Since the variation of the concrete strength, would typically affect the load of specimens with different concrete strength, the measured load values for the specimens should be normalized.
Is that for loads only or for deflections as well?
Can we ignore the normalization if the difference between the strength is slight?
I'm not sure of which test are you referring to. For me, I'm expert in pushout tests for steel-concrete composite structures. These tests are basically measurement of load vs displacement. I always use the average of four displacement readings. This is to count for nonhomogeneous concrete and eccentricity of load.
Can you explain the tests you are doing? I think it could be usefull to normalize the load by the area to receive the stress and to normalize the deflection by the height of the specimen to receive the strain.
Im testing RC slab against punching shear. I found that some slabs have compressive strenght higher than the control which is a refrence specimen. To do comparisons betwee the slabs and the control, a normalization can be used. But is it for just load or deflection as well?
If the geometry of the slab, the slab span or the position of the load ist variable, you should normalize the deflection as well to compare the results.
Thanks for your response Rene. Actually all the specimens are identical. My point is because of the compressive strength has an influence on the stiffness if the slab, so we might to normailize the deflection as well
what do you think??
You can try to divide the deflection by the compressive strenth and compare the results. This ist not an usual normalization and the resulting unit is absolutely senseless. But maybe you can explain the influence of the compressive strength or rather the Young's modulus on the deflection.
It depends on what you are really interested and how you normalize these quantities.
If you just want to report your experiemental work, I suggest that you should present the experimental results in detail and with the original measured values. It may be very useful for other researchers, who may have different thinking about the problem you are considering and they may want to use only the experimental values instead of your interpretations of your data.
For example, many shear strength equations (also punching shear strength) do not take the deformation into account. It is of cause the big shortcomings of this approach and as a result, these equations may work only for certain concrete members. A better approach is based on deformation, especially crack width (also indirectly considered through crack spacing), e.g. punching shear model according to Muttoni, in which the curvature is of interest.
By normalization according to the concrete strength, you have to face the fact that the concrete compressive strength is only a mechanical property of concrete, and it cannot represent for all properties of a concrete. The other important properties are Ec, fct, GF (here, the maximum aggregate size ag also very important for estimating GF, if a direct fracture energy test is not carried out).
Even when you are sure that the other influencing parameters are the same, the normalization according to concrete strength fc is also not easy. For example, You want to normalize Vc according to fc. How you define the relative value of shear strength now: v = Vexp/fc or v = Vexp/fc^0.5 or v = Vexp/fc^(1/3). Of course, you can also add geometrical parameter such as the depth of member d in your formula. But which one you will use for the formulation of the relative shear strength ?
You may realise that the equation v = Vexp/fc is not good, because the compressive strength is not really relevent to the failure of shear. the function of fc with a form fc^0.5 or fc^(1/3),... may decribe the shear strength tau_c in cracked concrete or the tensile strength of concrete. It is strongly dependent on your interpreation and I bet that you may be not sure which direction you will continue to explain your experimental results with the hope that it is also true in a long time (until someone finds something wrong).
I strongly suggest the following approach:
1. provide all the information of the original measured values of your test.
2. In the comparison and disscusion, you can normalize the values to find the common phenomenon of your test results (independent from Point 1).
Best regards,
Ngoc Linh Tran
- Badges
- Science topic