Skip to content

Commit

Permalink
Merge branch '202_nodal_capacity' into devel
Browse files Browse the repository at this point in the history
  • Loading branch information
@SanPen
SanPen committed Jun 3, 2024
2 parents 4ed6644 + 4c6ce18 commit 46329e0
Show file tree
Hide file tree
Showing 3 changed files with 58 additions and 33 deletions.
31 changes: 18 additions & 13 deletions src/GridCalEngine/Simulations/OPF/NumericalMethods/ac_opf.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -620,9 +620,8 @@ def ac_optimal_power_flow(nc: NumericalCircuit,
if opf_options.acopf_mode == AcOpfMode.ACOPFslacks:
nsl = 2 * npq + 2 * n_br_mon
# Slack relaxations for constraints
c_s = 1 * np.power(nc.branch_data.overload_cost[br_mon_idx] + 0.1,
1.0) # Cost squared since the slack is also squared
c_v = 1 * (nc.bus_data.cost_v[pq] + 0.1)
c_s = 1 * np.power(nc.branch_data.overload_cost[br_mon_idx] + 0.1, 1.0) # Cost squared since the slack is also squared
c_v = 1000 * (nc.bus_data.cost_v[pq] + 0.1)
sl_sf0 = np.ones(n_br_mon)
sl_st0 = np.ones(n_br_mon)
sl_vmax0 = np.ones(npq)
Expand Down Expand Up @@ -778,6 +777,14 @@ def ac_optimal_power_flow(nc: NumericalCircuit,
Sf = result.structs.Sf
St = result.structs.St
loading = np.abs(Sf) / (rates + 1e-9)

if opf_options.acopf_mode == AcOpfMode.ACOPFslacks:
overloads_sf = (np.power(np.power(rates, 2) + sl_sf, 0.5) - rates)*Sbase
overloads_st = (np.power(np.power(rates, 2) + sl_st, 0.5) - rates)*Sbase

else:
pass

hvdc_power = nc.hvdc_data.Pset.copy()
hvdc_power[hvdc_disp_idx] = Pfdc
hvdc_loading = hvdc_power / (nc.hvdc_data.rate + 1e-9)
Expand All @@ -787,9 +794,7 @@ def ac_optimal_power_flow(nc: NumericalCircuit,
tap_phase[k_tau] = tapt
Pcost = np.zeros(nc.ngen)
Pcost[gen_disp_idx] = c0 + c1 * Pg[gen_disp_idx] + c2 * np.power(Pg[gen_disp_idx], 2.0)

Pcost[gen_nondisp_idx] = c0n + c1n * np.real(Sg_undis) + c2n * np.power(np.real(Sg_undis), 2.0)

nodal_capacity = slcap * Sbase

if opf_options.verbose > 0:
Expand Down Expand Up @@ -885,19 +890,19 @@ def ac_optimal_power_flow(nc: NumericalCircuit,

if opf_options.acopf_mode == AcOpfMode.ACOPFslacks:
for k in range(n_br_mon):
if sl_sf[k] > opf_options.ips_tolerance:
logger.add_warning('Branch overload in the from sense',
if overloads_sf[k] > opf_options.ips_tolerance * Sbase:
logger.add_warning('Branch overload in the from sense (MVA)',
device=str(br_mon_idx[k]),
device_property="Slack",
value=str(sl_sf[k]),
expected_value=f'< {opf_options.ips_tolerance}')
value=str(overloads_sf[k]),
expected_value=f'< {opf_options.ips_tolerance*Sbase}')

if sl_st[k] > opf_options.ips_tolerance:
logger.add_warning('Branch overload in the to sense',
if overloads_st[k] > opf_options.ips_tolerance * Sbase:
logger.add_warning('Branch overload in the to sense (MVA)',
device=str(br_mon_idx[k]),
device_property="Slack",
value=str(sl_st[k]),
expected_value=f'< {opf_options.ips_tolerance}')
value=str(overloads_st[k]),
expected_value=f'< {opf_options.ips_tolerance*Sbase}')

