The lack of Proto:87 driving wheels has proven a major stumbling block for steam era modellers in Proto:87. By contrast, there is a wealth of driving wheels available in Proto:4. The question is, are they close enough to Proto:87 to be useful to us? The above diagram illustrates that in a first approximation, they are. The "depth" and "width" of the flange are very close, and a little judicious shaving could bring them into line.
The Theory - Roger Miener
Both the Proto:4 and the Proto:87 wheel are based on the same standard profile (BS276A).  In order to determine the feasibility of producing P87 wheels from P4, I produced a number of drawings comparing the two profiles, and the most interesting is shown in Figure 1 above. The results of these CAD-based experiments have perhaps turned out to be a bit more than just academic.  This comparison provides information that should be of use to anyone who contemplates using a commercially available P4 wheel as the starting point for creating a wheel that will be functional on trackwork built to P87 standards.
The drawing demonstrates the difference between the Effective Flanges of the P4 and P87 wheels when both are sitting on the same rail.  The effective flange is the distance from the back of the wheel to the gauge point of the rail when the wheel

is as close to the rail as possible without lifting the tread from the rail head [Brook-Smith, 1997].  So, if there is a villain in this piece, it is more likely to be the rail rather than the wheels.
The rail used for the comparison was Micro Engineering Code 70 flat bottom rail. This is a rail that will likely show up in wide use in P87 trackwork. Micro Engineering's specification for this rail is head width of 0.30 in (0.762 mm,) major top radius of 0.12 in (3.048 mm,) and a corner radius (rail shoulder radius) which defines an arc tangent to the rail's top radius and its vertical side of 0.004 in (0.102 mm.) For the drawing, the rail was oriented vertically - not canted. The wheels were also held to a vertical orientation. The treads of both wheels slope at 1:20.  The flange root radius for the P4 wheel is 0.21 mm and for the P87 wheel it is 0.18 mm.
When sitting on the specified rail, the Effective Flange (EF) for the P4 wheel was 0.438 mm and the corresponding dimension for the P87 wheel was 0.373 mm. Compare these numbers to the P4 specified min / max range for EF of 0.35 mm / 0.40 mm; and to the P87 specified min / max range for EF of 0.31 / 0.35 mm.  In each instance, the measured EF derived from this experiment falls outside the respective max limits associated with each wheel. This excess results from the large difference between the rail's shoulder radius and the each of the