Closed RickBrice closed 4 months ago
The hand calculations in issue #146 suggest that the integral equations may be correct, but there are likely other inaccuracies in the documented equations for the Ifc___PolynomialSpiralCurve types.
After learning of the correct repository for ifc rail unit tests, I am able to exactly match horizontal and vertical for the polynomial spirals. Have not yet compared unit test results with cant. I’ll close this issue with no changes required.
@peterrdf @apinzenoehler @evandroAlfieri
Problem The $\lambda(u)$ equations for IfcSecond, IfcThird, and IfcSeventhOrderPolynomialSpiral are incorrect. The equations are of the form:
$$\lambda(u) = C + \int{0}^{u}\cos \theta(t) dt x \int{0}^{u}\sin \theta(t) dt y$$
The integrals yield scalar values while C is a location with length units.
Solution(s) I think the equation should be $$\lambda(u) = C + L ( \int{0}^{u}\cos \theta(t) dt x \int{0}^{u}\sin \theta(t) dt y )$$ where $L$ is the arc length when $u=1.0$.
The documentation might also benefit from stating $u=1.0$ when $\theta=\frac{\pi}{2}$
Require schema changes?
✓
noRequire documentation changes?
✓
yes