PHYSICAL INFRASTRUCTURE → DATA → MODELS → INTELLIGENCE → ASSURANCE → ACTION

_{Electrical Engineering • Critical Infrastructure • Digital Twins • Physics-Informed AI • Industrial Systems • Robotics • OT Security}

_{Software-defined engineering for systems that exist in the physical world.}

I design and build software-defined infrastructure for complex physical systems.

The physical layer is the starting point.

Electrical infrastructure, power systems, traction systems, industrial assets, machines and networks become the foundation on which software, telemetry, semantic models, simulation, artificial intelligence, cybersecurity and autonomous systems are built.

The central problem is not another isolated application.

It is integration.

flowchart LR
    P[Physical Infrastructure]
    E[Electrical / Industrial Systems]
    T[Sensing + Telemetry]
    D[Protocols + Data]
    S[Semantic Models]
    DT[Digital Twin]
    SIM[Simulation]
    AI[AI / ML / Optimization]
    A[Policy + Assurance]
    R[Automation / Robotics]
    X[Physical Action]

    P --> E --> T --> D --> S --> DT --> SIM --> AI --> A --> R --> X --> P

flowchart LR
    A[PHYSICS] --> B[ELECTRICAL ENGINEERING] --> C[DATA + TELEMETRY] --> D[DIGITAL TWINS] --> E[SIMULATION] --> F[AI / ML] --> G[AGENTS] --> H[ASSURANCE] --> I[AUTOMATION] --> J[PHYSICAL SYSTEMS]

Electrical and physical reality provide the constraints; software provides the representation; data provides observability; AI provides intelligence; assurance governs action.

The portfolio is positioned around several of the most active technical directions now emerging across power-system engineering and critical infrastructure.

Examples include DOE's 2026 GridFM 2.0 project targeting vastly higher grid-scenario throughput; DOE/LLNL's Stormbreaker testbed for LLM and agentic-AI evaluation in power/OT environments; GE Vernova's active Dynamic Grid Foundation Model project; LF Energy's AINETUS, Grid2Op and OpenGridFM activities; NIST's ongoing trustworthy-AI profile for critical infrastructure; and the 2026 IEC 61850 series release. DOE GridFM 2.0 · DOE Stormbreaker · GE Vernova DynaGridFM · LF Energy AINETUS · NIST Critical Infrastructure AI · IEC 61850:2026

Modern infrastructure is no longer divided cleanly into hardware and software.

An electrical asset can simultaneously be:

physical machine · power-system component · real-time system · OT endpoint · cybersecurity boundary · telemetry source · digital-twin object · ML dataset · agent environment · operational decision surface

The engineering objective is therefore:

flowchart LR
    ASSET[Asset]
    STATE[State]
    MODEL[Model]
    TWIN[Digital Twin]
    INTEL[Intelligence]
    ASSURE[Assurance]
    ACTION[Action]
    FEEDBACK[Measured Feedback]

    ASSET --> STATE --> MODEL --> TWIN --> INTEL --> ASSURE --> ACTION --> FEEDBACK --> STATE

The loop is intentionally closed.

A useful platform should not stop at visualization.

It should be capable of:

observe → model → simulate → predict → optimize → assure → act → measure → learn

flowchart LR

    subgraph P["PHYSICAL"]
        PA[Physical Assets]
        EE[Electrical Engineering]
        OT[Industrial Networks]
    end

    subgraph D["DIGITAL"]
        TEL[Telemetry]
        DATA[Data]
        SEM[Semantic Model]
        DT[Digital Twin]
    end

    subgraph I["INTELLIGENCE"]
        SIM[Simulation]
        ML[Physics-Informed AI]
        OPT[Optimization]
        AG[Agents]
    end

    subgraph A["ASSURANCE"]
        SEC[Cybersecurity]
        POL[Policy]
        VER[Verification]
        RUN[Runtime Assurance]
    end

    subgraph O["OPERATION"]
        EDGE[Edge]
        CTRL[Control]
        ROB[Robotics]
        HUM[Human / Operator]
    end

    PA --> EE --> TEL --> DATA --> SEM --> DT
    OT --> TEL
    DT --> SIM
    DT --> ML
    DT --> OPT
    DT --> AG
    SIM --> SEC
    ML --> POL
    OPT --> VER
    AG --> RUN
    SEC --> EDGE
    POL --> CTRL
    VER --> ROB
    RUN --> HUM
    EDGE --> PA
    CTRL --> PA
    ROB --> PA
    HUM --> PA

Grimaldi.ca is the strategic and visual surface.

