Fix ZeroDivisionError in simple_efficiency when load_loss=0

[aman-coder03]

fixes a ZeroDivisionError in pvlib.transformer.simple_efficiency
when load_loss = 0.

when load_loss is zero, the transformer loss model reduces to a linear
relationship rather than a quadratic one. this PR handles that case
explicitly to avoid division by zero while preserving existing behavior
for load_loss > 0.

Fixes #2645

[cwhanse]

Suggested change

	if load_loss == 0:
	return input_power - no_load_loss * transformer_rating

I don't think this is necessary with the alternate form of the quadratic equation.

[aman-coder03]

i have removed it

[cwhanse]

Could you add a test with load_loss=0?

[aman-coder03]

sure!!

[aman-coder03]

thanks!! @cwhanse
i have added a test covering the load_loss=0 that checks the linear-limit behavior P_out = P_in − L_no_load * P_nom
please let me know if you would prefer this test to also cover array like inputs or assert behavior at additional edge cases (negative input power)

[cwhanse]

Let's keep the scope to addressing the division by zero issue. Test looks good.

Please open a new issue to extend testing to np.array and pd.Series types.

[aman-coder03]

thanks @cwhanse
sure i will open issues separately for each np.array and pd.Series

[echedey-ls]

thanks @cwhanse sure i will open issues separately for each np.array and pd.Series

@aman-coder03

Just one issue for both of them types is enough.