April 2017 Jeff Miller Composites presentation

[jeffrey.miller]

Starting thread ahead of the presentation for a change.  

Rob Schwartz and I will be building some test panels for destructive testing at the upcoming meeting.  I will be posting my math here as I figure out how much they should take before failure.  

I have done a previous presentation on the topic of composite battery boxes, when we were prepping for the Porsche modification.  That presentation was light on math, and heavy on costs and concepts.  
http://www.fveaa.org/forums/index.php/topic,1286.0.html

This time around it will be more math and destruction.  Can't have destruction without the math.

As usual around here we will be using the Kevlar left over from the Porsche and Dakota.  KEVLAR® 5oz from US Composites
http://www.uscomposites.com/kevlar.html
Their spec sheet is seriously light so we have to go back to the manufacturer of the yarn it is made out of.  
http://www.matweb.com/search/datasheet.aspx?matguid=77b5205f0dcc43bb8cbe6fee7d36cbb5&ckck=1


Physical Properties   Metric   English   Comments
Density                   1.44 g/cc   0.0520 lb/in³   
Water Absorption    3.5 %   3.5 %   As shipped; Typical moisture levels on yarn as shipped; they reflect values reached at normal, moderate temperature and humidity levels following fiber production, which is a wet process.
Moisture Absorption at Equilibrium    3.5 %   3.5 %   Equilibrium from Bone-Dry Yarn; Equilibrium values are determined by bone drying the fiber and condition at 75°F (24°C), 55% RH.
 
Mechanical Properties           Metric   English   Comments
Tensile Strength, Ultimate    3000 MPa   435000 psi   Breaking Tenacity
                                    3620 MPa   525000 psi   Epoxy-impregnated strands, ASTM D2343
Elongation at Break            2.4 %   2.4 %   
Tensile Modulus                    112 GPa   16300 ksi   
Tenacity                            2.08 N/tex   23.6 g/denier   
Poissons Ratio                      0.36   0.36   
 
Thermal Properties   Metric   English   Comments
Specific Heat Capacity    1.42 J/g-°C
@Temperature 25.0 °C   0.339 BTU/lb-°F
@Temperature 77.0 °F   
Thermal Conductivity    0.0400 W/m-K   0.278 BTU-in/hr-ft²-°F   
Maximum Service Temperature, Air    149 - 177 °C   300 - 351 °F   For long-term use
Shrinkage    <= 0.10 %   <= 0.10 %   In water at 212°F (100°C) and in air at 351°F (177°C)

Red highlighted is one of the key items here.  

Since this is the yarn, we will need to adjust for the fact that half the yarn is running in the wrong direction for our test.  So our tensile strength will be 262,500psi in a single direction for the fabric in question in our planned test.  Steel is in the 80,000 range.  Fabric total thickness is 0.0095 inches, so if we end up building something that is one foot wide, the tensile strength of one side will be 0.0095*12= 0.114 square inches multiplied by the ultimate tensile strength of the fibers that are running in the direction of interest gives us a ultimate tensile strength of 29,925 pounds per side.  Various people report that Kevlar is equally good in compression and tension so for our purposes I will assume it is equal. 

[jeffrey.miller]

Rob Schwartz and I built the two test pieces for the upcoming meeting last night.  At the very least Rob learned how much of a pain Kevlar is to cut, even with the right scissors. 



The foam we used is on the low end of strength compared to the ones that we have used in our builds.  It was donated by Bruce as the leftovers from his efforts building simulated lithium batteries for the design phase of his HupMobile, (he wasn't concerned about density for his purposes).  The fabric and epoxy were leftovers from both the Porsche and Dakota project.  Thanks everyone for the materials, and Rob for putting in the effort. 

It will be a destructive presentation! 
IF YOU ARE PLANNING TO SIT IN THE FRONT ROW PLEASE BRING SAFETY GLASSES!
I don't expect things to fly in that direction, the force will be down towards the floor, but unplanned stuff happens when you break things.   

[jeffrey.miller]

The panels are fully cured.
Weight of the thin one is 1 pound 1.35 ounces
Thick one is 2 pounds 9.06 ounces

They both contain roughly the same amount of Kevlar and epoxy, so the main difference in weight is the additional foam core material. 

[ted.lowe]

  ... but unplanned stuff happens when you break things.   

 

[bruce.jones]

Maybe we should do this outside?  Do you need any weights?

[jeffrey.miller]

I plan to use human volunteers and four bags of lead I have laying around.  If it isn't enough we will revisit the still standing panel another time.  I ran some initial estimates today, and the numbers are pretty big, but the real world isn't that perfect.

I think I am also going to start with some smaller scale tests built out of lower strength materials, to help show the different failure modes before we go for it with the high strength fabric parts.  

It would be good to have a bathroom scale, to weigh people on, can I get a volunteer to bring one?  
Bruce, if you have some weights in the 5 to 10 pound range, that could be useful for the small scale tests. 

If we are concerned about flying debris, I can bring a bag to put the test pieces in for the actual test, I was planning to lay one below the test pieces to ease any cleanup.  

[jeffrey.miller]

Core material is generally polystyrene
Specificially
http://www.foamular.com/foam/products/foamular-150.aspx
ASTM C578 Type X, 15 psi, 1.3 pcf density
They claim 15 psi, and I believe them. 

The panels are roughly 30x16 = 480 square inches x 15 pounds per square inch =  7200 pounds. 
Of course we aren't loading it up evenly like that.  It gets tough to calculate, but I estimate the core will fail, at the highest when 1/3rd of the area hits it's max loading, which is 2400 pounds.  This is because the stresses will be concentrated by the loading, and the overall design.  I don't have any actual math on core failure beyond guesswork at this point. 

