Repmag Part IV

Showing a single 65 × 35 × 25 NdFeB35 magnet entering a steel shield

Examination of the results obtained from magnetostatic and then dynamic analysis of the original Repmag idea discussed in Part I suggests that, although it might be expected and "reasonable" that the energy gained by each magnet entering the shield would be balanced by the energy lost by another magnet exiting the shield, this does not actually occur. Further modelling of this in the above linear model shows that quite a lot more energy is required to extract a closed-up pair of magnets from within a shield, than is gained in total when the two magnets enter the shield separately.

This is so, even though total flux from the closed-up pair of magnets is slightly less than the sum of flux produced by two separate magnets.

This energy difference is almost (but not quite) enough to negate the energy gained from subsequently allowing the magnets to separate outside of the shield.

This result is an example of the surprises that can be encountered in detailed magnetostatic modelling of even quite simple concepts.

Another such surprise (a pleasant one this time) was the extent of the energy reduction for two magnets repelling side-by-side in air, then closing up again within a steel shield. The result for two 65 × 35 × 25 NdFeB35 magnets is 5.706 J gained (in air) vs only 0.0427 J lost, for 0.5mm airgaps between the magnets and the shield. I had not expected so much reduction, given that there is about a threefold increase in B (flux density) when the magnets are inside the shield, and that the force exerted by a magnet is roughly proportional to B². Evidently the steel shield concept is very effective in preventing flux from returning around the sides of the magnets.

The above linear model is "ideal," with zero separation between a pair of magnets throughout the entire duration of their exit — in fact some separation does occur as the magnets exit in the Repmag model. Even a small amount of separation can make a significant difference to the overall performance.

However, I would have to find a definite solution for the entry/exit energy difference problem before going further with this concept. Otherwise, the modelling done so far suggests that the concept would work only from "second-order" effects, which would be too weak to justify building a physical prototype.

For splined-spider Repmag modelling

This model was built for future magnetostatic modelling of the splined-spider approach, with radial holes in the magnets, as shown in Drawing 2 last time.