Earthflow3D is a browser-based 3D digital twin demonstrating real-time operations intelligence for utility-scale solar + storage. This document catalogs all implemented features, maps each to real-world data sources, benchmarks against market leaders, and identifies the development gaps to close for a full pilot on a B&V-constructed site.
The global solar O&M market reached $14.5 billion in 2024 and is projected to grow at 8.4% CAGR to $32.6 billion by 2034 (Precedence Research). In the U.S. alone, with 221 GW of installed solar capacity and 246 GW projected through 2030, the addressable O&M services market will exceed $10 billion by 2034. Yet the market remains dominated by monolithic SaaS platforms (Stem/AlsoEnergy, Power Factors, GreenPowerMonitor) that require 6+ months to implement and offer no AI-driven intelligence, no environmental compliance, and no full-lifecycle coverage.
The B&V Business Case identifies a $135M+ 10-year revenue opportunity by combining B&V’s 100+ GW installed base and utility relationships with Earthflow’s Physics AI platform. B&V’s EPC contracts guarantee SCADA/DAS data handoff at COD — eliminating the single largest barrier to O&M platform adoption.
When we benchmarked Earthflow against the top 5 O&M platforms, we identified five critical feature gaps: day-ahead generation forecasting, BESS dispatch optimization, energy loss waterfall analysis, fleet-wide failure probability scoring, and soiling ROI economics. All five gaps have now been closed in the Earthflow3D demo. What remains is the transition from simulated data to live site data — a 3-6 month API integration and AI training exercise detailed in Chapter 8.
Key thesis: The gap from demo to production is an API integration exercise, not a fundamental research problem. Every metric displayed in Earthflow3D maps to available real-world data sources that are standard at commercial operation date (COD).
Earthflow3D is a browser-based 3D digital twin delivered as a single-file web application hosted on Firebase. It renders a photorealistic globe with real terrain, then overlays a fully interactive utility-scale solar + storage plant where every asset is clickable, inspectable, and monitored in real time.
Unlike traditional 2D SCADA dashboards, Earthflow3D provides spatial context — operators see where a degraded panel block sits relative to the substation, which inverter feeds which bus, and how the BESS units are positioned within the plant boundary. This spatial awareness accelerates fault isolation and maintenance planning.
| Asset Class | Quantity | Specification | Inspector Tabs |
|---|---|---|---|
| Panel Blocks | 16 | ~9.4 MW DC each, mono-Si bifacial, single-axis tracker | Overview, Performance, Health, Alerts, Work Orders, Inspection |
| Inverters | 5 | 30 MW AC each (SMA / TMEIC class) | Overview, Performance, Health, NERC CIP, Alerts, Work Orders |
| Substation | 1 | 150 MVA, 34.5 kV → 230 kV step-up | Overview, Performance, Health, DGA, Alerts, Work Orders |
| BESS Units | 2 | 50 MW / 200 MWh LFP each (100 MW / 400 MWh total) | Overview, Performance, Health, Dispatch, Alerts, Work Orders |
The demo showcases a 150 MW solar + 100 MW / 400 MWh BESS layout, but the Earthflow3D platform is fully configurable to any utility-scale site. The number of panel blocks, inverter count and capacity, substation configuration, BESS sizing, and plant boundary are all parameterized — not hardcoded. Any site layout can be modeled by providing the configuration data below.
Asset interconnections — which panel blocks feed which inverters, which inverters connect to which MV bus, and how the substation ties into the transmission grid — are defined via a site configuration that maps the electrical single-line diagram to 3D positions. This allows operators to trace fault propagation spatially: click a tripped inverter and see exactly which panel blocks are affected.
