Now, all these things I am telling is very simple to hear, on the drawing, on the, on a piece of paper actually, when you do experiments the resistance at low speeds is very low, so there is bound to be, if the measuring equipment has a small error, a constant error, then that reflects here maximum, because the quantity we are measuring is small.
So, the error part of the model experiments shows mostly at low speed region whether its Rn, Fn, or speed, whatever we plot the graph width, if there is an error it shows mostly at low speed and therefore- I mentioned before also- the estimation of k itself and the nature of the curve here is always a little suspect and this suspicion reduces if you go for bigger models in a bigger tank.