Heian Platform — 2026 Launch

ZWEICHAIN

平安京

Crowdsourced earthquake intelligence powered by Lagrangian Tensor Generating Logic. Buildings as data. Seismic physics as open science.

In 794 AD, Emperor Kanmu moved the capital from Nara (Heijo-kyo) to Kyoto (Heian-kyo), beginning 400 years of cultural flourishing. In 2026, Zweichain moves from Heijo to Heian — a new platform era for seismic intelligence.
Targeting researchers at
Caltech Seismo Lab · UC Berkeley SSA · USGS · NTT Data · Japan Cabinet Office · J-REIT Operators
Equation of Motion — Building Structural System
M · ẍ(t) + C · ẋ(t) + K · x(t) = f(t)
M

Mass matrix — floor inertia per story level. Diagonal for classical shear buildings.

C

Damping matrix — energy dissipation (Rayleigh form: αM + βK). Identified from shake response.

K

Stiffness matrix — structural coupling tensor. TGL identifies this from sensor data.

f(t)

Ground excitation — accelerometer time series. Input from ZweiMedia iOS app or IoT floor sensors.

$150M
REIT outperformance vs index
within 2 weeks of ranking
4,000
Buildings ranked across
60 Japanese REITs
#1
Best Proptech, Plug and Play
Tech Summit 2020
~$800M
NMZ token market cap
listed on CEX

Namazu 3.0
鯰 — Earthquake Intelligence

Smartphones and IoT floor sensors become a nationwide structural health monitoring network. Each building is treated as a coupled oscillator system — sensors on each floor, TGL algorithms identify the full structural tensor (K, C, M matrices) governing inter-floor seismic response.

The tensor, stored in an .nmz file, enables per-building digital twin simulation — not statistical grid estimates, but physics-first structural identification at building resolution.

TGL Algorithm REIT Intelligence Digital Twin CAT Bond Ready Real-Time Alerts
TGL — Floor Response Tensor (3-Story Example)
F3
k=1.12 · c=0.07 · peak×2.4
×2.4
F2
k=0.98 · c=0.07 · peak×1.7
×1.7
F1
k=0.86 · c=0.08 · peak×1.0
×1.0
GND
f(t) — ground excitation input
in
1

POST /v3/shake — accelerometer float array from iOS / IoT sensor

2

TGL computation — TensorFlow identifies K, C matrices via optimization

3

Generate .nmz — structural tensor file with building_class, model_id

4

Simulation — run 3.11 or Nankai scenario against the model

File Format

.nmz — Namazu Structural Tensor

A portable binary/JSON descriptor encoding the full structural identity of a building: mass, stiffness, and damping matrices plus sensor geometry and confidence scores. Named like NMZ-Z-W3B-1778780307.

NMZ-Z-W3B-1778780307.nmz .NMZ v1.0
{
  "model_id":      "NMZ-Z-W3B-1778780307",
  "tgl":            "Z-TGL",
  "building_class": "W3B",         // Wood, 3-story, Base-isolated
  "generated":      1778780307,      // UNIX epoch

  "M": [[1.00,0,0],[0,0.92,0],[0,0,0.85]],  // Mass matrix
  "K": [[2.24,-1.12,0],[-1.12,2.10,-0.98],[0,-0.98,0.86]],  // Stiffness
  "C": [[0.15,-0.07,0],[-0.07,0.14,-0.07],[0,-0.07,0.08]],  // Damping

  "sensors": [
    { "floor": 0, "lat": 35.0116, "lng": 135.7681, "device": "ios" },
    { "floor": 1, "lat": 35.0116, "lng": 135.7681, "device": "iot" },
    { "floor": 2, "lat": 35.0116, "lng": 135.7681, "device": "iot" }
  ],

  "confidence": 0.94,
  "rmse":       0.0031,
  "scenario_311_score":    "Excellent",
  "scenario_nankai_score": "Good"
}

The Zweichain Network

Six interconnected layers — from raw seismic sensing to blockchain settlement, digital twin simulation, and real estate investment intelligence.

Open TGL Competition

Submit your own Tensor Generating Logic algorithm via GitHub URL. Run sandboxed in Docker against real building shake data. Ranked by RMSE against ground-truth floor accelerations. Reward: NMZ tokens.

Scoring Function

RMSE(ŷ, y) = √( 1/n · Σ (ŷ_i − y_i)² )

Where ŷ is your algorithm's predicted floor acceleration (next 500ms window) and y is the ground-truth sensor reading.

Input: accelerometer time series per floor · Output: predicted acceleration for target floor. Docker container receives shake data via stdin JSON, returns predictions.

GitHub URL submission Docker sandboxed NMZ reward Kaggle-style
→ View TGL Leaderboard

Three Phases

PHASE 1
NOW

Building Structural Identification

Crowdsourced seismic sensing → per-building tensor (K, C, M) identification. .nmz file format. TGL open competition. REIT portfolio scoring.

Active
PHASE 2
2027

Planetary Propagation Tensor

Normal mode seismology at continental scale. Spherical harmonic decomposition. .ppt file format — from building tensor to planetary tensor.

Next
PHASE 3
2028+

QPU Acceleration

Encode the seismic tensor as a quantum Hamiltonian. VQE / QAOA optimization on QPU. Quantum-native structural system identification.

Horizon
NTT Data (NDA signed) Japan Cabinet Office Tribeca Capital SoftBank Board Gov't Bank IoT Approved
"In the Heian period, Kyoto became the center of science, art, and governance for four centuries. We move our platform — from Nara's Heijo to Kyoto's Heian-kyo — to build the infrastructure that makes every building in Japan seismically legible."
— Zweichain, May 2026
710 AD
Heijo-kyo
794 AD
Heian-kyo
2021
Heijo API
2026
Heian Platform

NMZ — Namazu Token

Merit-based reward system. TGL algorithm developers, building data contributors, and seismic sensor operators earn NMZ. Catastrophe bonds, REIT risk scoring, and insurance products settle on-chain via Zweichain smart contracts.

Market Cap
~$800M USD
Listed on CEX. NMZ token powers the Namazu ecosystem incentive layer.
TGL Reward
RMSE ranked
Leaderboard-based. Best algorithms earn NMZ. Open to anyone — merit only.
CAT Bond
On-chain settlement
Catastrophe bonds trigger automatically from .nmz tensor scores. Validated by Tribeca Capital (global bank consortium).
API Auth
nmz_live_ token
Bearer token auth. nmz_live_... for production, nmz_test_... for sandbox.