Add support for flow environments with body forces

[jcallaham]

Purpose

All of the current environments implement actuators as Dirichlet boundary conditions. This makes sense physically, but many numerical flow control configurations implement actuation as a body force instead (for the cylinder wake in particular). This has a few advantages, like not having feed-through effects from the actuation directly to the measurement and smoother estimates of the aerodynamic coefficients.

TODO

• [ ] Add notes to documentation about the challenges with BC-based controls (especially for co-located control and measurement)
• [ ] Modify body force implementation so it doesn't assume that the body force is just noise (i.e. eta could be actuation)
• [ ] Add alternate paths for including BC-based and body force-based controls in timestepping.
• [ ] Add BodyForceCylinder implementation following Noack et al 2011.