Offline-first operations
Local rules keep running when internet, VPN or cloud services are unavailable.
Edge AI product line
The charging site stays intelligent, even when the cloud is gone.
Edge AI moves local CPMS, the OCPP Broker, energy logic and AI diagnostics onto an edge device at the charging site. Critical decisions run locally, sessions stay protected and the cloud becomes a synchronization layer rather than a prerequisite for operations.
EDGE AI CPMS · OCPP · AI
Edge CPMS & OCPP Broker
The broker stabilizes WebSocket sessions, caches state and syncs cleanly when the cloud returns.
Local AI diagnostics
AI detects patterns in OCPP messages, MeterValues, charging power, grid draw and device status.
Architecture
Decide on site, monitor centrally.
The edge controller connects meters, charge points, local generation and optional loads such as heat pumps or batteries. Local decisions avoid cloud latency, while telemetry, alerts and reporting continue to synchronize with central systems.
Meter
Modbus · SunSpec · EEBUS
Modbus · SunSpec · EEBUS
PV / Battery
Surplus · Reserve
Surplus · Reserve
Edge AI Device EMS · CPMS · OCPP Broker · AI
Charge points
OCPP 1.6 / 2.0.1
OCPP 1.6 / 2.0.1
Cloud / Fleet
Sync · Alerts · Reporting
Sync · Alerts · Reporting
EV Charging Use Cases
Eight product modules for a charging site that keeps thinking locally.
Edge AI is not just one dashboard. It is local operating logic for charging infrastructure. Modules can be introduced individually or run together as an edge platform: from resilient fallback to OCPP diagnostics and predictive maintenance.
01
Local fault diagnostics
OCPP, meter data and charging curves explain each other.
Many charging faults are not isolated error codes. A vehicle reduces power, a charger sends delayed status messages, the meter sees a phase close to the site limit and the backend only sees a stopped session. Edge AI correlates these signals locally and turns them into an understandable incident.
- Root-cause hints for support and operations
- Local timeline from OCPP, MeterValues, energy and charger status
- Recommendations such as reducing profiles, checking firmware or isolating a charger
IncidentSession stops after load peak
OCPPStatusNotification delayed by 42 s
MeterL2 reaches 92 % of site limit
AILikely load-management conflict
02
Autonomous site fallback
When connectivity disappears, the site keeps operating.
The edge controller keeps local authorizations, limits, priorities and running OCPP sessions available. Messages are queued, decisions are logged and synchronization resumes later. “Cloud unavailable” no longer automatically means “charging unavailable”.
Cloud online VPN unstable Edge takes over Sync later
03
Predictive maintenance
Recurring field patterns become visible before outages happen.
Edge AI observes charging behavior and device health over time. Reconnect clusters, decreasing power, thermal derating, weak LTE links, phase anomalies or charger-specific firmware patterns become actionable maintenance hints.
Risk rising
04
AI-assisted smart charging
PV, battery, grid limit and fleet needs are balanced locally.
Classic load management only protects the grid connection. Edge AI can prioritize vehicles, respect departure times, use PV surplus, keep battery reserves, include heat pumps and still apply OCPP-compliant charging profiles.
PV42 kW
FleetPriority A
Grid120 kW limit
Battery30 % reserve
05
OCPP anomaly detection
Protocol quirks become visible and manageable.
In real charging parks, chargers do not always behave exactly like specifications promise. Edge AI detects unusual sequences, invalid MeterValues, status changes, command timeouts or vendor-specific deviations and suggests broker rules or configuration changes.
BootNotification StatusNotification MeterValues? RemoteStart Broker Rule 06
Support copilot for operators
Logs become a readable incident file.
The copilot automatically summarizes affected charge points, timeframe, sessions, probable cause, technical evidence and next action. That saves support time and makes field issues understandable beyond OCPP specialists.
Incident Summary EVSE 03 · 2 sessions · 18:14-18:47
Likely local load limit plus delayed charger status notification.
07
Depot and fleet optimization
The right vehicle gets energy at the right time.
Depots need availability, not just maximum charging power. Edge AI can combine route plans, minimum SoC, dwell time, vehicle group, electricity price and site limits locally. The cloud provides planning, the edge executes robust charging decisions.
Bus 12 · 06:20 · 78 % Van 04 · 07:10 · 64 % Service 2 · 09:30 · 41 % Pool · flexible · 38 %
08
Security at the edge
Unusual commands and configuration changes are caught locally.
Charging infrastructure is part of critical operations. Edge AI can detect suspicious remote commands, certificate issues, new firmware versions, unexpected charger configuration changes or unusual OCPP sequences and escalate them locally.
Certificate valid New firmware detected RemoteCommand outside maintenance window Audit trail secured locally
Why edge?
Resilience for sites that cannot wait.
Edge AI moves critical logic to the site: local rules, OCPP session protection, fallback routing, load management and diagnostics stay active when the external connection is degraded.
- Stable charging during poor field connectivity
- Priorities for vehicles, heat pumps, batteries and PV surplus
- Auditable local decisions with later cloud synchronization
Hardware & energy
Edge Controller
Local integration for meters, PV, batteries, charge points and loads.
View controllerOperations & OCPP
Edge CPMS
Operator functions, OCPP Broker, Digital Twin and synchronization at site level.
View Edge CPMSSupport & AI
AI Diagnostics
Local diagnostics for OCPP, charging power, meter data and recurring field issues.
View diagnosticsNext step