Losses section and sankey diagram doesn't make sense or add up

I think there is a bug in the losses calculator that SAM is using for detailed residential distributed Photovoltaic analysis.

I have attached a pdf of the sankey chart that SAM generated, it is called "Losses.pdf". I've also attached a pdf of what my excel calculations show based on the loss data shown in the sankey chart called "sam question.pdf"

I can't for the life of me figure out how the loss percentages lead to the kWh stipulated in the diagram.

For example, Nominal DC energy is listed as 4394 kWh.

There are 4 losses listed under nominal DC energy:

snow 3.008%
module 1.387 %
diodes and connections .498%
DC wiring .498%

these losses add up to 5.391%, which would reduce the make the net DC energy equal to 94.609% of 4394 kWh, or 4157.12 kWh. The number that SAM spits out for net DC energy however is 3526 kWh, or 82% of the nominal DC energy figure. See my excel

Can anyone explain to me why this is so far off?

My method works for getting to gross AC energy from net DC energy, and for getting to annual energy from gross ac energy (see "sam question.pdf"). Why doesn't this work when going from nominal DC energy to net DC energy?

Dear Tom,

Thank you for this feedback.

You are correct that the loss diagram for the detailed photovoltaic model does not flow as you would expect from top to bottom. This is a problem with the loss diagram that we will fix for the next version of SAM. In general, you can consider the diagram to give you a rough sense of the relative magnitude of losses from different mechanisms affecting the sytem's output.

The problem is that those mechanisms are represented in different parts of the model and are difficult to express in a coherent way as a single set of losses. For example the "snow" loss is calculated by the snow loss model, which calculates a DC loss percentage that applies to the output of the module model. The "diodes and connections" loss is a percentage that you specify on the losses page. The "module" loss is determined by the value of the module model's DC output in kWh and represents several different loss mechanisms like module cover optical and cell temperature effects. I realize this might be a bit unclear as an explanation, but hopefully you can see that it is not straightforward to estimate the loss percentages for the diagram.

Best regards,
Paul.