GitHub is the engineering surface.

Together:

flowchart LR
    WEB[GRIMALDI.CA<br/>Strategic / Visual]
    GH[GITHUB<br/>Engineering]
    R[Research]
    S[Software]
    SYS[Systems]
    EN[Energy]
    AI[Intelligence]
    ROB[Robotics]
    CI[Intelligent Infrastructure]

    WEB --> GH
    GH --> R
    GH --> S
    GH --> SYS
    R --> EN
    S --> AI
    SYS --> ROB
    EN --> CI
    AI --> CI
    ROB --> CI

flowchart LR

    subgraph PHYSICAL["PHYSICAL DOMAIN"]
        HV[High Voltage]
        MV[Medium Voltage]
        TP[Traction Power]
        GRID[Power Networks]
        IND[Industrial Assets]
        ROB[Robotics]
    end

    subgraph DIGITAL["DIGITAL DOMAIN"]
        TEL[Telemetry]
        PROT[Protocols]
        DATA[Data Platforms]
        SEM[Semantic Models]
        TWIN[Digital Twins]
        SIM[Simulation]
    end

    subgraph INTEL["INTELLIGENCE"]
        ML[Machine Learning]
        PINN[Physics-Informed AI]
        OPT[Optimization]
        RL[Reinforcement Learning]
        AG[AI Agents]
    end

    subgraph ASSURE["ASSURANCE"]
        SEC[Cybersecurity]
        VER[Verification]
        OBS[Observability]
        POL[Policy]
        RT[Runtime Assurance]
    end

    subgraph ACTION["ACTION"]
        EDGE[Edge]
        AUTO[Automation]
        CTRL[Control]
        OP[Operator Systems]
    end

    HV --> TEL
    MV --> TEL
    TP --> TEL
    GRID --> TEL
    IND --> PROT
    ROB --> TEL

    TEL --> DATA
    PROT --> DATA
    DATA --> SEM
    SEM --> TWIN
    TWIN --> SIM

    SIM --> ML
    TWIN --> ML
    ML --> OPT
    OPT --> RL
    RL --> AG

    AG --> SEC
    OPT --> VER
    ML --> POL
    SEC --> RT
    VER --> RT
    POL --> RT

    RT --> EDGE
    RT --> AUTO
    RT --> CTRL
    RT --> OP

The engineering surface is designed to connect with the broader open engineering ecosystem rather than exist as an isolated island.

LF Energy · GridAPPS-D · OpenEMS · PyPSA · pandapower · OpenDSS · HELICS

Grid2Op · OpenGridFM · AINETUS · SOGNO · Power Grid Model · PowSyBl · Dynawo · OperatorFabric

IEC 61850 · CIM · IEC 61968 · IEC 61970 · IEC 62351 · IEC 62443

HELICS · RTDS · OPAL-RT · MATLAB/Simulink · OMNeT++ · scientific Python

Digital Twins · Runtime Assurance · OT Security · Functional Safety · HIL · SIL · Edge Computing

The goal is not to duplicate established ecosystems.

It is to build the engineering layers that connect them.

flowchart LR
    STD[Open Standards] --> MODEL[Open Models] --> SIM[Open Simulation] --> DATA[Open Data] --> SW[Open Software] --> INTEL[Intelligent Engineering]

LF Energy's 2026 ecosystem activity is particularly relevant here: AINETUS is explicitly designed to integrate with Grid2Op, OperatorFabric and SOGNO; Grid2Op and OpenGridFM moved toward Incubation; SEAPATH reached Graduation; and additional projects are expanding AI, DER interoperability and operational tooling.

Electrical infrastructure forms the physical foundation.

flowchart LR
    ASSET[Electrical Asset] --> MODEL[Electrical Model] --> DATA[Digital Representation] --> OBS[Observable System] --> INTEL[Intelligent System]

Modern electrical-system software is increasingly moving toward:

digital substations → DER-aware semantic models → grid-forming / inverter-dominated systems → real-time operational data → AI-assisted planning and operations → validated decision support

IEC's 2026 IEC 61850 series includes new material such as IEC 61850-7-410:2026, while IEC 61850-7-420:2021 provides information models for DER and distribution automation.

