Concrete strength estimates from the rebound hammer can be affected by a number of variables including water-cement ratio, type of aggregate, proximity of aggregate, proximity of reinforcement, and others. Brencich et al. (attached) discuss these issues. Liu et al. (attached) discuss the use of detailed statistical analysis to improve the accuracy, and develop a correlation like you aim to do. Alyamac et al. (attached) discuss the combined use of the rebound hammer and the Windsor probe to improve the accuracy of strength estimates. Finally, the attached CEMEx Technical Bulletin provides detailed procedural steps for the use of the rebound hammer.
Concrete strength estimates from the rebound hammer can be affected by a number of variables including water-cement ratio, type of aggregate, proximity of aggregate, proximity of reinforcement, and others. Brencich et al. (attached) discuss these issues. Liu et al. (attached) discuss the use of detailed statistical analysis to improve the accuracy, and develop a correlation like you aim to do. Alyamac et al. (attached) discuss the combined use of the rebound hammer and the Windsor probe to improve the accuracy of strength estimates. Finally, the attached CEMEx Technical Bulletin provides detailed procedural steps for the use of the rebound hammer.
It is very approximate and very subjective., because its an impact resistance on the surface (it may be coarse aggregate, paste area or may re-bar). One way is that try to get as many observations on all faces and deal statistically, say using distribution (based on histogram shape), you can develop a statistical relationship (just a suggestion)
I think Mr. Arturo E. Schultz has been very exhaustive on this issue. I commend him. I just wanted to mention that because of the so many variables involved in the use of the hammer, for the rebound hammer strength results to be meaningful calibration must be done for each site. The hammer can , however, show accurately any variations in concrete strength.
you could simply cast samples of the special concrete under study and then take at least 10 readings per each side of the sample in both direction; the horizontal and vertical directions, and then take the samples and perform traditional compressive strength test on them or equivalent samples, finally you will reach a new curve correlating the rebounds with the compressive strength of your special concrete, this off coarse will be affected with the sample size and curing condition.