EDGESMARTGRID.CZ • AUTONOMOUS DISPATCHING

Edge AI Power Grid Orchestration and Control

GIP (Grid Intelligent Platform) hosted on EdgeSmartGrid.cz provides an autonomous control environment designed for orchestrating entire power networks and standalone substations. It combines RES forecast analytics, BESS optimization, and microsecond response during grid faults.

Edge AI & Forecasting

Local data processing on NVIDIA Jetson without cloud dependency. Real-time analysis of weather forecasts, load profiles, and dynamic rating directly inside the substation.

Hard Real-Time Control

Seamless integration with Beckhoff IPCs (TwinCAT 3) for millisecond loop cycles and FPGA hardware gates providing nanosecond reaction speeds using IEC 61850 (GOOSE/SV).

Uncompromising OT Security

Full alignment with IEC 62443 (SL 2/3), ISO 27001, and EU AI Act (Annex III) standards via hardware crypto-modules (Secure Box/TEE) and physical tamper shielding.

Hybrid Hardware Architecture

Deterministic safety combined with advanced artificial intelligence

NVIDIA Jetson / TEE Node

AI & Analytics Layer

Model Execution: DeepSeek-R1:70B quantised models running locally for network state orchestration.
Secure Environment: Hardware TEE (Trusted Execution Environment) protecting neural network weights and cryptography keys.
Predictive Dispatch: Dynamic adjustment of generation limits and BESS charging/discharging curves based on weather models.

Beckhoff PLC & FPGA Hard Guard

Deterministic OT Layer

Hard Real-Time: Beckhoff TwinCAT 3 executing critical safety loops independently of AI state.
FPGA Interlock: Sub-millisecond hardware override when AI confidence drops below safety threshold or grid voltage fails.
IEC 61850 Native: Native processing of GOOSE messages and Sampled Values without OS overhead.

Regulatory Framework & Standards

GIP platform is engineered to satisfy strict European infrastructure requirements

EU AI Act

High-Risk (Annex III)

Přísná shoda s článkem 14 (Lidský dohled) a článkem 15 (Přesnost, robustnost a hranice konfidence AI).

IEC 62443

Target Security Level 2/3

Bezpečný vývojový proces dle IEC 62443-4-1 a technické zabezpečení komponent dle IEC 62443-4-2.

NIS2 Directive

Kritická infrastruktura

Vestavěná detekce incidentů, hardwarové šifrování, audit dodavatelského řetězce a správa zranitelností.

ISO 27001

Norma ISMS

Kompletní řízení bezpečnosti provozního softwaru, vývojářských přístupů a životního cyklu kryptografických klíčů.

EU AI Act Article 15 Demonstrator

AI Confidence & Fallback Simulator

Adjust the AI Rate of Confidence slider to observe autonomous control handovers between AI and PLC/FPGA hardware logic.

AI Rate of Confidence (Confidence Score): 92%

50% (Uncertainty) 90% Safety Threshold 100% (High Certainty)

Autonomous AI Dispatch Active

AI model predicts optimal BESS dispatch and grid topology. Rate of confidence satisfies EU AI Act safety thresholds.

Active Controller: NVIDIA Jetson / TEE Agent Latency: < 15 ms