The math I understand for the Kevlar is showing that the kevlar will fail higher than 2400 though.  On the three inch thick panel, it is much much higher.  I need to redo my math and post that up at some point, but not tonight. 

Also, we have traditionally used the heavier 5 ounce fabric, they make a lighter fabric, but the cost per yard isn't proportional to it's strength so I just go for the heavy stuff and that way I am pretty confident that the fabric won't be the failure mode without doing a bunch of math. 

[jeffrey.miller]

It has been almost a week since the presentation, and I am just getting around to posting my analysis!  

I will open my analysis with the fact that the foam in the Porsche is twice the density of what we were testing with.  I don't have the spec on the foam in Ted's build as he had done most of the cutting before I got there.

The early tests with the low strength coverings failed the foam with the exception of the single layer of packing tape which stretched the tape enough to allow the foam to tear.  In retrospect that should have caused me pause.  In the case of the paper covering the top of the foam was compressing down and allowing it to fail.  It was hard to see that even where I was standing.  

What happened and why?  In my presentation I said that the Kevlar of the thin panel would fail at 1330 pounds and the thick panel material would fail at 7980 pounds.  I speculated on the strength of the foam that it would fail at 1/3 of the area under 15 PSI, but that speculation was very wrong.  The thin panel did fail at 118 pounds and the thick panel failed at 500 to 700 pounds depending on how the load was applied.  When we consider this it actually makes sense.  The stress applied to the fabric of the thin panel at 100 pounds is 1500 pounds force to the fabric.  Considering the top sheet is taking 750 pounds of compression at that point, we can start to figure out what happened.  The failure was when the compressing Kevlar buckled down in to the foam.  That affected an area less than an inch wide space across the 16" width of the panel.
I don't have the panels in front of me but it looked like it crushed about 3/4 inch wide strip of the foam in and allowed the Kevlar to buckle inward.  16" width panel x 3/4" crush area gives us a 180 pounds of crush force at 15 psi.  

The thicker panel lowers the forces on the fabric which inherently lowers the forces on the surface of the foam (which is where they both failed).  In my presentation I stated that the fabric could take 7980 pounds of force which is 6 times higher than the thinner panel, which is driven by it having more distance between the two sheets of Kevlar.  By lowering the force on the Kevlar we lowered the compression force on the foam that caused it to collapse.  It failed identically to the thin panel and appeared to have the same 180 pound force to crush the foam, but it took roughly six times the weight to do so.  This is due to the foam in the panel being 6 times as thick which lowered the stresses on the fabric by a factor of 6.  

In retrospect it makes perfect sense and don't skimp on the foam.    



If we had added more fabric, it would have taken more weight, but only because the fabric would have to buckle over a larger area due to it's greater thickness it would crush a larger area of foam due to it's internal rigidity.  The payoff would not be worth the expense though, using a stronger core material would yield better results at lower cost and lower weight.  

How strong of a core do we need if we were to repeat this test to get the fabric to fail?  That sounds like something to answer on the next post.  

[jeffrey.miller]

http://www.uscomposites.com/foam.html

Urethane foam offering up to 580 PSI in compression, makes the foam we used look like a piece of sandwich bread. 

Of course that comes at the cost of weight, the 580psi one has a density of 16 pounds per cubic foot!  If we made the thick panel completely out of the 16 pound the foam would have weighed 12.5 pounds.  Although in theory we could use multiple densities and build it up, assuming the part is really thick.  Their quote is excellent "Common Applications: This 16LB density foam is essentially as hard as a rock, you would need a hammer in order to make any dents in this product. "

They offer other strengths all the way down to 40PSI, which is still much higher than the 15 PSI we used the other night. 

[jeffrey.miller]

To calculate the proper foam to make the fabric the weak link we can take our experimental data and extract some lessons learned and then extrapolate. 

The foam failed when its crush strength was equal to roughly 10% of the compression load on the fabric.  (this data is only valid for this fabric in a single layer with Epoxy)

Considering that fact for a moment I previously calculated that the fabric had a strength of 39,900 in those panels.  That means the foam would have to withstand 3990 pounds of crush force over a 12 sq in area.  This is assuming the combination of the two things interact identically, but if we run with that.  That means that foam with 332.5 PSI in compression would be equal in strength to the fabric in this build.  I am extrapolating roughly a 20x value, so the margin of error is large to say the least.  If I was so inclined I could build a panel out of the 250 and 580 PSI Foam and see what happens, when I break them.  I would make them much smaller, so we don't need 8000 pounds to break them though. 

The early test strips were generally 2 inches wide.  If that was made out of one of these two higher density foams and the same single layer of fabric it would hold 1/8th the weight, so 2 inch wide 3 inch tall would hold just shy of 1000 pounds.  The foam would be dimensions 3*2*32, the heaviest foam would be just 0.9 pound, and the next foam down would be 0.45 of a pound.  The fabric and epoxy would add about 2 ounces to that.  One or both of those builds should cause the fabric to fail.

[jeffrey.miller]

At the meeting we briefly touched on lighter weight Kevlar material and I mentioned I don't bother because the cost is about the same as the heavier stronger cloth, I was wrong, it is actually more expensive!

http://www.uscomposites.com/kevlar.html

The 1.7 ounce is showing 28.50 per 38" wide yard.
The 5 ounce is showing  17.50 per 38" wide yard. 
Usually we end up buying the 60' wide roll which is presently 25.50 per yard at close to double the width.  More strength, less math, and considerably less cost.