Variation in Weight

Assuming that as you say the contents of the tanks has not changed it can only be that the behaviour of the load cell is changing with temperature,

I suggest you keep a record of the temperature and see how it co relates with the apparent change in weight and then have words with the manufactures of the load cells

As said, but also the temperature and properties of the resin can change as well!!!

John.

We have addressed importance of cardinal physics priinciples on CR4 before. The same applies here. In any chemical reaction, EXCEPT a nuclear reaction, the weight or mass of the products will always equal that of the materials used at the beginning of the reaction. Certainly, if you weigh ash after burning wood, it will weigh less than the wood did, but only because the mass of CO2 and water as combustion products had been lost to the atmosphere.

The same applies here. Unless your resin evaporates solvent through a venting system, the resin should not change weight overnight, despite any aging, unintended curing, or other changes which may take place.

More often than not, load cell based tank weighing systems suffer from torque moments introduced by piping, or by omission of mechanical provisions to isolate thermally induced expansion or contraction from the load cell or cells. A so-called diaphragm load cell will not respond to horizontal loading but only to vertical loading. That is accomplished by supporting the load bearing vertical force measurement member within the cell literally by a metal diaphragm at the top, which prevents lateral stress and strain from bending that member in the cell. Other types of load cells may not be fully compensated against lateral sensitivity, which may arise from horizontal force due to thermal expansion considerations.

There is also the possibility that the load cell or cells are defective. Try and establish a correlation between time, temperature, and indicated weight. You can buy inexpensive mechanical temperature chart recorders from companies like Omega, or use an approach based upon a PC, interface module or card and temperataure sensor used with the proper software.

If your system is based upon a single load cell, the other supports for the tank should be two only, such that all three are located more or less in an equilateral triangle configuration. The two non-measurement supports should comprise flexures, such that a very small amount of bending (in the low or sub-arc second range) can take place. Bearing saddles similar to those used to support bridge spans are sometimes used, but can suffer from "stick-slip" problems under rare conditions.

If each support has a load cell, the equipment used for signal conditioning sums them algebraically. In that case, torque introduced by piping stress will not matter much, because changes in loading distribution will be compensated between load cells. But vertical stress induced by piping will matter, and cannot easily be compensated for. Do you have flex joints in the piping?

Hope this helps.

Bernie Katz

Bernie:

An excellent reply ... but then you do know your load cells, among so many other things.

I still have in my possession (as a kind of souvenir) a set of the custom analog load cell/feed rate boards that you designed for us. Unfortunately the digital version you subsequently designed arrived after I had left or I would have definitely gotten hold of one of those once they went into production!

The loss of mass when energy is liberated is not limited to nuclear reactions, the law MC^2 applies to all energy although the effect is small and not easily measured.

Without getting into the twin paradox question, the apparent mass of a fast travelling object for a reference position observer increases by the root of one minus V squared over C squared. However, apparent mass from the standpoint of the travelling object itself remains the same.

Bernie Katz

Nearly everything Berniek and others have said is true there was one thing that isn't strictly true;

"We have addressed importance of cardinal physics priinciples on CR4 before. The same applies here. In any chemical reaction, EXCEPT a nuclear reaction, the weight or mass of the products will always equal that of the materials used at the beginning of the reaction."

The mass part is true, the problem is that weight is not a direct measure of mass. Weight is a measure of the effect of gravity on a mass and is compounded by things like buoyancy.

I will elaborate, A ship may have a mass of 20Gg for example but while it is floating in water it weighs nothing. This is because the weight is cancelled by the 20 Gg of water that it displaces. This can be stated mathematically as

Weight = g x (Mass_Object – Mass_Displaced)

Where

g = acceleration due to gravity

Mass_Object = mass of object

Mass_Displaced = mass of displaced substance

Now if you have an object that expands and contracts with temperature its volume will change so accordingly the volume of air it displaces will change and this will effect the buoyancy part of the equation. The question is could the volume change enough so that the mass of air it displaced changed by 20 to 80 lb.

This can be easily checked by calculating the volume of 20 to 80 lb of air and since the density of air is roughly 0.075 lbft^-3. That would give us a volume of 267 ft³ to 800 ft³ which is obviously not the case here so it must be one or a combination of the other factors.

The point I am trying to make however is that you can't always equate mass and weight, they are two quiet different things. Mass is a measure of an objects ability to resist a change in motion and is measured in Kg (Kilo grams) while weight is a really a measure of force and should be stated in N (Newtons). Unfortunately most people do not understand the difference and confuse the two. This is not helped by the fact that we use the effect of gravity to indirectly measure and objects mass.

Hi again. This is a reply to Masu. What you have said is true. However, I based my answer upon the problem at hand, in a practical sense. For example we all should know that W=M*G. The assumption was that gravity at the tank location was constant.

Certainly there are any number of real physical possibilities which enter into weight measurement. For example, there will be a very small amount of apparent weight change of the tank due to air density changes, since the tank displaces air. As an example, one pound mole of air in the English system is 259 cubic feet at standard temperature and pressure, while air has an average molecular weight of 29. So if the temperature changes from 70F to perhaps 40F, the change in density will be about 30/515, or about 5.8%. The figure "515" comes from the average temperature in that span (55F) referenced to absolute zero in the English system (-459F, or zero degrees Rankine). If the vessel displaces 259 cubic feet as an example, 5.8% of one pound mole of air (29 pounds) will represent the apparent change in vessel weight, in this case .058*29, or 1.68 pounds. The question concerned apparent weight changes of about 40 times that value.

