well i haven't done the experiment myself but i suspect you could get about twice as much power out of an existing set of magnets, by making a two layer winding as what is shown in your diagram, as compared with the single layer 3/4 arrangement that most people make.
But halbach arrays are not 100% efficient. meaning, when you double the volume of magnet, you won't get a 41% increase in air gap flux density. how much you will get depends on how many segments of magnets you fit in there and how they are placed and other details.
So for instance, if you have a 1T air gap flux density, a halback array might get you 1.3T rather than 1.41. while the difference doesn't seem like much, generators and motors operate according to the square of the flux density. If instead of making a halbach array, you increased the diameter of the rotor by 1.41 instead and used twice as many magnets, you would have a machine that would generate more than twice as much power at the same rpm.
Halbach arrays are good for reducing inertia and weight, making very uniform magnetic fields. where weight and inertia and volume are not a problem they are not cost effective for electric motors.
however it is probably cost effective to make a halbach array from ceramic magnets, because ceramic magnets are cheaper than magnet wire. so doubling the volume of ceramic magnets in a motor might be cheaper than making the motor larger. where as for neodymium at 7$ per cubic inch.. its cheaper to make the motor larger than to use more magnets.
also take a look at the Novatorque motor, except instead of using a halbach array, they are using geometry to fit a larger magnet for a smaller cross section of core.
