AC → DC → across the line → AC.

The teaching grids use balanced sinusoidal fundamentals at 50 hertz (Hz), meaning 50 cycles per second, and 220 kilovolts (kV) line-to-line root mean square (RMS). Each phase is 120° apart. The displacement power factor cos(φ) is 1: fundamental current is in phase with its corresponding phase voltage. For this case only, real power P = √3 × V_LL,rms × I_line,rms × cos(φ). The symbol φ is the voltage–current phase angle. This formula is not a switching-waveform harmonic or total-power-factor calculation. The waveforms below label a mathematical time axis, not the animation's slowed viewing speed.

Distinct high-voltage and converter-side bushings terminate the three phases. Windings exchange energy through alternating magnetic flux, not through a conducting bridge between circuits. In this simplified power balance the transformers are ideal, with zero assigned losses. Their ratios are scenario-matched abstractions, not fixed winding or tap specifications valid over every DC-voltage setting. Actual choices also depend on converter modulation margin, insulation, harmonics and grid conditions. Bushings insulate a live terminal from the tank; radiators and the conservator show oil-cooling/expansion functions, not a checked thermal design.

Three phase legs each contain an upper arm and a lower arm: six arms total. Each half-bridge submodule has two insulated-gate bipolar transistors (IGBTs), antiparallel diodes and a capacitor. An inserted cell contributes its capacitor voltage; a bypassed cell contributes approximately zero. Arm reactors limit rapid current change; control must also balance stored capacitor energy. The cutaway compresses each arm to six visible teaching cells, not a real installation's cell count or voltage rating. In this sending direction, controlled AC current draws real power from the grid. Half-bridge cells do not inherently block a DC fault through their diode paths; this normal-operation lesson does not model fault protection. Semiconductor/reactor losses become heat and require cooling.

This is a symmetrical monopole, not a bipole with independent pole converters and not an earth-return circuit. V_DC is the difference between the two conductors. The default sending value is 320 kV pole-to-pole: nominally +160 and −160 kV around a balanced common-mode reference. The reference/grounding controls are not simulated. For equal conductors, loop resistance R_loop = 2rL includes both full-length power conductors; r is resistance per kilometre per conductor, and L is one-way length. P_s,DC is sending DC power; V_s and V_r are sending/receiving line pair voltages. I_DC = P_s,DC/V_s; ΔV = I_DC R_loop; P_line_loss = I_DC²R_loop. Protective earthing and any shield wire are not the return and are not included in R_loop. The displayed line is drastically shortened; changing kilometres changes the calculation, not the scenery's scale.

The transmission calculation treats the interface as a four-terminal isolated power-balance block: output voltage is 0.75 times input voltage and output power is 0.98 times input power, both invented assumptions. Output current must change: I_out = P_out/V_out; it is not the line current. Its input and output negative rails are not connected. The separate dual-active-bridge (DAB) teaching circuit uses two active full bridges, a high-frequency transformer and coupling inductance. A relative bridge phase shift controls transfer. Its invented 48→96 V example is a lower-power switching principle, not equipment that handles this page's MW/kV link. Transmission-scale modular topology, insulation, balancing, protection and cooling would require a different, qualified design. The standard mode bypasses the entire stage with both conductors, not a parallel live branch.

The ideal staircase below demonstrates a modulation reference, not a diode rectifier's six-pulse waveform. Cell insertion changes in complementary upper/lower arms around an AC phase midpoint. A real MMC needs closed-loop current control, capacitor-energy balancing, filtering and coordination with the receiving grid; none is solved here. The steady-state loss factor is an invented 99% converter efficiency. Grid-side AC voltage/current measurements describe the balanced sinusoidal fundamental after the ideal transformer, not the internal switching voltage. The receiving network is assumed stiff at 50 Hz; this lesson does not simulate grid-forming control, synchronisation or a black start.

The same ideal-transformer simplification applies here: transmitted real power is preserved through the transformer; voltage adaptation changes line current. The physical illustration distinguishes bushings, cooling radiators, a tank, insulators and foundations. It does not represent full measurement wiring, surge protection, switching sequences, clearance coordination or an as-built yard. In the DC/DC mode a scenario-matched receiving transformer is assumed; the drawn model is not a fixed transformer proven suitable for every operating-point control setting.

The source fixes sending-grid real power and sending DC voltage for this calculation. Line current follows from those choices; receiving voltage follows from the loop's resistive drop. It is not a constant-receiving-voltage load-flow solver. Overall efficiency η = P_received/P_sent uses real power at the two AC grids, with all included losses subtracted exactly before display rounding. Reactive power is set to zero, ideal transformers have zero assigned losses, and auxiliary, corona and transient losses are omitted. Real plants have additional losses and constraints; this ledger is neither a performance prediction nor dispatch advice.