Solar Fleet Intelligence

SUNSPOT

Every Photon Counts. Every Anomaly Costs.

Real-time solar panel fleet monitoring with GBDT-powered anomaly detection. Predict degradation, classify faults, optimize yields. From 100 to 100,000+ panels — catch issues in 24 hours, not 90 days.

12K+
Panels/Second
8ms
P99 Latency
3-8%
Yield Recovery
$12K+
Saved/MW/Year
Go 1.22+
Flutter Dashboard
Lattice/MP
TimescaleDB
NFSv4

The Hidden Cost of Blind Operations

Solar operators lose 3-8% annual yield to undetected issues.

Quarterly Inspections

  • Soiling goes undetected for months
  • Micro-cracks spread before discovery
  • Inverter drift wastes energy silently
  • Hot spots cause permanent damage
  • $12,700+ lost per MW per year

SUNSPOT Monitoring

  • 24-hour anomaly detection
  • GBDT expected vs actual yield
  • C4.5 fault classification
  • ROI-ranked maintenance queue
  • Mobile dashboard for field crews

Platform Screenshots

Real-time fleet visibility across all your installations

SUNSPOT Command Center
Command Center Dashboard
ML Analytics Studio
ML Analytics Studio
Executive ROI Dashboard
Executive ROI Dashboard
Global Fleet Operations
Global Fleet Operations

System Architecture

From panel telemetry to actionable insights

Ingest

Panel Telemetry

Store

TimescaleDB

Analyze

Lattice/MP GBDT

Detect

Anomaly Scores

Alert

Flutter Dashboard

SOLAR FLEET Inverter Gateway 15-min intervals Weather Station Irradiance + Temp GNSS Module Panel locations SUNSPOT ENGINE (GO) Anomaly Detection GBDT expected vs actual 12,000+ panels/second Fault Classification C4.5 Decision Trees Soiling, cracks, inverter Fleet Clustering K-means grouping Performance profiles OUTPUTS Flutter App Mobile Dashboard Maintenance ROI-ranked queue Alerts Webhook + Email Lattice/MP Model training PostgreSQL TimescaleDB Time-series data NFSv4 Mount Telemetry files

Detection Capabilities

AI-powered monitoring that never sleeps

Anomaly Detection

GBDT-powered expected vs actual yield comparison. Detect underperformance within 24 hours instead of quarterly inspections.

Fault Classification

C4.5 decision trees diagnose soiling, micro-cracks, inverter drift, and hot spots. Know what's wrong, not just that something is.

Fleet Clustering

K-means groups panels by performance profile. Identify systemic issues affecting entire strings or zones.

Drift Monitoring

Detect distribution shifts in irradiance→yield curves. Catch gradual degradation before it compounds.

Maintenance Queue

Priority-ranked work orders with ROI estimates. Fix what matters most first. Maximize return per service call.

Mobile Dashboard

Flutter app for field technicians. Real-time alerts, panel maps, work orders. Everything needed on-site.

Technical Specifications

Built for scale and speed

12K+
Panels/Second
8ms
P99 Latency
2.1s
10K Batch Scan
~120MB
Memory
~2KB
Per Panel/Day
15min
Telemetry Interval
REST
API Protocol
K8s
Deployment

Deployment Scenarios

From rooftop installations to utility-scale farms

Utility-Scale Farms

Monitor 100,000+ panels across multiple sites. Fleet-wide anomaly detection, centralized maintenance scheduling, executive dashboards.

Commercial Rooftops

C&I installations with complex string configurations. Per-building analytics, lease compliance reporting, tenant billing integration.

Residential Portfolios

Manage thousands of residential systems. Homeowner-facing dashboards, automated warranty claims, installer dispatch.

O&M Providers

Multi-client fleet management. SLA tracking, performance guarantees, predictive maintenance, client reporting.

Stop Losing Yield to Invisible Problems

Deploy SUNSPOT and recover 3-8% annual production.

Lattice/MP integration
Custom model training
Flutter mobile app
On-prem or cloud