Thank you for the detailed explanation and additional images, that is very helpful.
A simpler approach to replacing the chain is to add two springs, one on each side of the sprocket tangent point, the other end to each hook. A spring can be scoped to a cylindrical face on a sprocket tooth, and the other end to a face of the hook. You can move around the coordinates after scoping to get the chain to be in a straight line parallel to the Y axis if you want. The chain has an axial spring rate of some kN/mm per meter of length. You might need to request that data from the chain manufacturer, or measure it in the lab. The downside of using a spring instead of link elements to represent the chain is the spring has no mass, which affects a Modal analysis slightly.
Now the real mechanism has a low rate spring (compared with the chain), on the other side of the hook. Replace the cylindrical structure that represents the real spring with a Spring connection, and insert the proper spring rate. This is a spring that might have a pretension applied that pulls on the muscle through the chain and sprocket. You can apply a pretension to this spring.
Modal analysis doesn't include loads, so applying forces is not allowed in a Modal analysis. You would use a Static Structural model to apply loads if you are interested in the deformation of deformable parts. Or you might use a Rigid Dynamics model to apply loads if you are interested in the motion of rigid parts and the forces acting through them. Let's say you are in one of those.
Instead of applying a force to the weight, insert a Gravity load into your model. But make sure that the properties of the weight show that it has the correct mass. If not, you can add a point mass at the center of the face to make up the difference if it is too light, or you can create a custom material and adjust the density until the mass is correct. The second technique must be used if the initial mass is too heavy.
If you want to plot muscle tension vs. shaft angle, that is most easily done in a Rigid Dynamics model. Let's make a very simple model that you can get some similar results from easily. Make a copy of the file you are working in so you can go back to the original file later. In Workbench, RMB on cell 1 Modal and Replace with... Rigid Dynamics. Open the geometry in SpaceClaim and create a new component and call it Arm. Move all the solids such as the weight, the rod, the disk, the shaft and the sprocket into that component called Arm. That is one rigid body. You can pick all the other solids in the Structure and set to Ignore for Physics so that they don't transfer.
Now open the Model cell. In Mechanical, insert a Revolute joint to ground on the shaft body. Insert a Standard Earth Gravity load and check that it is pointing in the correct direction. Add a Joint load to the Revolute joint: type: Displacement and Magnitude, Function, =10*time. Under Analysis settings, if you want to move 180 degrees, enter 18 seconds for the End time. Solve the model. Into the Solution, put a Total Deformation result and you can animate to see the Arm rotate. Add a Probe, Joint, Total Moment to the Solution branch. Now you can plot Torque vs time. Add a Probe, Joint, Relative Rotation. Now you can plot Torque vs Angle. Of course you could have obtained the same plot using Excel and a few simple formulas.
I understand this is not exactly what you want in the end, but you have to get some experience with successfully building, solving and obtaining results from a model first. Let me know if you have success building this simple Rigid Dynamics model. You can get the same plot from Static Structural, it would just take a lot more time to solve the model.