I have a time-dependent function (F(t)) that returns a vector (in my case, a load vector in R^3) together with the corresponding position vector of the load. The cross product can be evaluated at specific time instances, but not symbolically over the entire load cycle.

That is generally fine — I do not actually need a symbolic expression of the complete load cycle. However, for fatigue and lifetime calculations of certain components, I need to compute an equivalent force/load value using an integral.

I expected the built-in integral function to evaluate the expression numerically point by point, but instead I get the same error as when trying to evaluate (Q(t)) symbolically.

For a trivial example like this, the integral could of course be calculated manually. But with much larger and more complex load-cycle definitions, this quickly becomes impractical.

My current workaround is to define a time vector using range() and then calculate the sum divided by the number of time steps. This works reasonably well, but it does not feel like the ideal solution.