| Data Required | Source | Purpose |
|---|---|---|
| Electrical single-line diagram | EPC / Owner | Asset interconnections (panel → inverter → MV bus → substation → grid) |
| Site layout / plot plan (CAD/GIS) | EPC civil drawings | 3D asset positioning, plant boundary, access roads, drainage |
| Equipment specifications | OEM datasheets | Asset dimensions, capacity ratings, model-specific parameters |
| GPS coordinates per asset | As-built survey or DAS config | Precise geolocation of each panel block, inverter pad, substation, BESS container |
| Tracker row geometry | EPC / tracker OEM (NEXTracker, Array Technologies) | Row pitch, azimuth, ground coverage ratio (GCR) for panel block positioning |
| BESS container layout | BESS OEM / EPC (Fluence, Tesla, BYD) | Container positions, AC/DC coupling topology, thermal setback zones |
For higher-fidelity digital twins, the site geometry can be augmented with 3D point cloud data from LiDAR scans or photogrammetric drone surveys. Point clouds provide exact as-built geometry that goes beyond what CAD drawings capture: actual panel tilt angles, rack heights, cable tray routing, terrain contours, and vegetation clearance zones.
Drone-captured orthomosaics and thermal imagery — from providers like Skydio and Veerum — can be overlaid directly onto the 3D model, enabling visual inspection and thermal anomaly detection without dispatching personnel to the field. Combined with Earthflow’s satellite-based detection capabilities (Earthflow Detect), this creates a multi-resolution inspection pipeline from orbit to ground level.
| Data Type | Capture Method | Resolution | Use Case |
|---|---|---|---|
| LiDAR point cloud | Aerial LiDAR or terrestrial scanner (Skydio, Veerum) | 1–5 cm | Precise terrain model, rack geometry, vegetation encroachment measurement |
| Photogrammetric mesh | Drone survey (Skydio, Veerum) | 2–10 cm | Photorealistic 3D model, visual condition assessment |
| Thermal orthomosaic | Drone with IR camera (Skydio, Veerum) | 5–15 cm | Hotspot detection, string-level fault identification |
| RGB orthomosaic | Drone survey (Skydio, Veerum) | 1–5 cm | Visual overlay on 3D terrain, soiling assessment, vegetation mapping |
Earthflow3D monitors 4 asset classes with a total of 40+ metrics. Each asset type has a dedicated multi-tab inspector showing real-time values, historical trends, and predictive analytics. Data ingestion frequencies are fully configurable based on what is available from the site’s SCADA, BMS, DAS, and sensor infrastructure — from sub-second streaming for critical alarms, to 15-minute intervals for standard telemetry, to daily or monthly cadences for trending and predictive models. The update rates shown below represent typical configurations; actual frequencies depend on the data source and site-specific availability. The goal is to build asset health models and condition indicators that surface degradation trends, anomalies, and failure risks before they escalate into unplanned downtime or costly repairs — shifting O&M from reactive to predictive.
| Metric | Units | Update Rate | Category |
|---|---|---|---|
| Performance Ratio | % | 15 min | Performance |
| Degradation Rate | %/year | Monthly | Health |
| Soiling Loss | % | 15 min | Performance |
| Erosion Risk Score | 0-100 | Daily | Environmental |
| NDVI (Vegetation Index) | 0-1 | Weekly | Environmental |
| Availability | % | 15 min | Performance |
| Alarm Level | 0-3 | Real-time | Alerts |
| String Currents (24 strings) | A | 15 min | Performance |
| Specific Yield | kWh/kWp | Daily | Performance |
| Capacity Factor | % | 15 min | Performance |
