The error is likely resulted from the algebraic loop inside the PV Array as described in this page:
By default the "Break algebraic loop in internal model" parameter is not selected. This means that the PV Array has an algebraic loop in it. If you use the block in a discrete system using large sample times, this algebraic loop is required to get an iterative, accurate solution for the highly nonlinear diode characteristics. For continuous system, when the PV array block is connected to a detailed power electronic converter where real switches are simulated, you need to specify a small sample time to get accurate resolution in PWM pulse generation (for example, 1e-6 sec with a 5 kHz PWM inverter).
If you do not want to break the algebraic loop and you know that your model is correct, please consider the following:
1) Reducing the error tolerances (Relative tolerance and Absolute tolerance in Solver tab of the Configuration Parameters) of the solver as suggested by the error message. This might slow down the computation.
2) Changing the solver to ode23tb which typically works better when you have Simscape components in your model.
Additionally, please be aware of the measurement filter time constant in the advanced tab for PV Array blocks. If the time constant is zero, the solver will have a harder time solving the model with an algebraic loop.
Best Answer