Adriaan,
I would like to hear any more thoughts you may have about the dropping CP at wind speeds higher than 25 kph.
The first explanation that the safety (furling) system is causing it is plausible, but that should be evident in the TSR measurements.
If the wind turbine has no safety system, it means that the rotor will be perpendicular to the wind. In this case, the maximum Cp which can be obtained if the rotor is loaded such that it runs at the optimum tip speed ratio, will slightly increase at increasing wind speed. This is caused by the increasing Reynolds numbers as increasing Reynolds numbers result in lower drag/lift ratios. However, at a certain rotational speed, the maximum torque level of the generator will be reached and this means that the generator is no longer able to load the rotor such that it turns at the optimum tip speed ratio. Then the rotor will turn at a higher tip speed ratio and the tip speed ratio will be higher as the wind speed is higher. This results in decrease of the real Cp and the reduction of the Cp is more as there is a larger distance in between the real tip speed ratio and the optimum tip speed ratio. You need to know the Cp-lamba curve of the rotor to find this reduction.
If the wind turbine has a safety system which turns the rotor out of the wind at high wind speeds, the maximum Cp is reduced because of the yaw angle delta. This effect becomes important for yaw angle larger than about 20 degrees. Realise that the power goes down with cos^3 delta but that the rotational speed goes only down with cos delta. If the safety system limits the rotational speed below the value for which the generator has reached its maximum torque level, the reduction of Cp because of the yaw angle is the main effect. However, if the rated wind speed is rather high, both effects together will result in decrease of the Cp at high wind speeds.
I have looked at the measured efficiency curves for 24 V and they look strange to me if I compare them with curves I have measured in KD 78. What I have measured is that a curve for a certain voltage and a certain way of rectification starts at the rotational speed for which the open DC voltage of the generator is equal to the open battery voltage. From this point, the efficiency curve rises very strongly and it reaches a maximum at a rotational speed which is about a factor 5/4 higher. Then it decreases, mainly because of copper losses in the generator winding but this part of the curve is no straight line but a curved line with the hollow side above. So I don't understand how you can have measured straight efficiency lines.
If you have accurate generator characteristics for the correct load and way of rectification, you can find the generated Pmech-n curve of the rotor by dividing the measured Pel-n curve of the generator by the correct generator efficiency. But if you have used a wrong efficiency curve of the generator, you will also make a mistake in the calculated Cp.
But the biggest mistake is made if the undisturbed wind speed has been measured incorrectly as the power is proportional to V^3. Correct measuring of the wind speed is very difficult in the field. You should do it at about three rotor diameters in front of the rotor at the height of the rotor axis but then you can take only the measurements for which the wind direction is about in line with the line through the wind meter and the rotor axis. Another problem is that it takes a certain time until the measured wind has reached the rotor plane. So you can take only measuring points for which the wind speed has been constant for a rather long time. Even if you have taken only measuring points for which all measuring conditions are right, you still will find a rather strong variation the the Cp value for a certain tip speed ratio.