Add test for start input values

[mabijkerk]

In principle, the start value of all inputs should work for all datasets. In the past, we have encountered issues where this was not the case. For example, an efficiency > 100% even though the maximum value of the input was set at 100%. It would be good to add a check for this.

This seems a mechanical turk test, since it is not only related to the datasets, but also to the simulation of the input in the future year. We would preferably write a test that for all datasets extracts the start value of all inputs and then sets the values in a scenario. The test should then not produces any calculation errors.

What do you think of this idea @noracato. Notifying @kaskranenburgQ.