Machine intelligence becomes more useful for physical systems when it understands the systems it is modelling.

flowchart LR
    PHYS[Physics] --> DATA[Observations] --> MODEL[Hybrid Model] --> PRED[Prediction] --> OPT[Optimization] --> DECISION[Decision]

AI should not replace engineering constraints.

AI should operate inside them.

A major current direction is the combination of:

foundation models + graph structure + time-series data + physical constraints + simulation + uncertainty + operator workflows

DOE's September 2026 GridFM 2.0 initiative is explicitly pursuing large-scale foundation-model approaches for grid planning and expansion, while an active GE Vernova/Georgia Tech/PNNL project is developing a Dynamic Grid Foundation Model using generative AI, time-series modelling and graph neural networks.

The interesting problem is not simply whether an AI system can reason.

It is whether the system can reason inside a governed engineering environment.

flowchart LR
    HUMAN[Human / Operator] --> AGENT[AI / LLM / Agent] --> POLICY[Policy / Capability Boundary] --> TOOL[Validated Tool Interface] --> PHYS[Physics / Safety Constraints] --> ASSURE[Runtime Assurance] --> ACTION[Industrial / Physical Action] --> OBS[Observability] --> HUMAN

Agentic AI is moving into dedicated critical-infrastructure test environments rather than remaining purely a conversational technology.

DOE and Lawrence Livermore's 2026 Stormbreaker testbed is specifically designed to evaluate LLMs and agentic AI in power-system and OT environments. LF Energy's AINETUS similarly places AI decision support inside an established simulation and operator ecosystem instead of treating the agent as a standalone application.

A digital twin should move beyond static visualization.

flowchart LR
    STATE[Live State] --> TEL[Telemetry] --> TOPO[Topology] --> SEM[Semantic Model] --> PHYS[Physical Equations] --> SIM[Simulation] --> PRED[Prediction] --> OPT[Optimization] --> SCEN[Scenarios] --> OPS[Operator] --> ACTION[Controlled Action]

flowchart LR
    MODEL[Digital Model] --> TWIN[Digital Twin] --> OP[Operational Twin] --> INTEL[Intelligent Twin] --> AGENT[Agentic Cyber-Physical System]

The strongest digital-twin architectures are converging with:

live operational state + communications + controls + simulation + AI + security + decision support

Digital-twin work in critical infrastructure is therefore increasingly about dynamic operational models rather than 3D visualization alone. DOE research programs and recent utility-AI initiatives reflect this shift.

flowchart LR
    SPEED[Fast] --> LATENCY[Low Latency] --> DETERMINISM[Determinism] --> ASSURANCE[Operational Assurance]

flowchart LR
    ASSET[Asset] --> PROTOCOL[Protocol] --> SEMANTIC[Semantic Model] --> TWIN[Digital Twin] --> SIM[Simulation] --> AI[AI / Optimization] --> DECISION[Decision]

The objective is not protocol collection.

It is interoperability across heterogeneous infrastructure.

flowchart LR
    C[Confidentiality] --> I[Integrity] --> A[Availability] --> S[Physical Safety] --> ST[System Stability]

IEC 62351 · IEC 62443 · NERC CIP · NIS2 · EU Cyber Resilience Act · MITRE ATT&CK for ICS

AI assurance for critical infrastructure is becoming a first-class engineering concern. NIST's 2026 Trustworthy AI in Critical Infrastructure work explicitly addresses AI systems operating at the intersection of AI, IT, OT, ICS, cybersecurity and physical infrastructure, including deterministic behaviour, explainability, graceful degradation and fail-safe operation.

flowchart LR
    PHYSICAL[Physical Assets] --> EDGE[Edge<br/>RT / AI / OT] --> REGIONAL[Regional Services] --> CLOUD[Cloud<br/>Data / ML / AI] --> OPERATORS[Operators + Engineering] --> PHYSICAL

Docker · Kubernetes · Terraform · Ansible · Linux · AWS · Azure · GCP · NGINX

Containerization · Infrastructure as Code · GitOps · Edge deployments · Distributed services · Observability · Secrets management · Automated testing · Secure supply chains · Reproducible deployments

