Simulations in analysis of LOW calibration requirements

This Feature consists of investigating 2 key topics:

• Constraining the rate of change and magnitude of random errors between receive paths within a station

From Stefan W: "I think we want to move away from the requirements that are based on the assumption of a 10-minute calibration cadence. As I argue in my paper, we probably want to limit the rate of change of random gain variations to a point that DD calibration at array level is able to deal with. Practically, this implies that we balance that rate of change to intrinsic effects. As we discussed last year, we should also put a limit on the magnitude to which these random errors may grow, even at a limited rate, to ensure that they do not grow so large that they cause an unacceptable level of decorrelation in the station beamformer."

Refer to section 2 "Stability"

https://confluence.skatelescope.org/display/TDT/Calibration+Requirements+Work

In particular,

https://confluence.skatelescope.org/display/TDT/Number+of+calibrators+vs+SNR

• Accuracy limits on (station) beam model prediction.

From Stefan W: "I think that the discussions on calibration transfer and beam model accuracy are related as flux calibration transfer is where accurate beam model prediction is needed. In LOFAR, we have demonstrated that, owing to DD calibration, one can get away with a reasonably poor beam prediction within the target field. I think this will relax some of those requirements considerably."

Refer to section 3 "Beam model accuracy"

https://confluence.skatelescope.org/pages/viewpage.action?pageId=106465087

The goal of this Feature is to investigate these 2 keys topics in order to update the related L1 and L2 requirements.