| Curtailment Loss | MWh | 15 min | Revenue |
| Temperature Loss | % | 15 min | Performance |
| Clipping Loss | MWh | 15 min | Revenue |
| Irradiance (GHI/POA) | W/m² | 15 min | Weather |
| 12-Month Degradation Trend | Sparkline | Monthly | Health |
| Failure Probability (P30/P60/P90) | % | Daily | Predictive |
| Metric | Units | Update Rate | Category |
|---|---|---|---|
| AC Power Output | MW | 15 min | Performance |
| Conversion Efficiency | % | 15 min | Performance |
| Internal Temperature | °C | 15 min | Health |
| Status | Online/Fault/Offline | Real-time | Operations |
| Firmware Version | String | Static | NERC CIP |
| NERC CIP Compliance | Status | Audit cycle | Cybersecurity |
| Efficiency Curve (load vs %) | Sparkline | 15 min | Performance |
| DC Input Voltage | V | 15 min | Performance |
| Grid Frequency | Hz | 15 min | Grid |
| Total Harmonic Distortion (THD) | % | 15 min | Power Quality |
| IGBT Junction Temperature | °C | 15 min | Component Health |
| Fan Speed | RPM | 15 min | Cooling |
| Capacitor ESR | mΩ | Daily | Component Health |
| Power Derating | % | 15 min | Performance |
| Fault Count (lifetime) | # | Daily | Reliability |
| IGBT Degradation Trend (12mo) | Sparkline | Monthly | Predictive |
| Remaining Useful Life (RUL) | Years | Monthly | Predictive |
| Failure Probability (P30/P60/P90) | % | Daily | Predictive |
| Metric | Units | Update Rate | Category |
|---|---|---|---|
| Transformer Capacity | MVA | Static | Config |
| High-Side Voltage | kV | 15 min | Performance |
| Grid Connection Status | Status | Real-time | Operations |
| Load Factor | % | 15 min | Performance |
| Revenue Meter Reading | MWh | 15 min | Revenue |
| Oil Temperature | °C | 15 min | Health |
| Winding Hotspot Temperature | °C | 15 min | Health |
| DGA Status | Normal/Warning/Alarm | Continuous | Health |
| Bushing Power Factor | % | Monthly | Health |
| Breaker Operations Count | # | Event | Reliability |
| Cooling Fan Status | On/Off/Fault | Real-time | Cooling |
| DGA Gas Trending (H&sub2;, CH&sub4;, C&sub2;H&sub2;, CO) | ppm, 12mo sparkline | Monthly | Predictive |
| Failure Probability (P30/P60/P90) | % | Daily | Predictive |
| Metric | Units | Update Rate | Category |
|---|---|---|---|
| State of Charge (SoC) | % (TOU arbitrage pattern) | 15 min | Performance |
| Power (charge/discharge) | MW (+/−) | 15 min | Dispatch |
| Cell Temperature | °C | 15 min | Health |
| State of Health (SoH) | % | Daily | Health |
| Cycle Count (lifetime) | # | Daily | Health |
| Round-Trip Efficiency (RTE) | % | Daily | Performance |
| Cell Voltage Spread | mV | 15 min | Health |
| Coolant Flow Rate | L/min | 15 min | Cooling |
| DC Bus Voltage | V | 15 min | Performance |
| Insulation Resistance | MΩ | Daily | Safety |
| Capacity Fade Curve (36mo) | % SoH, sparkline | Monthly | Predictive |
| Failure Probability (P30/P60/P90) | % | Daily | Predictive |
| Augmentation Status & Timeline | Date / % | Monthly | Planning |
| Remaining Useful Life (RUL) | Years | Monthly | Predictive |
Beyond asset monitoring, Earthflow3D includes four advanced operations intelligence modules that transform raw telemetry into actionable revenue and maintenance decisions. Each module is fully configurable to the site’s specific market, equipment, and operational context — the ISO node and pricing zone, forecast provider, loss model parameters, alert thresholds, and maintenance cost assumptions are all adjustable per site. This ensures the intelligence layer adapts to any utility-scale plant, whether it operates in CAISO, ERCOT, PJM, or any other ISO market.
