Some variable bounds are set multiple times

[odow]

I'm looking at making these variable bounds lazy:

PowerOperationsModels.jl/src/ac_transmission_models/AC_branches.jl

Lines 211 to 234 in 0b29d0d

	function branch_rate_bounds!(
	container::OptimizationContainer,
	::DeviceModel{B, T},
	network_model::NetworkModel{<:AbstractPowerModel},
	) where {B <: PSY.ACTransmission, T <: AbstractBranchFormulation}
	time_steps = get_time_steps(container)
	net_reduction_data = get_network_reduction(network_model)
	all_branch_maps_by_type = net_reduction_data.all_branch_maps_by_type
	for var in _get_flow_variable_vector(container, network_model, B)
	for (name, (arc, reduction)) in PNM.get_name_to_arc_map(net_reduction_data, B)
	# TODO: entry is not type stable here, it can return any type ACTransmission.
	# It might have performance implications. Possibly separate this into other functions
	reduction_entry = all_branch_maps_by_type[reduction][B][arc]
	# Use the same limit values as FlowRateConstraint for consistency.
	limits = get_min_max_limits(reduction_entry, FlowRateConstraint, T)
	for t in time_steps
	@assert limits.min <= limits.max "Infeasible rate limits for branch $(name)"
	JuMP.set_upper_bound(var[name, t], limits.max)
	JuMP.set_lower_bound(var[name, t], limits.min)
	end
	end
	end
	return
	end

and I ran into an issue: this function loops over multiple occurrences of the same var[name, t].

Is this expected (and just a minor performance issue to be re-setting existing bounds)? Or is it a bug somewhere?

[jd-lara]

This is an unexpected side effect of the reductions if two branches are reduced in series. cc. @m-bossart