for i in range(npq):
if sl_vmax[i] > opf_options.ips_tolerance:
Expand Down
6 changes: 4 additions & 2 deletions src/GridCalEngine/Utils/NumericalMethods/ips.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -334,7 +334,7 @@ def interior_point_solver(x0: Vec,

feascond = calc_feascond(g=ret.G, h=ret.H, x=x, z=z)
converged = error <= gamma

maxdispl = 0
error_evolution = np.zeros(max_iter + 1)
feascond_evolution = np.zeros(max_iter + 1)

Expand Down Expand Up @@ -414,7 +414,7 @@ def interior_point_solver(x0: Vec,
feascond = max([g_norm, max(ret.H)]) / (1 + max([np.linalg.norm(x, np.Inf), z_norm]))
gradcond = np.linalg.norm(lx, np.Inf) / (1 + max([lam_norm, mu_norm]))
error = np.max([feascond, gradcond, gamma])

maxdispl = np.max(np.r_[dx, dlam, dz, dmu])
z_inv = diags(1.0 / z)
mu_diag = diags(mu)

Expand All @@ -432,6 +432,7 @@ def interior_point_solver(x0: Vec,
print("INEQ:\n", ineq_df)
print("\tGamma:", gamma)
print("\tErr:", error)
print("\tMax Displacement:", maxdispl)

# Add an iteration step
iter_counter += 1
Expand All @@ -449,6 +450,7 @@ def interior_point_solver(x0: Vec,
print(f"\tF.obj: {ret.f * 1e4}")
print(f"\tErr: {error}")
print(f'\tIterations: {iter_counter}')
print(f'\tMax Displacement: {maxdispl}' )
print(f'\tTime elapsed (s): {t_end - t_start}')
print(f'\tFeas cond: ', feascond)

Expand Down
54 changes: 36 additions & 18 deletions src/trunk/acopf/acopf_run.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -8,6 +8,7 @@
from GridCalEngine.Simulations.NodalCapacity.nodal_capacity_options import NodalCapacityOptions
import numpy as np
import pandas as pd
import math
from GridCalEngine.enumerations import NodalCapacityMethod


Expand Down Expand Up @@ -398,21 +399,31 @@ def casehvdc():
cwd = os.getcwd()

# Go back two directories
new_directory = os.path.abspath(os.path.join(cwd, '..', '..', '..'))
new_directory = os.path.abspath(os.path.join(cwd, '..', '..', '..', '..'))

file_path = os.path.join(new_directory, 'Grids_and_profiles', 'grids', 'entrada_a_aopf.raw')
file_path = os.path.join(new_directory, 'REE Grids', 'entrada_a_aopf.raw')

grid = gce.FileOpen(file_path).open()

for gen in grid.generators:
gen.qmin_set = -0.8 * gen.Snom
gen.qmax_set = 0.8 * gen.Snom

options = gce.PowerFlowOptions(gce.SolverType.NR, verbose=False)
power_flow = gce.PowerFlowDriver(grid, options)
power_flow.run()

opf_options = gce.OptimalPowerFlowOptions(solver=gce.SolverType.NONLINEAR_OPF, acopf_mode=AcOpfMode.ACOPFslacks,
opf_options = gce.OptimalPowerFlowOptions(solver=gce.SolverType.NONLINEAR_OPF, acopf_mode=AcOpfMode.ACOPFstd,
verbose=1, ips_iterations=150, ips_tolerance=1e-8)
pf_options = gce.PowerFlowOptions(solver_type=gce.SolverType.NR, verbose=3)
run_nonlinear_opf(grid=grid, pf_options=pf_options, opf_options=opf_options, plot_error=True, pf_init=True)

