Particle Size Distribution and Standard Deviation

It will be actually some fun trying out his data and analysing these. At least the shape of the curve would give us some idea .

I agree wholeheartedly

As you people banter over the maths involved here please keep in mind another variable:

I assume the particle size distribution you are measuring is that of a powder material. The physical method to grab the powder must also be carefully performed to ensure a representative sample is obtained. Larger particles obviously have more mass than small powders, however large powders will also have a lighter bulk density due to the increase in void space between the particles. Whenever the particles are in motion a fair amount of segregation / separation will take place.

In a pneumatic conveying line the large particles will stay to the outside of an elbow while the fine powders will hug the inside. If you dump powders into a hopper, the fine powders will stay under the fill stream while the large particles will bounce to the outside of the hopper.

Please elaborate on your sampling method.

This is also statistics (we are in statistics not maths )

The sample should be representative of the population.

We are not in the bulk powder sampling business, we are in the business of helping people solve problems, and for this problem the OP wants to know the method for calculating the Standard Deviation of a set of data already generated.

When the OP asks for advice on Sampling Methods I am sure there are more than able members of the forum to tackle that problem.

I was interested in finding out what method you are using and how you thought the method would have impacted the Statistical Analysis method being suggested.

However, I am thinking that if you are interested in being advised about a good method of Sampling, perhaps you would want to submit such as a post.

1. Yes the OP requested a method for calculating the Standard Deviation. That is why I regarded my original posting to be "off topic" and appropriately marked it as such.

2. I too consider myself to be in the business of helping people. However my experience is in powder handling, not statistics, and have investigated many applications where the plant statistician complained of receiving non-representative samples. The collection method will cause sampling errors so is a very relevant issue.

3. The OP did not state the process where the sample was being obtained (from a static pile/ hopper, moving mechanical/ pneumatic conveyor, etc. The behavior of powder particles in motion will be variable relative to the mechanical process involved. There is no "standard sampling method". View the sampling equipment links to understand the varied methods to apply samplers.

4. Should the OP forward any details on his particular application, and directly request my assistance, I will gladly review the application and forward specific areas for improvement.

5. I assume I ruffled your feathers by my "maths" comment. It obviously was a clumsy attempt at humor and hereby apologize if I offended any statisticians.

You did not ruffle my feathers by the comment of "maths", actually I was waiting all day yesterday to read about your sampling method; and when I did not hear from you, I offered my final suggestion. That you then came back after I have offered the final suggestion was what I thought inconsiderate or distracting.

Agreed, the Sampling Method would impact the results that obtains and therefore the method of analysis. However, the use of Wiebull Distribution works for all results that conforms to any of the standard probability distribution functions, so irrespective of the result, the Wiebull should work. You should try it.

By the way, have you ever performed repeatability and reproducibility analysis on the various sampling methods? I should think that the method that satisfies both Repeatability and Reproducibility should become recommended standard. What do you thing?

The specific timing of my previous posting was due to my being onsite a chemical plant, not any unwillingness to reply to your comment.

The "sampling methods" I am referring to is how to mechanically extract a sample of material from an isolated point on the application that will be representative of the entire process.

As stated before my background is in powder handling and not statistics. I will give you the benefit of the doubt that you are genuinely interested in my opinion, and not furtively attempting to discredit my experience. If it is the latter I will respectfully opt out of this thread.

Delmar,

I assure you that I have no interest in discrediting your experience under any circumstance. I was interested in any method that I thought you were using. I have actually always been interested in Entrained Bed Systems such as is the powder system you were describing. I have actually developed a full mathematical description and analysis of such system in the past, hence my interest.

Even more pertinent, I have, in the past, actually set up a Quality Assurance System for a company that supplied to Ford Motor Company, and used the concepts of Repeatability and Reproducibility. You said you were interested in knowledge so I thought I am relate to you the concepts just in case you might find it useful in making future decisions on Sampling Methods.

My regrets if I have come across as trying to discredit your knowledge/experience. I have recently decided not to respond to too many people either just because of that. Anyway, if you should feel that I can be of help at anytime, do please feel free to contact me at "editor@gbanalysts.com" , and I should help however, I can.