PostgreSQL · TimescaleDB · InfluxDB · Redis · MongoDB · Neo4j · Cassandra · Kafka · RabbitMQ · NATS

flowchart LR
    EVENTS[Events] --> STREAMS[Streams] --> TIMESERIES[Time Series] --> GRAPH[Asset Graph] --> SEMANTICS[Semantic Layer] --> FEATURES[Real-Time Features] --> MODELS[Models] --> DECISIONS[Decisions]

Time-series databases · Event sourcing · Streaming architectures · Graph databases · Telemetry pipelines · Digital-thread architectures · Historical replay · Event correlation · Asset knowledge graphs · Real-time feature pipelines

flowchart LR
    LOGS[Logs] --> METRICS[Metrics] --> TRACES[Traces] --> EVENTS[Events] --> TELEMETRY[Telemetry] --> MODELS[Model Predictions] --> AGENTS[Agent Actions] --> OPERATORS[Operator Decisions] --> OBS[Unified Observability]

Prometheus · Grafana · OpenTelemetry · Elasticsearch · Loki · distributed tracing · structured logging

The objective:

make system behaviour observable during operation and explainable after the fact.

Complex cyber-physical systems need environments where ideas can be tested before touching physical infrastructure.

flowchart LR
    POWER[Power Simulation] --> NETWORK[Network Simulation] --> CPS[Cyber-Physical Simulation] --> AGENT[Agent Simulation] --> RL[RL Environment] --> HIL[Hardware-in-the-Loop] --> SIL[Software-in-the-Loop] --> SCENARIO[Scenario Generation] --> ADV[Adversarial Testing]

HELICS · OMNeT++ · RTDS · OPAL-RT · MATLAB/Simulink · Python scientific computing

Power-system simulation · Cyber-physical simulation · Network simulation · Agent-based simulation · RL environments · Digital twins · HIL · SIL · contingency analysis

Research and engineering around LiDAR-based perception and autonomous inspection.

LiDAR · Computer vision · Sensor fusion · Point clouds · SLAM · Localization · Object detection · 3D reconstruction · Uncertainty estimation · ROS 2 · Autonomous navigation · Path planning · Robotic inspection · Edge inference

flowchart LR
    SENSOR[LiDAR / Cameras] --> FUSION[Sensor Fusion] --> PERCEPTION[Perception] --> LOCALIZATION[Localization] --> TWIN[3D / Digital Twin] --> PLANNING[Path / Action Planning] --> INSPECTION[Infrastructure Inspection]

Technology interests:

Three.js · WebGL · WebGPU · Unity · Unreal Engine · Blender · Plotly · Dash · React

flowchart LR
    ASSET[Asset] --> STATE[Live State] --> 3D[3D View] --> TELEMETRY[Telemetry] --> SIM[Simulation] --> WHATIF[What-If Analysis] --> OPERATOR[Operator Interface]

Applications:

3D digital twins · Substation visualization · Asset visualization · Network topology · Live telemetry · Simulation playback · Operator dashboards · Spatial interfaces · Engineering visualization · Interactive what-if analysis

One of the largest infrastructure problems is not missing data.

It is missing shared meaning.

flowchart LR
    ASSET[Physical Asset] --> TELEMETRY[Telemetry] --> PROTOCOL[Protocol] --> MODEL[Semantic Model] --> TWIN[Digital Twin] --> SIM[Simulation] --> AI[AI / Optimization] --> DECISION[Decision]

Relevant technologies:

CIM · IEC 61968 · IEC 61970 · IEC 61850 · knowledge graphs · ontologies · RDF · graph databases · semantic APIs

The target is a digital thread capable of surviving vendors, protocols and software generations.

flowchart LR
    HUMAN[Human] --> AGENT[AI / LLM / Agent] --> POLICY[Policy Layer] --> VALIDATE[Validation] --> ASSURE[Runtime Assurance] --> PHYSICAL[Physical System]

_{GridOS — digital-twin and intelligent infrastructure engineering environment}

The repositories are not intended to be isolated software projects.

