InvestmentFlowBlock does not add ep_cost & offset to objective if Investment object is NonConvex

[jfreissmann]

Short description:
While probing the Investment object for the Issue #1172 and PR #1173, it seems like another issue came to the attention of @maltefritz and me: When e.g. a Converter has a solph.Investment as the nominal_capacity with ep_cost, offset and nonconvex set, while the Flow itself is not set to be NonConvex (e.g. when the min parameter should be enforced), the ep_cost and offset are not added to the objective function. See below for an example of a Converter producing the error:

hp_expensive = solph.components.Converter(
    'heat pump expensive',
    inputs={b_el: solph.Flow()},
    outputs={
        b_th: solph.Flow(
            nominal_capacity=solph.Investment(
                ep_costs=100,
                offset=2000,
                maximum=1,
                minimum=0,
                nonconvex=solph.NonConvex()
            ),
            variable_costs=1
        )
    },
    conversion_factors={b_el: [3]}
)

To Reproduce:
You can find a full example in the gist below:

https://gist.github.com/jfreissmann/a7165f9af517a1a8518a70e809a20186

The objective function produces from the energy system model above looks like this:

min
objective:
+50 InvestmentFlowBlock_invest(heat_pump_cheap_heat_bus_0)
+1 flow(heat_pump_expensive_heat_bus_0)
+1 flow(heat_pump_cheap_heat_bus_0)

It does not include the ep_cost or the offset of the hp_expensive Converter. Therefore, the solution of the optimization is to build the hp_expensive Converter for free and only pay its variable_cost to cover the heat demand. This also can be observed via the 'sequences' of the 'heat bus' node. The 'scalars' do not even exist.

Setup:

• OS: Windows 11
• Python 3.11
• oemof.solph dev branch latest