I am trying to simulate a low-temperature differential (LTD) Stirling engine.

The scenario goes like this: Some gas is confined inside the engine cylinder and is heated from below due to the bottom plate resting on a hot surface. The resulting temperature increase causes a rise in pressure, and if everything is configured correctly, the pressure forces should be able to drive piston motion, allowing the pressure to evolve consistently with the changing volume. I'm having trouble making the 2 pistons move in their own sinusoidal trajectory with same stroke length while having a 90 degree phase offset between the two. It seems that i could make both pistons moveable by enabling the In-Cylinder option in dynamic mesh, but i cannot preserve the 90 degree phase shift this way. The 6DOF option also does not seem to have any capability on its own to set the piston movement conditions.

The yellow faces mark a symmetry boundary condition.

Pistons moving as they naturally should with an udf assigned to certain walls (top and bottom displacer piston walls and power piston wall) with layering and smoothing enabled. Maybe i could make use of a similar udf, that only sets the conditions and movement trajectory for the pistons, making the pistons free to move under changing gas pressure?