Fix the Vin prediction using expected k-factor not measured one

The predicted slope for the VI curve depends on the chips and is calculated based on the measured mirror factor (kShunt). https://gitlab.cern.ch/atlas-itk/pixel/module/module-qc-analysis-tools/-/blob/main/src/module_qc_analysis_tools/cli/SLDO.py#L216

This should be changed to always use the expected average kShunt.