From the documents

• docAll 211 rod positions, their types (АР, УСП) and the 12 height-detector cells, from Karpan's maps for 00:39 and 01:22:30 (pp. 362–363), cross-checked against INSAG-7 Fig. II-6. Plant grid labels 66…12 / 11…65.
• docPre-accident axial flux shape (INSAG-7 Fig. II-5, curve 4); coolant flow 56 800 → 51 000 m³/h during the run-down; rundown pumps 13, 14, 23, 24 and their trips at 01:23:42.6–43.0 (Karpan App. 3).
• docTeletype: two RTA-80s printing 7 characters per second, time without a leading zero ("1.23.39 АВАР ЗАЩИТА 5"), cut off when power failed (Karpan pp. 359–361). DREG: low-priority 2–4 s cycles numbered "№NNNД", "=0" marks a signal clearing, 9 rod positions logged once a minute (INSAG-7 p.77).
• docPRIZMA: SKALA recalculated core parameters about every 5 minutes; the duty engineer printed PRIZMA every two hours on odd hours and carried it to the control room (Karpan p.350). ORM was never on a continuous display.
• docThe Архив tab reproduces the surviving teletype and DREG lines of 26 April 1986.

Reconstructed or invented

• recRequest device: the two keyboards, the two display sets (value, deviation from the nearest setpoint, code), the code grammar, the ПРМ/ПП/АП keys, ИЗМ and ЗАП, error codes 301 and 303, and the codes П0N0000, Т0N7100, К0S0000, К5N0000 and К0S/К0G + channel follow the Chornobyl Family reconstruction, built with former Chernobyl staff. It models the modernized SKALA of unit 3, which ran until 2000, so the 1986 codes of unit 4 may have differed in detail; the builders say the principles stayed largely the same. In the control room the keyboard and the displays were separate pieces; here they share one panel, as on the computer-hall console. Every operator had this device, and SKALA's engineers had the same one. The plant grid, which never uses the digits 8 or 9, confirms that channel coordinates are octal. Reading ПС/АС as the falling setpoints is a guess, and the unlabelled keys do nothing here.
• invNot an emulator: the V-3M instruction set and SKALA software are not public. All other request codes, the subprogram numbers (except 0301), the PRIZMA summary lines, the rod cartogram layout, the wording of the teletype's setpoint lines and the choice of the 9 logged rods are this simulator's own. The PRIZMA channel map follows the form shown in the Chornobyl Family printer video, without the per-channel flow and boiling margin the model lacks. The tube styling of the displays is a design choice.
• invPhysics: 28-slice axial one-group diffusion (quasi-static) with point kinetics, per-slice void, Doppler, xenon/iodine and a burnup profile; every rod's absorber, graphite displacer and water columns at its real depth, weighted by its radial position. Rods are radially averaged, so the local character of the excursion (rows 12–30, Karpan p.380) is not captured.
• invTuned, not derived: migration area, water-column absorption, burnup profile, and a boiling-crisis threshold after which channels rupture, rods jam and steam floods the core. The xenon profile at 01:22:30 is solved from the measured flux shape because it was never measured (INSAG-7 p.81).
• invFuel-channel positions are inferred (every lattice cell inside the core not used by a rod or detector). Their colouring is an illustrative radial estimate, not a 2D solution. The 175 plain rods are 139 РР manual, 12 ЛАР local-automatic and 24 АЗ emergency rods, which the sources do not tell apart. 24 of them read exactly 7 cm in both maps; they are probably fully withdrawn, but their scattered positions rule them out as the АЗ group, which would be laid out symmetrically like the АР and УСП rods.

Validation against the record

INSAG-7 (p.142) notes that none of the three Soviet 3D models reproduced the overpower signal within 3 seconds, and that every model was extremely sensitive to the starting axial distribution. This one is too: lowering the water-column absorption by 7 % (0.30 → 0.28) stops the 01:22:30 state from exploding, and raising it by 7 % makes the 00:39 rods trip. Matching the record here comes from tuning a far simpler model to the documents. It is not proof that the model is right.

Sources

• IAEA, INSAG-7, Russian edition, 1993 (PDF)
• Н.В. Карпан, «Чернобыль. Месть мирного атома», Kyiv, 2005
• DREG protocol transcript, L. Sturm and R. Plaskov (CC BY-NC 4.0)
• V. Alekseev, Чернобыль
• Chernobyl Witness, DREG parameter graphs
• Hackaday on the Chornobyl Family SKALA videos
• Habr, СКАЛА: ЭВМ ЧАЭС

Thank you

This reconstruction rests on the shoulders of experts and their work (and some impressive AI to tie it together).

• Nikolai Karpan, whose book «Чернобыль. Месть мирного атома» preserved the rod maps, shift logs and recorder data the core model is built on.
• The IAEA's International Nuclear Safety Advisory Group, whose INSAG-7 report supplied the measured flux shape, the rod positions to cross-check against, and a frank account of the reactor's design flaws.
• Chornobyl Family, whose restoration of real SKALA hardware shows how the console, printers and DREG actually behaved.
• Linus Sturm and R. Plaskov, for transcribing the DREG protocol of the last 40 seconds.
• Claude (Anthropic), the AI assistant that read the Russian sources side by side with the English ones, extracted the rod map from the book's typography, cross-checked it cell by cell against INSAG-7, and ran many simulation runs to test the model against every figure in the record.

Working this way was a lesson in what AI can do to aid historical reconstruction. Its value was not knowing the answers, but holding scattered Russian and English sources together, reconstructing them into tabulated data, and testing them against each other far faster than one person could. Every finding was checked against the source documents, and the gaps are marked accordingly.