Real-time BESS dispatch optimization driven by ISO locational marginal pricing (LMP).
| Feature | Description |
|---|---|
| LMP Price Curve | 24-hour CAISO SP15 locational marginal price (96 data points at 15-min intervals), rendered as interactive sparkline with peak/trough annotations |
| Optimal Dispatch Schedule | Visual schedule strip showing charge (green), discharge (amber), and idle periods. Algorithm charges during solar trough & low LMP, discharges during evening peak & high LMP |
| Revenue Stacking | Breakdown of daily BESS revenue across 3 streams: energy arbitrage, capacity payments, and frequency regulation. Card-based display with per-stream values |
| Optimization Savings | Side-by-side comparison of optimized dispatch vs naive TOU strategy. Demonstrated savings: +38-64% vs naive |
| Curtailment Risk | Proactive curtailment probability indicator based on generation forecast vs grid absorption capacity |
24-hour ahead generation forecast with confidence bands, weather correlation, and ramp event detection.
| Feature | Description |
|---|---|
| Confidence Band Sparkline | Forecast center line with upper/lower bounds (±8-15%). Actual generation overlay in green when available. SVG band rendering via custom bandSparkline() |
| Weather Cards | GHI (W/m²), cloud cover (%), wind speed (m/s), ambient temperature (°C) — key forecast input drivers |
| Ramp Event Alerts | Detects rapid generation changes (>20% in 30 min) from cloud transients. Badge shows timing and severity |
| Forecast Accuracy | MAPE (Mean Absolute Percentage Error) and bias tracking. Target: <8% MAPE for day-ahead |
| Actual vs Forecast | When live data is available, actual generation is overlaid on forecast for real-time accuracy monitoring |
Energy loss analysis, clipping quantification, soiling economics, and underperformer identification.
| Feature | Description |
|---|---|
| Energy Loss Waterfall | Nameplate capacity → net output through 5 loss stages: temperature, soiling, clipping, degradation, and downtime. Horizontal bar chart with cumulative loss visualization |
| Clipping Loss Tracker | 24-hour clipping profile (DC generation vs AC inverter limit) with daily revenue impact in $/day |
| Soiling ROI Calculator | Computes cleaning payback period (days) based on soiling loss rate, daily MWh output, PPA price, and cleaning cost. Shows break-even date and annual ROI |
| Worst Performers | Table of lowest-performing panel blocks ranked by performance ratio, with root cause classification (soiling, shade, degradation, tracker fault) |
Aggregate health scoring and predictive maintenance intelligence across all asset classes.
| Feature | Description |
|---|---|
| Fleet Health Bar | Always-visible bar in the Active Site panel showing aggregate health scores for Panels, Inverters, BESS, and Substation (0-100 scale, color-coded) |
| Failure Probability Gauges | P30 / P60 / P90 failure probability for every asset. Drives predictive maintenance scheduling — schedule repairs before P60 exceedance |
| Degradation Trending | 12-month degradation curves for panels, inverters (IGBT), BESS (capacity fade), and substation (DGA gases). Sparkline rendering with trend extrapolation |
| Maintenance Recommendations | Auto-generated action items with ROI: "Clean panels — 4.2 day payback", "Schedule IGBT replacement — RUL 2.1 years", "DGA retest — H&sub2; trending up" |
The Site Performance drawer provides a multi-tab view of the entire plant’s operational status, accessible via a persistent toggle in the bottom toolbar.
| Tab | Content | Data Window |
|---|---|---|
| Live | Generation, Grid Export, Availability, Revenue, BESS SoC, Net Dispatch — all as 48-hour sparklines | 192 data points (15-min intervals) |
| Forecast | 24h generation forecast with confidence bands, weather cards, ramp alerts, forecast accuracy metrics (MAPE, bias) | 96 forecast points + actuals overlay |
| Optimize | Energy loss waterfall, clipping sparkline, soiling ROI calculator, worst-performing blocks table | 24h clipping + cumulative losses |
The left-side Active Site panel provides at-a-glance site status:
| Alert | Asset | Severity | Type |
|---|---|---|---|
| Communication Loss | INV-03 | Critical | Connectivity |
| Accelerated Degradation | Block C-02 | Warning | Performance |
| Tracker Stow Fault | Tracker D-04 | Warning | Mechanical |
| NERC CIP Audit Due | Substation | Info | Compliance |
| Cell Voltage Imbalance | BESS-B | Warning | Battery Health |
| Vegetation Encroachment | Block A-04 | Info | Environmental |
| Work Order | Type | Priority | Status | Asset |
|---|---|---|---|---|
| INV-03 Communication Restore | Corrective | Urgent | Assigned | INV-03 |
| Block C-02 IV Curve Test | Predictive | High | Scheduled | Block C-02 |
| Q1 Panel Cleaning (South Zone) | Preventive | Medium | In Progress | Blocks A-01 to A-08 |
| Annual Thermal Inspection | Inspection | Medium | Planned | All Panels |
| BESS-A Coolant System Check | Preventive | High | Scheduled | BESS-A |
The solar O&M monitoring market is dominated by 5 established players. Earthflow3D’s feature set was designed to match or exceed these platforms while introducing capabilities none of them offer.