opf_options = gce.OptimalPowerFlowOptions(solver=gce.SolverType.NONLINEAR_OPF, ips_tolerance=1e-8,
ips_iterations=50, verbose=1, acopf_mode=AcOpfMode.ACOPFslacks)
res = run_nonlinear_opf(grid=grid, pf_options=pf_options, opf_options=opf_options, plot_error=True, pf_init=True,
optimize_nodal_capacity=False,
nodal_capacity_sign=-1.0,
capacity_nodes_idx=np.array([10]))
#run_nonlinear_opf(grid=grid, pf_options=pf_options, opf_options=opf_options, plot_error=True, pf_init=True)
print('')

def caseREE():
"""
Expand All @@ -423,11 +434,15 @@ def caseREE():
# Go back two directories
new_directory = os.path.abspath(os.path.join(cwd, '..', '..', '..', '..'))

# file_path = os.path.join(new_directory, 'REE Grids', 'entrada_a_aopf.raw')
file_path = 'C:/Users/J/Documents/ree_opf/entrada_a_aopf.raw'
file_path = os.path.join(new_directory, 'REE Grids', 'entrada_a_aopf.raw')
# file_path = 'C:/Users/J/Documents/ree_opf/entrada_a_aopf.raw'

grid = gce.FileOpen(file_path).open()

for gen in grid.generators:
gen.qmin_set = -0.8 * gen.Snom
gen.qmax_set = 0.8 * gen.Snom

disp_areas = ['A11', 'A15']
dict_bus_lims = {'21215': [230, 225],
'11055': [410, 405],
Expand All @@ -436,20 +451,23 @@ def caseREE():
'15005': [410, 405],
'15015': [410, 405]}
tol = 1e-4
vm_cost = 1e2
vm_cost = 1e4
i = 0
for gen in grid.generators:
if gen.bus.area.name in disp_areas:
# P limits -> restrict them very close to P
print(f'Select generator {i}')
gen.Pmax = gen.P + tol
gen.Pmin = gen.P - tol
#gen.Pmax = gen.P #+ tol
#gen.Pmin = gen.P #- tol
# Tanmax -> set pf close to 0 to get large tanmax
gen.Pf = tol
#gen.Pf = tol
else:
gen.enabled_dispatch = False
gen.Pmax = gen.P + tol
gen.Pmin = gen.P - tol

#gen.enabled_dispatch = False
gen.Pmax = gen.P #+ tol
gen.Pmin = gen.P #- tol
#gen.Qmax = abs(math.tan(math.acos(gen.Pf)) * gen.P)
#gen.Qmin = -abs(math.tan(math.acos(gen.Pf)) * gen.P)

# i += 1

Expand All @@ -472,7 +490,7 @@ def caseREE():
power_flow = gce.PowerFlowDriver(grid, options)
power_flow.run()

opf_options = gce.OptimalPowerFlowOptions(solver=gce.SolverType.NONLINEAR_OPF, acopf_mode=AcOpfMode.ACOPFstd,
opf_options = gce.OptimalPowerFlowOptions(solver=gce.SolverType.NONLINEAR_OPF, acopf_mode=AcOpfMode.ACOPFslacks,
verbose=1, ips_iterations=100, ips_tolerance=1e-8)
pf_options = gce.PowerFlowOptions(solver_type=gce.SolverType.NR, verbose=3)
run_nonlinear_opf(grid=grid, pf_options=pf_options, opf_options=opf_options, plot_error=True, pf_init=True)
Expand Down Expand Up @@ -508,13 +526,13 @@ def case_nodalcap():
# linn5bus_example()
# two_grids_of_3bus()
# case9()
case14_linear_vs_nonlinear()
#case14_linear_vs_nonlinear()
# case14()
# case_gb()
# case6ww()
# case_pegase89()
# case300()
# casepegase13k()
# casehvdc()
# caseREE()
#casehvdc()
caseREE()
# case_nodalcap()

0 comments on commit 46329e0

Please sign in to comment.