They form a connected research and engineering surface.

flowchart LR

    EE[Electrical Engineering]

    EE --> PHY[physics-informed]
    EE --> GRID[GridOS]
    EE --> DER[DERIM]
    EE --> ROB[robot-lidar-fusion]

    PHY --> TWIN[Digital Twin]
    GRID --> TWIN
    DER --> TWIN
    ROB --> TWIN

    TWIN --> SIM[Simulation]
    TWIN --> AI[Physics-Informed AI]

    SIM --> AI
    AI --> NB[NeuralBridge]

    NB --> AGENTS[Agentic Systems]
    AGENTS --> ASSURE[Runtime Assurance]
    ASSURE --> CONTROL[Controlled Action]

    CONTROL --> EE

flowchart LR
    PHYSICAL[Physical Domain] --> MODEL[Model] --> SIMULATE[Simulate] --> INTELLIGENCE[Intelligence] --> ASSURE[Assure] --> ACT[Act] --> MEASURE[Measure] --> LEARN[Learn] --> PHYSICAL

The portfolio is deliberately designed to connect to established technical foundations.

flowchart LR

    subgraph OPEN["OPEN ENGINEERING ECOSYSTEM"]
        CIM[CIM]
        IEC[IEC 61850]
        LF[LF Energy]
        GRIDAPPS[GridAPPS-D]
        OPENEMS[OpenEMS]
        PYPSA[PyPSA]
        PP[pandapower]
        ODSS[OpenDSS]
        HELICS[HELICS]
    end

    subgraph GRIMALDI["ENGINEERING SURFACE"]
        PHY[physics-informed]
        GO[GridOS]
        DER[DERIM]
        NB[NeuralBridge]
        ROB[robot-lidar-fusion]
    end

    subgraph SYSTEM["SYSTEM LEVEL"]
        TWIN[Digital Twin]
        AI[Physics-Informed AI]
        AGENTS[Agentic Systems]
        SEC[Cybersecurity]
        OBS[Observability]
    end

    CIM --> PHY
    IEC --> GO
    LF --> DER
    GRIDAPPS --> GO
    OPENEMS --> DER
    PYPSA --> PHY
    PP --> PHY
    ODSS --> PHY
    HELICS --> PHY

    PHY --> TWIN
    GO --> TWIN
    DER --> TWIN
    NB --> AGENTS
    ROB --> TWIN

    TWIN --> AI
    AI --> AGENTS
    AGENTS --> SEC
    SEC --> OBS

Integration rather than reinvention.

That distinction matters in infrastructure.

flowchart LR
    A[Physical Infrastructure]
    B[Electrical Engineering]
    C[Industrial Protocols]
    D[Telemetry]
    E[Semantic Infrastructure]
    F[Digital Twin]
    G[Simulation]
    H[Scientific AI]
    I[Optimization]
    J[Agentic Systems]
    K[Cybersecurity]
    L[Runtime Assurance]
    M[Human / Operator Interface]
    N[Automation / Robotics]

    A --> B --> C --> D --> E --> F
    F --> G
    F --> H
    G --> H --> I --> J --> K --> L
    L --> M
    L --> N
    M --> A
    N --> A

PyPSA · pandapower · OpenDSS · GridAPPS-D · HELICS · Grid2Op · Dynawo · Power Grid Model · PowSyBl

LF Energy · SOGNO · OpenEMS · OpenFMB-related architectures · CIM · IEC 61850 ecosystems

Grid foundation models · Physics-informed ML · Scientific ML · Graph-based grid intelligence · Reinforcement-learning environments · Agentic decision support

Digital twins · Runtime assurance · OT cybersecurity · Functional safety · HIL · SIL · Edge computing

flowchart LR
    ASSET[Physical Asset] --> ENG[Engineering Model] --> DATA[Telemetry / Data] --> SEM[Semantic Model] --> TWIN[Digital Twin] --> PRED[Simulation / Prediction] --> AI[AI / Optimization] --> ASSURE[Assurance / Policy] --> ACTION[Control / Action]

A physical asset progressively becomes richer in digital representation.

That is the digital thread.

