How would I fix the PV kWh annual output or efficiency from the kWp? I cannot see how to change the 13% efficiency for a generic cell for example or how to fix the kWh annual basis. For my project location, 1.5MWh for a 1kW cell is a good rule of thumb and I want to make this the generic option. I then want to create a sensitivity to shading to reduce this output by 20% due to tree cover. How can I do this?
The 13% efficiency parameter for the PV is only used if you've selected consider temperature effects, and it's not used how you might expect. HOMER only uses the efficiency parameter to calculate the PV cell temperature (see attached article "How HOMER Calculates the PV Cell Temperature").
HOMER doesn't use the cell efficiency to calculate PV output, since you already specify the PV size in terms of nominal capacity. Consider a HOMER PV component with 1 kW of rated capacity and a derating factor of 80%. In STC (1 kW/m2 incident radiation), the panel will produce 800 W (1 kW * 80%). The efficiency figure does not affect this calculation. I've attached an article from the HOMER help (How HOMER Calculates the PV Array Power Output) that describes this in more detail.
I can imagine a couple of ways to set the annual production to a specific amount. I think the most straightforward method would be to adjust the derating factor so that the simulation matches your empirical data. You can run a simulation with your solar resource and 1 kW of PV and any load. You can set the Maximum annual capacity shortage to 100% to get a "feasible" result. In the results, look at the total annual electric production (click the "Column choices" button and check the box for "Annual electric production"). You can then adjust the PV derating factor to get the result you expect. I found that changing the derating factor to 63% yields 1.5 MWh annually per 1 kW of PV capacity for your location.
You can now use this derating factor in your full models. You can reduce the output 20% again with the derating factor. You can change the value, or run a sensitivity analysis with several derating values. To derate by another 20%, multiply that new derating factor by 0.8 (0.63 * 0.8 = 0.5) to find the derating factor for the shading case. With 50% derating factor I now get 1.2 MWh of annual production for 1 kW of PV. Of course, adjusting the derating factor doesn't give a perfect representation of shading, since shading can change the daily output profile, and will vary over the course of the year. This approximation would be good enough in many cases, but if the hour-by-hour PV power output profile is really important, you might want to use a tool like PVsyst to calculate the hourly effects of shading and changes over the course of the year. You can export GHI or PV power output (you can use solar DNI in order to specify an hourly time-series power output profile) from PV syst to use in HOMER.