Convergence for power flow analysis shows inconsistent properties

Paulnkk commented 4 months ago

Feature Checklist

[x] Searched the issues page for similar reports
[x] Read the relevant sections of the documentation
[x] Browse the tutorials and tests for usefull code snippets and examples of use

Issue

Hey,

I am currently running power flow analysis in order to generate the admittance matrix Ybus:

# Convert MATPOWER case to pandapower network
# mat_file_path is file to tree files
net = pc.from_mpc(
        mat_file_path, f_hz=50, casename_mpc_file="mpc", validate_conversion=False
    )
# Run power flow to calculate the Ybus matrix
pp.runpp(net)
Ybus = net._ppc["internal"]["Ybus"]
# Extract the conductance (G) and susceptance (B) matrices
G = Ybus.real
B = Ybus.imag

Now I am facing the issue, when I run those lines for the "case14Tree.m" that I receive the following error message:

/Users/PaulKandora/opt/anaconda3/envs/copopf/lib/python3.12/site-packages/pandapower/create.py:5049: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation.
  net[table] = pd.concat([net[table], dd[dd.columns[~dd.isnull().all()]]], sort=False)
/Users/PaulKandora/opt/anaconda3/envs/copopf/lib/python3.12/site-packages/pandapower/converter/pypower/from_ppc.py:199: RuntimeWarning: divide by zero encountered in divide
  max_i_ka = ppc['branch'][:, 5]/ppc['bus'][to_bus, BASE_KV]/np.sqrt(3)
/Users/PaulKandora/opt/anaconda3/envs/copopf/lib/python3.12/site-packages/pandapower/converter/pypower/from_ppc.py:210: RuntimeWarning: divide by zero encountered in divide
  c_nf_per_km=(ppc['branch'][is_line, BR_B]/Zni[is_line]/omega*1e9/2).real,
/Users/PaulKandora/opt/anaconda3/envs/copopf/lib/python3.12/site-packages/pandapower/converter/pypower/from_ppc.py:210: RuntimeWarning: invalid value encountered in divide
  c_nf_per_km=(ppc['branch'][is_line, BR_B]/Zni[is_line]/omega*1e9/2).real,
/Users/PaulKandora/opt/anaconda3/envs/copopf/lib/python3.12/site-packages/pandapower/converter/pypower/from_ppc.py:277: FutureWarning: Setting an item of incompatible dtype is deprecated and will raise an error in a future version of pandas. Value '[]' has dtype incompatible with int64, please explicitly cast to a compatible dtype first.
  branch_lookup.loc[~is_line, "element"] = idx_trafo
numba cannot be imported and numba functions are disabled.
Probably the execution is slow.
Please install numba to gain a massive speedup.
(or if you prefer slow execution, set the flag numba=False to avoid this warning!)
/Users/PaulKandora/opt/anaconda3/envs/copopf/lib/python3.12/site-packages/pandapower/build_branch.py:164: RuntimeWarning: invalid value encountered in divide
  branch[f:t, BR_R] = line["r_ohm_per_km"].values * length_km / baseR / parallel
/Users/PaulKandora/opt/anaconda3/envs/copopf/lib/python3.12/site-packages/pandapower/build_branch.py:165: RuntimeWarning: invalid value encountered in divide
  branch[f:t, BR_X] = line["x_ohm_per_km"].values * length_km / baseR / parallel
/Users/PaulKandora/opt/anaconda3/envs/copopf/lib/python3.12/site-packages/pandapower/build_branch.py:194: RuntimeWarning: invalid value encountered in multiply
  b = 2 * net.f_hz * math.pi * line["c_nf_per_km"].values * 1e-9 * baseR * length_km * parallel
/Users/PaulKandora/opt/anaconda3/envs/copopf/lib/python3.12/site-packages/pandapower/build_branch.py:206: RuntimeWarning: invalid value encountered in multiply
  branch[f:t, RATE_A] = max_load / 100. * max_i_ka * df * parallel * vr
/Users/PaulKandora/opt/anaconda3/envs/copopf/lib/python3.12/site-packages/pandapower/build_bus.py:487: RuntimeWarning: invalid value encountered in divide
  v_ratio = (ppc["bus"][bus_lookup[s["bus"].values], BASE_KV] / s["vn_kv"].values) ** 2 * base_multiplier
/Users/PaulKandora/opt/anaconda3/envs/copopf/lib/python3.12/site-packages/pandapower/pypower/makeBdc.py:97: RuntimeWarning: invalid value encountered in divide
  b = stat / branch[:, BR_X]  ## series susceptance
/Users/PaulKandora/opt/anaconda3/envs/copopf/lib/python3.12/site-packages/scipy/sparse/linalg/_dsolve/linsolve.py:293: MatrixRankWarning: Matrix is exactly singular
  warn("Matrix is exactly singular", MatrixRankWarning)
/Users/PaulKandora/opt/anaconda3/envs/copopf/lib/python3.12/site-packages/pandapower/pypower/makeYbus.py:82: RuntimeWarning: invalid value encountered in divide
  Ysf = stat / (branch[:, BR_R] + 1j * branch[:, BR_X])  ## series admittance
/Users/PaulKandora/opt/anaconda3/envs/copopf/lib/python3.12/site-packages/pandapower/pypower/dSbus_dV.py:34: RuntimeWarning: invalid value encountered in divide
  diagVnorm = sparse((V / abs(V), (ib, ib)))
Traceback (most recent call last):
  File "/Users/PaulKandora/PyCharm_projects/CopOpf/copopf/test_soc_on_large_instances.py", line 124, in <module>
    opf_data = extract_opf_data('Tree_Library/case14Tree.m')
               ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/Users/PaulKandora/PyCharm_projects/CopOpf/copopf/extract_opf_data_from_matlab.py", line 13, in extract_opf_data
    pp.runpp(net)
  File "/Users/PaulKandora/opt/anaconda3/envs/copopf/lib/python3.12/site-packages/pandapower/run.py", line 250, in runpp
    _powerflow(net, **kwargs)
  File "/Users/PaulKandora/opt/anaconda3/envs/copopf/lib/python3.12/site-packages/pandapower/powerflow.py", line 86, in _powerflow
    _ppci_to_net(result, net)
  File "/Users/PaulKandora/opt/anaconda3/envs/copopf/lib/python3.12/site-packages/pandapower/powerflow.py", line 190, in _ppci_to_net
    raise LoadflowNotConverged("Power Flow {0} did not converge after "
pandapower.powerflow.LoadflowNotConverged: Power Flow nr did not converge after 10 iterations!

I assumed that the issue was related to dimensionality of the dataset but when I run the code on higher dimensional cases such as "case30Tree.m", I receive the correct results. Probably, it is not the best approach to generate admittance matrix (grateful if you know a better way to do so) but not sure how to resolve this issue.

Thanks in advance and best regards,

Paul

Label

[ ] Relevant labels are selected

SteffenMeinecke commented 4 months ago

Dear @Paulnkk, I agree that numerical issues might arise when trying to solve the power flow with the converted grid data. You can try to run pp.diagnostic(net) after receiving the data from matpower conversion. Maybe some r, x, or b values are extreme, for some reason.

Paulnkk commented 4 months ago

@SteffenMeinecke thanks a lot for the suggestion, I will try the approach you proposed and try to fix the issue

e2nIEE / pandapower