| Feature | Earthflow3D | Stem / AlsoEnergy 32.5 GW |
Power Factors 25 GW |
Raptor Maps | GreenPowerMonitor 20 GW |
|---|---|---|---|---|---|
| 3D Digital Twin | ✓ | ✗ | ✗ | ✗ | ✗ |
| Generation Forecast | ✓ + confidence bands | ✓ | ✓ | ✗ | ✓ |
| BESS Dispatch Optimization | ✓ LMP-based | ✓ Athena | ✗ | ✗ | ✗ |
| Loss Waterfall | ✓ | ✓ | ✓ | ✓ | ✓ |
| Fleet Failure Probability | ✓ P30/P60/P90 | ✗ | ✗ | ✓ | ✗ |
| Soiling ROI Calculator | ✓ | ✓ | ✓ | ✗ | ✓ |
| DGA Gas Trending | ✓ 4-gas 12mo | ✗ | ✗ | ✗ | ✗ |
| BESS Capacity Fade | ✓ 36mo curve | ✓ | ✗ | ✗ | ✗ |
| Physics-Informed AI | Unique | ✗ | ✗ | ✗ | ✗ |
| Agentic AI (Cirra) | Unique | ✗ | ✗ | ✗ | ✗ |
| Environmental Intelligence | Unique | ✗ | ✗ | ✗ | ✗ |
| NERC CIP Cybersecurity | ✓ with B&V | ✗ | ✗ | ✗ | ✗ |
| Full Lifecycle (Dev → Ops) | Unique | ✗ | ✗ | ✗ | ✗ |
Every metric displayed in Earthflow3D maps to a real-world data source that is available at COD (commercial operation date). The table below shows the complete mapping.
| Feature | Real Data Source | Access Method | Owner | Est. Cost | Earthflow Physics AI |
|---|---|---|---|---|---|
| LMP Pricing | CAISO OASIS / ERCOT / PJM | Configurable — Public REST API | Public | Public API | ✓ Queue Pressure Score (QPS) already in Earthflow |
| Generation Forecast | Solcast, SolarAnywhere, or Earthflow Physics AI | Configurable — Commercial API or Earthflow irradiance-to-yield model | 3rd Party / Earthflow | $0–$8K/yr/site | ✓ Irradiance → yield model, confidence bands |
| Weather Data | Site instrumentation, sensor packages, Tomorrow.io, OpenWeatherMap, NREL | Configurable — Site sensors, API (already in Cirra) | Owner / 3rd Party | Site sensors or $0–$5K/yr API | ✓ 3-source weather engine in Cirra |
| Panel SCADA | String monitors (SMA, SolarEdge, Tigo) | Configurable — DAS API or Modbus | Owner via DAS | Included at COD | — |
| Inverter Telemetry | OEM SCADA (SMA, TMEIC, Power Electronics) | Configurable — Modbus registers | Owner at COD | Included at COD | — |
| BESS Cell Data | BMS (Fluence, Tesla, BYD, CATL) | Configurable — BMS cloud API | Owner via OEM | Included at COD | — |
| Substation DGA | Online DGA (Qualitrol, Vaisala) or lab | Configurable — API or PDF parse | Owner | $5-15K/yr | — |
| Soiling Data | Soiling stations (Kipp & Zonen) or satellite | Configurable — IoT sensor or API | Owner | $2-5K/yr | — |
| Clipping Loss | Calculated: DC power − AC power | Configurable — Math on existing SCADA | Derived | Derived from SCADA | — |
| Work Orders | CMMS (Maximo, SAP PM, eMaint, Fiix) | Configurable — REST API or CSV | O&M Provider | Existing license | — |
| Revenue Meter | Fiscal meter + DAS | Configurable — SCADA export | Owner | Included at COD | — |
| Curtailment Signals | ISO curtailment signals + SCADA | Configurable — ISO API + SCADA | Public + Owner | Public API | — |
| NDVI / Vegetation | Sentinel-2, Planet Labs, drone imagery | Configurable — Satellite API or upload | Public / 3rd Party | Public or $0–$3K/yr | ✓ NDVI, vegetation classification, RUSLE erosion, species ID |
| Fire Risk | USGS, NIFC, NASA FIRMS | Configurable — REST API (already in Cirra) | Public | Public API | ✓ Physics-based fire risk model |
| Grid Interconnection Queue | PJM, MISO, CAISO, SPP, ERCOT | Configurable — REST API (already in Earthflow) | Public / ISOs | Public API | ✓ QPS scoring, 50K+ projects tracked |
