I can't find the possible hydro parameter to match my load.
 

When modeling small hydro power in HOMER, you need to be careful to match the size of the hydro turbine (the design flow rate) to the available hydro resource.  It is usually best to start with the hydro resource.  First, estimate the stream flow to whatever degree of detail you can, and then estimate the required residual flow, which is the quantity of water that you cannot use in the hydro power system because of ecological or other concerns.  The quantity of water available to the hydro turbine is the hydro resource minus the residual flow.

Once you have entered the hydro resource inputs, you can specify the hydro turbine inputs.  The terrain determines the available head, but the system designer has control over the design flow rate.  This is a critical value because most hydro turbines can only operate within a certain range of flow rates.  If you choose too large a turbine, meaning one whose design flow rate is much higher than the available flow rate, the available flow may never be sufficient to operate that turbine.  That is what is happening in your file.

Note the inputs immediately below the design flow rate:  the minimum and maximum flow ratios.  If the design flow rate is 100 L/s and the minimum flow ratio is 50%, as they are in your base case, then the turbine can only operate when the available stream flow exceeds 50 L/s.  In most of your sensitivity cases the available stream flow (the gross stream flow minus the residual flow) that never happens so the turbine never produces power.

In the attached modified model, I have set the gross stream flow to a constant 20 L/s, and the design flow rate of the hydro turbine to 20 L/s as well.  Combined with an available head of 25m, that allows the turbine to produce a steady power output of a little under 4 kW.  Combined with a battery/inverter, that is enough to serve the load with only a small amount of capacity shortage.

I also cleared out all your sensitivity variables for simplicity.  I think you were using sensitivities to search for feasibility, and they can often help with that, but you had an impractically large number of sensitivity cases.  Often it is better to start simple and achieve feasibility before moving into sensitivity analysis.