Microgrid & islanding
Diesel island steady state (V-f master, PQ sources), bus transfer (BATS) with inrush and spare loading, island relay settings against the GB/T 33593 limits, and the DER hosting capacity of the busbar.
Select or create a scheme to open this page
Not computed — press Run
Genset as the island V-f master with PV/BESS/wind as PQ injection; load, overload and voltage per busbar.
Not computed — press Run
Three steady-state conditions (pre-fault / lost supply / after transfer) with recovered load, inrush and spare-source loading.
Not computed — press Run
Maximum DER that the busbar can host before the binding constraint is hit.
Not computed — press Run
U/f measurement against the GB/T 33593 thresholds, with each check state (trip / no trip / unknown).
About Microgrid & islanding
Four studies of the engine drive this page, all behind study switches: study.dieselIsland (the genset as the V-f slack of the island - the island case is solved by the same load-flow solver, genset output and bus voltages come straight from it, IEC 60034 for the machine and ISO 8528 practice for the droop estimate), study.busTransfer (BATS - the pre-fault, dead and post-transfer steady states, each solved by the same solver, with load lost and recovered, inrush peak, spare-source loading and the honest rule that a de-energised island is reported as null volts rather than a mathematical phantom voltage, GB/T 14285-2006), study.hostingCapacity (how much distributed generation the feeder can accept - stepwise real load flow plus bisection on the three constraints of bus voltage, branch thermal loading and short-circuit level; default voltage limit 1.07 pu, thermal limit 100 %) and study.islandRelay (anti-islanding settings against GB/T 33593-2017 with defaults Umax 1.10 pu, Umin 0.85 pu, fmax 50.5 Hz, fmin 49.5 Hz).
Standby and island operation is where a design loses its utility back-up: the genset is overloaded, the transfer never closes, the bus never recovers, or too much PV is connected for the feeder to accept. Reviewers ask for exactly these four verdicts, and the protection settings that trip the island must be shown to be effective - otherwise the scheme cannot be commissioned.
Input = topology plus studyInput.gridOperation (mode island, island master tag) and studyOptions.enableDieselIsland / enableBusTransfer / enableHostingCapacity / enableIslandRelayCheck. Chain: the island mode makes the genset node a V-theta slack with a near-zero internal impedance, all other DERs inject as PQ and the utility grid is treated as disconnected, so the same load-flow solver yields genset P/Q/S, current and bus voltages; the transfer study removes the main source path, solves the dead state, closes the tie and solves again to give the recovered load, the inrush peak and the spare loading; hosting capacity repeats the load-flow solve at every search step (iterations equal solver calls, auditable from the result) and bisects on the first binding constraint; the relay study reuses those island voltages and compares them with the settings, deriving frequency from the droop estimate and flagging it as droop-derived. Linkage: genset kVA moves the island voltages and the recovered load; the tie arrangement moves the inrush and the spare loading; DER size moves the hosting capacity, the fault level and therefore the protection settings.
Four lazy engine cards. Every "unknown" status below is the engine refusing to judge (missing rating/overload data) — no pass is ever invented.
Scheme has no topology yet — nothing to assess
Maximum DER active power the feeder can host; every search step is a real power flow (no sensitivity approximation)