Not simply another:

AI application · dashboard · digital-twin visualization · grid simulator · automation framework · robotics repository

The larger direction is the integration of all of them.

flowchart LR
    PHYSICS[Physics] --> ELEC[Electrical System] --> DIGITAL[Digital Representation] --> TWIN[Digital Twin]
    TWIN --> SIM[Simulation]
    TWIN --> DATA[Data]
    SIM --> AI[AI / ML / RL]
    DATA --> AI
    AI --> POLICY[Policy + Assurance] --> ACTION[Controlled Action] --> INFRA[Physical Infrastructure] --> PHYSICS

flowchart LR
    PHYSICS[PHYSICS] --> CORE
    ELEC[ELECTRICAL ENGINEERING] --> CORE
    DATA[DATA] --> CORE
    SOFTWARE[SOFTWARE] --> CORE
    AI[AI] --> CORE
    SECURITY[CYBERSECURITY] --> CORE
    AUTOMATION[AUTOMATION] --> CORE
    HUMAN[HUMAN ENGINEERING] --> CORE

    CORE["INTELLIGENT<br/>CYBER-PHYSICAL<br/>INFRASTRUCTURE"]

The objective is not to remove the engineering discipline beneath the software.

It is to make that discipline computable.

flowchart LR
    PHYSICS[Physics] --> TWIN[Digital Twins] --> AI[AI / ML]
    TWIN --> SIM[Simulation]
    AI --> AGENTS[Agentic Systems]
    SIM --> AGENTS
    AGENTS --> ASSURE[Runtime Assurance]
    ASSURE --> SEC[Cybersecurity]
    SEC --> INFRA[Physical Infrastructure]
    INFRA --> PHYSICS

The resulting class of systems is different:

Software that understands the physical systems it operates around.

This portfolio tracks the evolution from:

flowchart LR
    RULES[Rules] --> MODELS[Physics Models] --> ML[Machine Learning] --> FM[Foundation Models] --> AGENTS[Agentic Systems] --> ASSURED[Assured Cyber-Physical Intelligence]

Current industry and research signals include:

Grid foundation models DOE's GridFM 2.0 initiative is targeting dramatically higher planning and scenario-analysis throughput using AI.

Dynamic grid foundation models GE Vernova's active DynaGridFM project combines generative AI, time-series modelling and graph neural networks for proactive grid decision support.

Agentic AI for power/OT DOE and LLNL's Stormbreaker testbed is explicitly focused on evaluating LLMs and agentic AI in power-system and OT environments.

Open-source AI operations LF Energy's AINETUS integrates AI decision support with Grid2Op, OperatorFabric and SOGNO; OpenGridFM is progressing through LF Energy's project lifecycle.

Digital-substation evolution IEC's 2026 IEC 61850 series release demonstrates that the interoperability layer itself continues to evolve.

Critical-infrastructure AI governance NIST is developing a dedicated Trustworthy AI in Critical Infrastructure profile covering AI/IT/OT/ICS intersections and operational properties such as determinism, explainability and graceful degradation.

flowchart LR
    WEB[GRIMALDI.CA<br/>Strategic / Visual]
    GITHUB[GITHUB<br/>Engineering]
    RESEARCH[Research]
    SOFTWARE[Software]
    SYSTEMS[Systems]
    ENERGY[Energy]
    INTELLIGENCE[AI]
    ROBOTICS[Robotics]
    INFRA[Intelligent Infrastructure]

    WEB --> GITHUB
    GITHUB --> RESEARCH
    GITHUB --> SOFTWARE
    GITHUB --> SYSTEMS
    RESEARCH --> ENERGY
    SOFTWARE --> INTELLIGENCE
    SYSTEMS --> ROBOTICS
    ENERGY --> INFRA
    INTELLIGENCE --> INFRA
    ROBOTICS --> INFRA

English · German · Spanish · Mandarin

The future of infrastructure will not be defined by AI alone.

It will be defined by the integration of:

Physics + Electrical Engineering + Software + Data + AI + Cybersecurity + Automation + Human Engineering

The systems worth building are those that can operate across these domains without losing the properties that make physical infrastructure trustworthy:

determinism · safety · resilience · observability · interoperability · explainability · accountability

VINCENZO GRIMALDI
_{Electrical Engineering · Cyber-Physical Systems · Critical Infrastructure · Digital Twins · Physics-Informed AI · Intelligent Engineering}

grimaldi.ca · Vincenzo.grimaldi.engineering@gmail.com