| Satellite Detection | Mapbox, Sentinel-2, Google Earth Engine | Configurable — Satellite API (already in Earthflow) | Public / 3rd Party | Public or $0–$2K/yr | ✓ Earthflow Detect: GeoAI, U-Net, YOLO, Gemini |
| Erosion Risk | USLE/RUSLE soil & slope data, NRCS | Configurable — REST API (already in Cirra) | Public | Public API | ✓ RUSLE-based erosion model |
| Site Assessment | Multi-source: geospatial, environmental, grid, weather | Configurable — Aggregated via Cirra agent | Earthflow | Included | ✓ 25-tool agentic AI assessment |
| Construction Forecast | Weather + activity models (6 activities) | Configurable — Physics AI model | Earthflow | Included | ✓ 6-activity weather impact model |
| Development Cost | NREL ATB, regional multipliers, site conditions | Configurable — Physics AI model | Earthflow | Included | ✓ Parametric cost estimator |
| PDF Report Generation | All Earthflow data sources | Configurable — Automated via Cirra agent | Earthflow | Included | ✓ 4 report types, branded PDFs |
Several data integrations are already operational in the broader Earthflow platform (used by Cirra AI agent):
The Earthflow3D demo demonstrates all critical O&M intelligence features with simulated data. Transitioning to a real-site pilot requires four development phases focused on API integration, real-time pipelines, predictive models, and dispatch optimization. The estimated total pilot investment is ~$350K across all phases — approximately 0.3% of the $135M 10-year revenue opportunity identified in the B&V Business Case. Target: Q2 2026 pilot on a B&V-constructed utility-scale solar + storage plant.
| Task | Demo Status | Pilot Requirement | Priority |
|---|---|---|---|
| SCADA Adapter | Simulated | Connect to DAS provider API (AlsoEnergy / SMA / meteocontrol) | P0 |
| Weather API | Built | Already operational in Cirra (Tomorrow.io + OpenWeatherMap + NREL) | Done |
| LMP Pricing | Simulated | CAISO OASIS REST API (free, public, well-documented) | P0 |
| Generation Forecast | Simulated | Solcast API integration | P1 |
| Task | Demo Status | Pilot Requirement | Priority |
|---|---|---|---|
| Data Ingestion Service | Simulated | 15-min interval polling from SCADA → Firestore | P0 |
| Historical Backfill | Simulated | 12-month data migration for degradation trending | P0 |
| Alert Engine | Simulated | Real alert thresholds from SCADA values (not demo alerts) | P1 |
| CMMS Integration | Simulated | Bi-directional sync with Maximo / SAP PM | P1 |
| Task | Demo Status | Pilot Requirement | Priority |
|---|---|---|---|
| Panel Degradation Models | Simulated | Train on actual IV curves + performance data | P1 |
| DGA Analysis | Simulated | Duval triangle classification from lab reports or online DGA | P1 |
| BESS Capacity Fade | Simulated | BMS data → SoH trending with augmentation timeline | P1 |
| Failure Probability (AI) | Simulated | AI models trained on maintenance history + sensor data | P2 |
| Task | Demo Status | Pilot Requirement | Priority |
|---|---|---|---|
| Real LMP Integration | Simulated | Live CAISO / ERCOT / PJM pricing feeds | P0 |
| BESS Dispatch Engine | Simulated | Optimal charge/discharge scheduling from price signals | P0 |
| Revenue Stacking | Simulated | Arbitrage + capacity + ancillary services revenue tracking | P1 |
| Curtailment Avoidance | Simulated | Grid signal integration for proactive dispatch | P2 |
The ideal pilot site should be a recently-constructed or Year 1 operating B&V solar + storage project with modern DAS and BMS infrastructure.
