PowerFlows.jl unified develop branch
#53
Conversation
Those sections being
Add fields to `PowerFlowEvaluationModel`s to support further PSI integration
Move reimporting code to PowerSystems, refine metadata format, misc code cleanup
| ) | ||
|
|
||
| # Shapes to reuse | ||
| zeros_bus_time = () -> zeros(Float64, n_buses, time_steps) |
There was a problem hiding this comment.
There is no need to have anonymous functions do this. The compiler can't infer this until call time. Just use zeros(Float64, n_buses, time_steps)
src/newton_ac_powerflow.jl
Outdated
| data.bus_angles[pq, t] .= angle.(V[pq]) | ||
| data.bus_angles[pv, t] .= angle.(V[pv]) | ||
| else | ||
| data.bus_activepower_injection[:, t] .= NaN64 |
| @@ -270,7 +258,7 @@ function _power_redistribution_ref( | |||
| @debug "Remaining residual $q_residual, $(PSY.get_name(bus))" | |||
| p_set_point = PSY.get_active_power(device) + p_residual | |||
| PSY.set_active_power!(device, p_set_point) | |||
| p_limits = PSY.get_reactive_power_limits(device) | |||
| p_limits = get_reactive_power_limits_for_power_flow(device) # TODO should this be active_power_limits? It was reactive in the existing codebase | |||
src/ac_power_flow_jacobian.jl
Outdated
| @@ -24,18 +24,19 @@ function PolarPowerFlowJacobian(data::ACPowerFlowData, x0::Vector{Float64}) | |||
| end | |||
|
|
|||
| function _create_jacobian_matrix_structure(data::ACPowerFlowData) | |||
| time_step = 1 # TODO placeholder time_step | |||
There was a problem hiding this comment.
Resolution: the NLSolve implementation of AC power flow does not support multi period, so @rbolgaryn will implement code to throw an error if the user tries to do that, and then all these time_step placeholders can be deleted.
There was a problem hiding this comment.
Right now, it is not even possible that the NLSolve is used with multiperiod power flow (see below). I think we can just safely remove it without raising an error.
# Multiperiod power flow
function solve_powerflow!(
data::ACPowerFlowData;
kwargs...,
)
pf = ACPowerFlow() # todo: somehow store in data which PF to use (see issue #50)There was a problem hiding this comment.
time_step can remain in the functions because it is possible to provide time_step as an optional input to set which time step to use for the data
The
hrgks/psse_exporter_psy4branch has become a unifieddevelop-type branch for lots of new PowerFlows.jl functionality. It now contains the PSS/E exporter implementation as well as some new power flow implementations from @rbolgaryn. Pull requests include:PowerFlowDatafields for multi-period AC PF #78plus some more commits only in this branch.