Safety First
Because your battery is an investment, not an experiment.
Batteries are chemistry.Chemistry doesn't forgive.
Push a lithium cell too hard and it degrades silently. Over-discharge and capacity fades. Thermal runaway is rare but real. A $15,000 investment demands respect.
Aggressive trading algorithms can destroy battery health chasing short-term gains. We built AET differently.
The optimizer proposes.The safety system disposes.
Profit never compromises battery health.
Four Independent Layers
Every setpoint must pass all four gates before reaching hardware. Any layer can veto. Defense in depth means no single point of failure.
Policy Compiler
Human intent to Machine constraintsYour policy compiles to machine-checkable constraints. Priority tiers: Safety (inviolable), Regulatory (legal), Economic (preferences). The compiler proves consistency before the system starts.
Physics Simulation
Predict before actingA digital twin simulates every proposed setpoint before execution. Models thermal dynamics, efficiency curves, SOH-adjusted capacity. If simulation shows any limit violation, the setpoint is rejected.
Formal Verification
Mathematically provenInterval arithmetic proves safety even with measurement uncertainty. State-space reachability analysis ensures the system cannot enter dangerous configurations. Mathematical proof, not probabilistic hope.
Guard Modules
Runtime protectionIndependent runtime guards—Thermal, Boundary, SOH—each with veto power. Even if simulation and verification pass, guards can reject setpoints based on real-time conditions.
Self-Learning Digital Twin
The model adapts to your specific battery. Online learning updates parameters as the battery ages. What worked at 100% SOH might damage at 80%.
Online parameter identification. Tracks internal resistance, capacity fade, and efficiency changes in real-time.
State estimation with uncertainty bounds. Knows what it doesn't know. Conservative when uncertain.
Thermal aging prediction. Every degree of temperature, every percentage of SOC has a degradation cost.
Hard Invariants
These limits are never violated. Not for any price. Not for any profit.
| Invariant | LFP | NMC |
|---|---|---|
| Max cell temp | 60C | 55C |
| Derating start | 50C | 45C |
| SOC bounds | 2-98% | 2-98% |
| Low SOH C-rate | 0.3C | 0.3C |
| Min rest time | 30s | 30s |
Maximize profit.Maximize lifespan.
These aren't competing goals—they're aligned. A battery that dies early costs more in lost revenue than any single trade could gain. The optimizer knows this.
Every trade is evaluated against its lifetime cost. Short-term greed is filtered out. What remains is sustainable profit that compounds over years, not days.