N-1 & transfer

This page answers what happens when one branch is lost, and whether the load can be picked up again. Two engine products are involved: the study "n1", which trips branches one at a time and reports dead load together with the SAIDI and SAIFI reliability indices, and studyOptions.enableN1Transfer in engines/n1-transfer.js, which for every N-1 scenario also searches the tie-switch combinations that could restore supply (single switch first, two switches only if one is not enough, with a hard cap on the combinations), solves the load flow for real for each candidate, and returns the feasibility, the violation list, the buses left without supply and an overall summary. Doing nothing is always kept as a candidate, and a busbar in a connected component that contains no source is reported as having no meaningful voltage rather than being given a phantom number.

N-1 is the language of reliability: can the scheme survive the loss of one transformer or one feeder, how much load is dropped, how long is the interruption, and does a tie switch or a spare incoming line have to be added. It is a standard requirement in substation design review and it feeds the reliability indices a client asks for in a tender. It also protects the designer from the opposite mistake — closing a tie switch that restores supply but overloads the remaining transformer, which is exactly what the candidate-by-candidate load-flow evaluation exposes.

Input: the topology with the branch switch states, the list of branches to trip (if none is given, every energised branch is scanned), the candidate tie switches (if none is given, the normally open branches are identified automatically), and the thresholds (branch loading limit default 100%, voltage band default ±7%, that is 0.93 and 1.07 pu) → chain: a branch is tripped, the existing load flow is solved for real, the connected components are examined topologically and any component without a source is marked as not energised so its busbars are excluded from the voltage, violation and loss statistics, then each candidate tie-switch combination is closed and the load flow is solved again → the combinations are grouped by whether they restore every dead busbar, the group that does is decided first (supply first, and it is reported honestly if the restored network ends up overloaded rather than being suppressed), otherwise the group containing doing nothing is used, and inside a group the order is the fewest violations, then the smallest total violation magnitude, then the lowest network loss, then the fewest actions, then the lexicographic plan string → output: feasibility, the violation list, the islanded buses and the summary, together with the SAIDI and SAIFI indices from the reliability study. Linkage: adding a tie switch adds restore options and changes the feasibility verdict; a cable section or transformer capacity change decides whether the restored network is overloaded; the load level decides how much load is at risk and moves the reliability indices; multiple sources change which buses are considered energised and how many restore paths exist. Approximations, all disclosed: every candidate combination is solved with the real load flow and no sensitivity approximation is used; when a branch has no rating or no ampacity its loading percentage is null and it is flagged as unknown instead of being judged; the combination search is capped and the truncation is flagged.