Yes, there are some simple rules. I assume that you want to design an axial flux generator with no iron in the coils. In chapter 9 of my public report KD 341 I give the optimum coil configuration for the 3-phase, 1-layer winding of an 8-pole generator. One basic rule is that the average pitch in between the left and the right leg of a coil must be the same as the pitch in between the heart of a north pole and the heart of a south pole. I have used rectangular magnets and trapezoid coils but you can also use circular magnets and circular coils. The advantage of using circular coils is that the flux variation in the wires is not as suddely as for rectangular magnets and the generated voltage in one phase is therefore more sinusoidal than for rectangular magnets.
If the armature has 8 poles, the angle in between two adjacent magnets is 45°. If the coil legs are radial, the angle in between the legs of two adjacent coils is 15° (see upper picture figure 5). However, more copper can be used in a coil if the angle in between the legs of two adjacent coils is chosen 0° (see lower picture figure 5).
There must be a certain distance in between adjacent magnets at the pitch circle to prevent that rectangular magnets are touching each other at the inner side. I have chosen magnets with a width of 20 mm. The pitch circle has a diameter of 90 mm and so the pitch in between the heart of adjacent magnets is pi * 90 / 8 = 35.3 mm. So this is 15.3 mm larger than the width of the magnets. This choice results in a minimum distance in between adjacent magnets of about 4 mm which is okay. If the magnets are mounted closer to each other, there will be a relatively large magnetic flux flowing directly from one magnet to its neighbour and this flux doesn't flow through the air gap in between the magnets on the two different armature sheets. This will reduce the generated voltage.
If the same magnets would be used in a 16-pole generator, one needs about the double pitch diameter and one also needs twelve in stead of six coils. The angle in between two adjacent magnets now becomes 22.5° and so the minimum distance in between the magnets becomes somewhat larger. Doubling of the number of magnets and coils means that the frequency doubles for the same rotational speed. So the generated voltage doubles with a factor two because of the increased frequency and with a factor two because of the increassed number of coils. So the voltage increases by a factor four if coils with the same number of turns per coil are used. The power also increases by a factor four. So increase of the number of armature poles has a strong effect on the power. But lesser poles makes the generator simpler and in practice, eight poles is the minimum for a generator with a 3-phase, 1-layer winding.
If you use two armature sheets with magnets at the inner side, there are two magnets for one air gap. The larger the air gap for a certain total magnet thickness, the lower the magnetic flux in the air gap. The generated voltage is proportional to the flux density in the air gap. The larger the air gap, the more copper can be used for the coils. I think that an air gap equal to the total magnet thickness is an acceptable choice if neodymium magnets are used. This results in a flux density in T which is half the remanence Br. Calculation of the flux density in the air gap is easy if it is assumed that the steel sheets are that thick that the iron isn't saturated. An example of such calculation is given in chapter 3 of KD 679. This report describes an 8-pole generator with circular magnets and magnets on only one armature sheet.
About a year ago I have started the post "Design of a PM-generator" in which I have explained much more. But I can't find this post because this title gives problems of you look for it at "search".