| Requirement | Ideal Specification | Minimum Acceptable |
|---|---|---|
| Solar Capacity | 100-200 MW DC | 50 MW DC |
| BESS Capacity | 50-100 MW / 200-400 MWh | 25 MW / 100 MWh |
| SCADA / DAS | AlsoEnergy, meteocontrol, or SMA Data Manager with REST API | Any DAS with API or Modbus export |
| BMS | Fluence, Tesla Megapack, or BYD with cloud API | Any BMS with data export |
| Substation Monitoring | Online DGA monitor (Qualitrol, Vaisala) | Quarterly lab DGA reports |
| Meteorological Station | On-site pyranometer + anemometer + thermometer + soiling station | Pyranometer + temp sensor |
| CMMS | Maximo, SAP PM, eMaint, or Fiix with REST API | Any CMMS with CSV export |
| Network | Fiber backhaul from site to cloud | Cellular (4G/5G) backhaul |
| ISO Market | CAISO (SP15, NP15) or ERCOT | Any US ISO with public LMP data |
| Data Access | EPC handoff of all SCADA credentials + BMS access at COD | Owner-provided read-only API keys |
| Milestone | Timeline | Deliverable |
|---|---|---|
| Pilot site selection | Q2 2026 | Signed data access agreement |
| SCADA + LMP integration | Week 1–4 | Live telemetry flowing to Earthflow3D |
| Historical backfill | Week 4–8 | 12-month trending and degradation curves |
| Alert engine + CMMS | Week 8–12 | Real alerts and bi-directional work orders |
| Predictive AI models | Week 10–18 | Failure probability, DGA analysis, capacity fade |
| Dispatch optimization | Week 14–20 | Live LMP dispatch, revenue stacking, curtailment avoidance |
| Full pilot operational | Week 20 | Complete O&M digital twin with real data |
Earthflow3D demonstrates a market-competitive suite of O&M intelligence features today. The demo includes capabilities that match or exceed every major competitor in the solar monitoring space, while introducing three features no competitor offers: Physics-Informed AI, Agentic AI (Cirra), and full development-to-operations lifecycle coverage.
| # | Action | Owner | Target |
|---|---|---|---|
| 1 | Select pilot site from B&V portfolio (COD or Year 1 asset, solar + BESS) | B&V / Orbyfy | Q2 2026 |
| 2 | Execute data access agreement — SCADA, BMS, met station, CMMS | B&V Legal | Q2 2026 |
| 3 | Begin Phase 1 API integration (SCADA adapter + LMP pricing) | Orbyfy Engineering | Q2-Q3 2026 |
| 4 | Deploy real-time pipeline and historical backfill (Phase 2) | Orbyfy Engineering | Q3 2026 |
| 5 | Train predictive models on site data (Phase 3) | Orbyfy AI Team | Q3-Q4 2026 |
| 6 | Launch dispatch optimization with live LMP (Phase 4) | Orbyfy Engineering | Q4 2026 |
| 7 | Full pilot review and commercial go-to-market decision | B&V / Orbyfy | Q4 2026 |