Anyone has any hint for me? I looked at the standard IEEE C57.12.91, but could not find a specific example for a Zig-Zag transformer.

Since the "load" losses should relate directly to the winding resistance, and the this load should consist entirely of zero sequence current, maybe you can approximate using this value. Once you have the resistance and the impedance you should be able to compute the X, and then the X/R.

You have got the Z0 in Ohms. Unless you get the X/R ratio from the manufacturer it is my understanding that you cannot calculate X0/R0.

@Tony2802,
I have verified your data and found these:
1) The information on the short-time neutral current of 3500A for 5 seconds and the continuous neutral current rating does not correspond to the IEEE 32 numbers that I found. The continuous current if we use IEEE 32 will yield 105A. The continuous current for a 10-second time rating is 3% of the ground fault (3500 X 0.03 = 105A). Instead, your data fits with a grounding duty for 1 minute (7% of ground fault current, 3500 X 0.07 =245 ~ 250A).
2) You can use the no-load and full-load losses to calculate the resistance. no-load losses = core losses and full-load losses = core losses + copper losses. Therefore your copper losses = 18,241 - 1951 = 16,263 W. Equate those losses to 3I²R and you get R which is equal to 0.813 ohms.
3) Using the Z₀ = 7.5 ohms, you'll get x₀ = square root (7.5²-0.813²) = 7.46 ohms
I got x₀/r₀ = 9.2