There are similarities with a wind turbine. The wind velocity is monitored and the torque and turbine speed is obtained from a previously defined look up table. I would like to use an adaptive algorithm for the hydro turbine. Perhaps every 5 minutes alter the torque and hence the turbine speed very slightly to see if the generated power increases or decreases. If the speed is optimum at the top of the efficiency curve the slight increase and decrease in speed should both cause a slight drop in output power. Otherwise the speed needs to be increased or decreased to achieve the optimum and the test performed again. In essence a finite differentiation is performed to determine the maximum. Has anyone experience of using such an algorithm with a variable speed hydro turbine ?