You can parallel diodes but I would not count on being able to use two 20A to carry 30A. One diode will tend to initially carry slightly more current than the other, which will cause it to heat to a higher temperature. This will lower its forward drop and cause it to carry even more of the load. Thus one diode will tend to hog the current. Better to have each diode rated at least for the maximum current you want to carry.
This phenomenon is not as much of a problem as is commonly believed. There is some decrease in current sharing due to differences in diodes, but it is minimal.
Here is the result of an experiment. I took 2 identical 20 amp diode bridge assemblies, the type where individual diodes are mounted in a package, electrically isolated from the package, but on the same thermal substrate, which is intended to be mounted on a heatsink. I did not mount the assemblies on a heatsink.
I used a large 2 kW transformer, with a 120 VAC primary and a 12 VAC secondary. I connected one diode from each assembly in series with a .001 ohm shunt to monitor current, connected the 2 diodes (plus shunt) in parallel, with heavy wires, and placed the combination across the secondary of the transformer. With a variac on the primary, I could gradually increase the current through the diodes which were simply shorted across the secondary. I avoided any excess resistance in series with the diodes, which might act as ballast, tending to equalize currents. I captured scope images of the diode currents, with different colors for the two currents..
I first increased the diode currents to about 30 amps peak, starting with cold (at room temperature) diodes, and capturing the current traces before the diodes had time to warm up significantly. The first image shows this situation.
I then waited several minutes until the diode packages were blazing hot, about 120 degrees C according to a thermocouple measurement. The second image shows this result.
Finally, I allowed the diode assemblies to cool back down to ambient. Then I used a hot air gun to heat ONE of the diode assemblies to about 100 degrees C, with no current in the diodes. Then I quickly applied about 30 amps peak to each diode and captured the scope trace. This result is shown in the third image.
It can be seen that there is good current sharing between two diodes in two different packages, even when the diodes are allowed to heat up without heat sinking, at fairly high currents.
Even with the extremely unfavorable case of pre-heating one diode to 100 degrees, the other remaining at room temp, the degree of mismatch in currents is not much, 29 amps vs. 33 amps peak.
As Hero999 said, if the diodes are all in one package, they will tend to be matched, and the thermal substrate they are mounted on will not allow any one diode to get much hotter than any other. And, the normal resistance of connecting wiring will tend to ballast the diodes; I was careful to avoid this effect as much as possible by using short, heavy, connecting wires.
The problem you describe is a non-problem for this type of bridge rectifier assembly. A user can safely use two diodes from the same bridge in parallel, with the effective current rating doubled. Proper heatsinking is required, but considering that a bridge rated for 20 amps is rated that way with all 4 diodes carrying current, if only two diodes are used I would think that perhaps even somewhat more than 20 amps will be within ratings.