Potential issue in get_bus_branch_connectivity_matrix()

[JosepFanals]

Hi Santiago,

Francesca Rossi, a PhD student at CITCEA, noticed the get_bus_branch_connectivity_matrix() method may not work as expected. The thing is, branch_list groups branch data for lines, transformers, etc. Therefore, the row index of Cf and Ct may be erroneous since the k index is not cumulative. I am pasting below a fraction of the original code (L3490 of multi_circuit.py):

        for branch_list in self.get_branch_lists():
            for k, br in enumerate(branch_list):
                i = bus_dict[br.bus_from]  # store the row indices
                j = bus_dict[br.bus_to]  # store the row indices
                Cf[k, i] = 1
                Ct[k, j] = 1

Instead, Francesca proposes to modify it for:

        k_len_cum = 0
        for branch_list in self.get_branch_lists():
            for k_1, br in enumerate(branch_list):
                k = k_1 + k_len_cum
                i = bus_dict[br.bus_from]  # store the row indices
                j = bus_dict[br.bus_to]  # store the row indices
                Cf[k, i] = 1
                Ct[k, j] = 1
            k_len_cum += len(branch_list)

From what I could check, the test suite runs equally on both cases, but I do believe the second block of code is the right one. What do you think?

[SanPen]

Hi, indeed there was a bug and I've just fixed it in devel with e197cd8.

However this function is not used for anything critical because it "lives" in the MultiCircuit.

Please advise Francesca that she should use Cf and Ct from the NumericalCircuit instead. Those account for branch and bus active and are the ones used for numerical processes.

[RossiFran]

Hi Josep, Hi Santiago,

Thank you for fixing the bug and for the clarification!