Bernie Katz

Just an update to the former air displacement example: At 80000 pounds, specific gravity of the resin would need to be about 5 for a 259 cubic foot tank. If the specific gravity is 1, the tank size would be such that it occupies about 1300 cubic feet. The example change in air density would be 5 times the example given, 1.68 pounds, or 8.4 pounds, which remains a relatively small fraction of the total error.

Bernie Katz

Masu:

It is largely because of people like you, Syphrum, and Berniek that make CR4 so interesting to me.

We (if I may flatter myself by inclusion) are of an analytical nature, grounded in physics. These sometimes "hair-splitting" posts are not only intellectually stimulating, but serve as a reminder that while we all indulge in mental and verbal shortcuts (as the human mind is wont to do) we should not be oblivious to that fact, lest we overlook something that turns out to be significant in a particular case.

The weight/mass thing while very much confused by the general public can also befuddle us at times as a byproduct. We routinely use electronic balances (which are typically force balances rather than mass balances) to read out in kilograms, and, depending on the nature of our use, we go happily on our way because while we understand the difference, the context tells us it is not significant in that particular case. And therein lies the potential trap, laying in wait for the particular case where it could be significant.

I took Berniek's reference to the law of conservation of mass to be to the classic "idealized closed system" based on the fact he was addressing a mixed audience plus the fact that I had the pleasure of working with him some years past, on a load cell related project and I am sure he was aware of all the subtleties.

Likewise, when I point out that strictly your buoyancy equation is not correct either, even though I understood exactly what you meant and it was a good point.

The trap you fell into was in algebraically simplifying your equation by factoring out "g". Strictly speaking, "g" is not the same due to the relative displacement of the masses involved, since a change in volume of a liquid in a tank (the particular resin in the OP) translates into a change in the elevation of the center of mass. And, the weight density (as opposed to mass density) of a buoyant medium changes with elevation, especially in the case of a compressible medium such as air which also changes in mass density. Buoyancy is a quality directly related to the relative weights, and therefore only indirectly to mass. (I am aware of the potential to read in circular logic here, but in a "weightless" environment, there is no buoyancy, while mass remains unaffected) But I'm sure you know all this. We understand "g" to vary even though we routinely treat it as an algebraic constant and typically refer to it as such as a matter of convenience. The more correct buoyancy equation therefore is:

Weight = (Weight_Object – Weight_Displaced)

As to a ship floating on water "weighing nothing", well that of course depends on how we "weigh" it and how we define our system, weight being as you say an indirect quantity. Load cells or force transducers summing the total force on the bottom of a canal lock containing a fixed volume of water will very definitely register the increase in force (weight) caused by a ship lowered into it. We would therefore be correct in defining the weight of a ship as:

Weight_Ship = (Weight_Total – Weight_{Water Alone})

Knowing as we do that weight is a variable quantity, indirectly determined, and affected not just by relative positions of the particular gravitational masses, but other masses, and forces acting on the body (ie: centripetal), and on or "near" the earth, tidal forces as well.

However, again I clearly understood your point, and also understand that I am "preaching to the choir".... And I do like this choir.

I'm sure you could "split hairs" further also, but we both understand the central issues and that in addressing a mixed audience we take some "shortcuts" in how we express some of them, while explaining in great detail others, all the while not being sure exactly who we are addressing and what they might immediately grasp, and what not, as so many disciplines are represented here.

Regards,

Greg

Syhprum:

My apologies for misspelling your name in several replies over time.

Greg

Is the same tank showing the variance regularly or does it happen with all of the tanks? Also, what is the tolerance spec on the sensor? 80 pounds of fluctuation on an 80,000 pound capacity tank seems pretty small.

By the way,

I have observed that a ferite anizotropic permanent magnet put on analitic balance [type Sartorius,made in Germany,in years 1970] has two uncertain values.

Nic

The Sartorius electronic balances I used, circa 1984 used field coils (a specialized form of electromagnets) internally. By measuring the force required to support an object placed on the weighing platform, they calculated and displayed a weight value.

They were quite accurate and repeatable, within their specs, but if your balance used the same methodology, it would seem that your magnet interfered with their field magnets, although it could have been a purely mechanical interference caused by interaction with some magnetic part of the balance beneath the platform, especially if the platform was of a non-magnetic stainless steel.

Condesation in the tank?

Variation of thermal stress on tank supports?

From a practical rather than a theoretical standpoint.

1. The percent error is small.

2. Temperature effects can alter the load cell behaviour slightly, insulation can help keep the cell temperature due to excitation more constant.

3. Differential temperature change introduces stresses to support members, once again insulation reduces the error. Afternoon sun will bend a U beam when it heats one flange for instance and spray on urethane insulation slows the heat gain enough to allow the temperature to equalise across the section. This is especially evident when using bulk stress measurement as a load indicator but affects other designs depending on the installation.

If it isn't isolated look at insulating it.

One possibility is that the pipes connected to the tanks in question are expanding and contracting so as to place an additional stress on the load cell with variations in temperature. The additional stress manifests as a perceived change in weight.

Try changing some of the piping near the tanks for hose, and slackening off any rigid pipe supports in the vicinity of these tanks, whatever is appropriate.