How do you prove and explain the accuraccy of your models ?
If it is an existing system, you can compare against the test results on the real system. This would be the foundation to build up further. If it is a completely new system, you will need to check whether the power consumption is within your expected range. For example, if the range was between 0.5 and 1 W of power but the simulation results shows 5 W, then you must use the model to understand why the number is so high. Again, this is a part of system design.