MATLAB: Does a model containing disk friction clutch fail to solve when both ports of the locked clutch are provided with equal velocity inputs from an Ideal Angular velocity source

clutchconstraindiskfrictionlockSimscape Drivelinetorque

In my model, both ports of a "Disk Friction Clutch" are fed in with equal angular velocities using "Ideal Angular Velocity Source" blocks. When the clutch is locked, i.e. the control pressure signal value exceeds the Engagement threshold pressure, the model errors out with the following message:
An error occurred while running the simulation and the simulation was terminated
Caused by:
  • ERROR: Trouble solving algebraic equations in differential-algebraic system. Singular iteration matrix encountered with step size 0.2 at time 0.0. Try tightening the error tolerances. If the problem persists, check the model structure and values of parameters.
However when the difference in speeds is significant, the model runs fine. Why can't I specify equal velocities as inputs to the clutch?

Best Answer

The issue here is that the system is over-constrained. The velocity sources force velocity at a node. The clutch (when locked) forces the velocities on its ports to be equal.
When the input velocities are equal and input pressure exceeds the threshold pressure, the clutch attempts to lock. The Simscape engine cannot reduce the network, as the values of the speeds are externally provided (through Simulink) and fails.
When the difference in input speeds is large enough, the clutch provides friction, but does not lock. This does not constrain the speed between its Base (B) and Follower (F) ports. When the speeds are equal but the input pressure is lower than the threshold, this works because the pressure is small enough that the clutch does not lock. This also does not constrain the speed between its B and F ports.
Physical shafts are usually provided with a torque, rather than an input velocity. Using torque instead of velocity sources as input, in combination with appropriate inertia blocks should reduce the overall constrains on the system.
Related Question