for torque measurements.

You should try to specify the needed accuracy!

Try and you will see where to improve.

RHABE

Thanks Rhabe,

It looks like I need to do some experimenting these next few weeks and determine my needed accuracy and to follow up on PW's suggestion and determine change of Ph vs Change in time. Perhaps combining viscosity and Ph measurements would help even more in determining the progress in the reaction.

For in-process viscosity, you can either purchase, or adapt a small mixer motor, or possibly the main mixer motor, such that it sends a continuous signal to a recorder or a computer, wherein as the viscosity increases, the load on the motor increases and it draws more current. Calibrate the set up, and with a curve in hand, you can instantly read a viscosity. It shouldn't cost a lot if you can do it yourself. The industrial in-process viscometers are pretty pricy if you're paying out of pocket.

Thanks Cardio. I was thinking I could determine viscosity via the main mixer motor. I'm wondering if anyone has done this and found it effective. Would motor temperature effect my current readings and thus throw off my viscosity measurement? Could current measurements be used from an AC motor effectively or would a PMDC be a better choice? I lack experience on motor characteristics at this depth...

Yes, if you could easily measure the the load to a motor w/paddles you could determine the viscosity. You could also develop your own version of a Saybolt Universal Seconds (SUS) test where you measure the time it takes to travel X distance. Instead of dripping with use of gravity, create a baseline at some set pressure and measure the volume or weight at a set amount of time to determine the change in viscosity.

I've personally worked on a design for Detroit Diesel that was later shopped to Mercedes with Lucas Engineering taking over the design work on it. The inventor used a piezo crystal with 2 sets of capacitor plates screened on both sides of the crystal. One was trapped in a vacuum for reference and the other exposed to the viscosity of oil. Depending on the thickness, the oscillations would vary in amplitude. Very creative design, but costly to produce and the quality sensor we were developing for them was sufficient for their application at a greatly reduced cost.

One peizo might work for me. I'm looking for change in viscosity with respect to time (this is starting to sound like a calculus lesson!). So if I could use the original viscosity as an initial condition and compare this to a later reading, it might just work. That might be complicated on the control end of things, but might be more consistent with a change in temperature.

How would you go about reading the feedback from the peizo oscillations?

By the way, I really want to thank everyone for their input so far. It has been incredibly helpful!

It was back in the mid 90s, so I'm having trouble at the moment remembering the exact circuit, but I can help with the general concept.

The circuit used a Voltage Controlled Oscillator (VCO) chip set at the frequency of the piezo. You need to create a sweep width range in the circuit which is what I'm fuzzing on at the moment. It is needed in order to detect the signal. There is also a tank circuit which attached off of the VCO output. Which of the 2 came first I can't remember, but after that the signal is fed into an op amp circuit to give you a usable output.

In the design I was working with, the VCO was set at 3Mhz with a sweep range of +/- 200 khz. Since your working with ambient temps, the capacitor used to set the frequency on the VCO can be a standard ceramic. Due to the temps I was testing with, I had to find a capacitor with the opposite temp characteristics of the VCO in order to keep the main signal within the sweep range.

Fantastic! I will check some of my resources and see how it might go together.