Cloud Intelligence Engine — Cortex
AI Solar Generation Forecasting & Asset Health Analytics for Renewable Plants
Cortex provides cloud analytics and solar generation forecasting for utility-scale and commercial solar power plants. It combines high-resolution telemetry with local weather models to calculate sensorless soiling baselines, track degradation trends, and detect inverter faults before generation is lost.
Operational scope: Forecasting models, soiling baseline calculations, and anomaly alerts are configured around plant-specific Single Line Diagrams (SLD), inverter telemetry, and site parameters. For regulatory scheduling frameworks, refer to the Central Electricity Regulatory Commission (CERC).
The Operational Challenge
Why Generic Regional Weather Models Fall Short on Indian Solar Plants
Standard renewable monitoring dashboards often apply generic regional weather forecasts over broad 15–25 km geographic grids. In India’s diverse solar belts—from high-dust arid environments in Rajasthan and Gujarat to monsoon transition corridors in Maharashtra and Tamil Nadu—localized plant behavior requires models calibrated to actual site conditions.
SLDC Schedule Deviations
Generic weather models cover broad geographic areas and frequently miss local morning ramp rates or rapid afternoon convective cloud cover. When actual plant generation diverges from declared day-ahead schedules, generators face heavy Deviation Settlement Mechanism (DSM) penalties under state grid codes.
Mistimed Cleaning Schedules
Dust accumulation in semi-arid zones causes performance drops that are often mistaken for permanent module degradation. Without clear separation between soiling and irreversible panel aging, cleaning teams get dispatched either too early (wasting water and labor) or too late (losing peak generation).
Late-Stage Inverter Trips
Threshold-based SCADA alarms typically notify operators only after an inverter has already shut down or tripped offline. Electrical warning signs like thermal derating, string voltage divergence, and cooling fan degradation build up days before the actual failure occurs.
Core Capabilities
Three Practical Analytical Engines for Solar & Hybrid Assets
Cortex operates on telemetry acquired by the Synapse Edge Controller, translating electrical measurements into actionable day-ahead schedules, loss diagnostics, and targeted maintenance recommendations.
Module 01 — Scheduling
Day-Ahead SLDC Solar Generation Forecasting
Physics-informed generation forecasting calibrated against plant-specific single line diagrams, historical generation trends, and site weather parameters.
- 15-Minute Block Granularity: Formatted directly for State SLDC and CERC scheduling requirements.
- P10 / P50 / P90 Probability Bounds: Configurable risk curves helping asset managers select optimal declaration targets.
- DSM Risk Reduction: Helps operators adjust declared capacity during expected curtailment and cloudy periods.
- Site-Specific Calibration: Accounts for local tracker behavior, module tilt, and site terrain characteristics.
Module 02 — Loss Analytics
Research-Driven Soiling Analytics
Calculates dust and soiling impact using empirical photovoltaic performance models and inverter electrical baselines, separating temporary loss from long-term degradation.
- Soiling vs. Aging Separation: Isolates cleanable dust deposition from irreversible annual panel degradation trends.
- Sensorless Methodology: Eliminates the need for capital-intensive dedicated optical soiling stations or reference cells.
- Economic Cleaning Schedules: Identifies when the financial value of recovered energy exceeds the cost of module washing.
- Warranty Verification: Provides clean degradation baselines to support manufacturer warranty reviews.
Module 03 — Health Analytics
Inverter Health & Anomaly Detection
Pattern-based monitoring across electrical parameters to flag abnormal operational behavior before component failures lead to plant downtime.
- Early Thermal Derating Warnings: Detects abnormal heat buildup and cooling fan anomalies before inverters curtail power.
- String-Level Discrepancy Alerts: Identifies underperforming strings, loose connections, or blown string fuses.
- Clipping & Limit Tracking: Distinguishes expected DC/AC oversizing clipping from equipment power limitation faults.
- Contextual Work Orders: Generates prioritized issue summaries for site O&M teams with estimated power impact.
System Roles & Boundaries
Technical Specifications & Operational Integration Scope
Cortex operates as a cloud analytics and intelligence layer. It complements field controllers and SCADA systems without interfering with physical protection or fast local control loops.
| Operating Layer | Cortex Functional Role | System Responsibility & Execution Boundary |
|---|---|---|
| Telemetry Inputs | Ingests Modbus/IEC measurements from Synapse Edge, ambient site temperatures, and historical generation logs. | Edge data acquisition, local polling, and buffering remain managed by Synapse hardware. |
| Analytics Kernel | Cloud-hosted empirical models processing telemetry into 15-minute scheduling blocks and diagnostic metrics. | Non-intrusive cloud execution; no computational load on plant RTUs. |
| Forecasting Horizon | Day-Ahead (96-block SLDC curve), Intra-Day adjustments, and rolling trend estimates. | Generates declaration recommendations; final SLDC schedule submission remains operator-approved. |
| Control Boundary | Outputs advisory notifications, soiling triggers, and generation curves to Clarity UI. | Sub-second grid control (Volt-VAR, zero-export) is executed by Solar PPC; protection remains with switchgear. |
| Multi-Asset Scope | Utility-scale solar, commercial rooftop PV, wind farms, and hybrid asset performance tracking. | For battery storage dispatch algorithms, refer to the BESS EMS Solution. |
Onboarding Path
Data Requirements & Model Setup Process
To configure and calibrate Cortex analytics for a solar or hybrid site, the onboarding process follows four straightforward steps:
1. Plant Baseline Inputs
Review single line diagrams (SLD), module and inverter datasheets, site coordinates, tilt and azimuth layout, and available historical generation records.
2. Field Data Pathway
Connect field equipment via the on-site Synapse edge controller or configure secure API ingestion from existing compatible dataloggers and SCADA gateways.
3. Model Setup & Calibration
Calibrate empirical performance models, soiling baseline calculations, and scheduling curves against historical plant behavior and local weather characteristics.
4. Dashboard & Report Setup
Activate automated daily SLDC schedule exports, economic cleaning threshold notifications, and equipment health views in the Clarity Control Centre.
Technical Governance & Standards Alignment: This product architecture is reviewed by the Enercog Engineering Governance & Systems Architecture Team (Last Reviewed: August 2026). Analytical workflows align with Indian grid codes including the CERC Deviation Settlement Mechanism (DSM), CEA Technical Standards for Grid Connectivity, and IEEE 1547. Implementation scope is confirmed per project SLD and FAT/SAT acceptance criteria.
Frequently Asked Questions
Frequently Asked Questions on Solar Generation Forecasting India & Cortex AI
How does solar generation forecasting in India help reduce DSM penalties?
Under CERC and State Electricity Regulatory Commission (SERC) Deviation Settlement Mechanism (DSM) regulations, solar generators face financial deviation charges when actual plant generation diverges from the declared day-ahead 15-minute schedule. Cortex provides plant-calibrated forecasting curves with P10/P50/P90 probability bounds, helping operators declare more reliable schedules and minimize deviation charges.
How does sensorless soiling analytics calculate dust loss without specialized sensors?
Cortex utilizes research-driven empirical photovoltaic models that cross-reference inverter DC/AC performance ratios with historical generation baselines and meteorological reference data. Because soiling losses fluctuate with weather and recover immediately after cleaning or rain, while panel degradation follows a continuous long-term trend, the software mathematically isolates the two factors without requiring dedicated optical soiling sensors.
Does Cortex require dedicated hardware at the solar plant?
No dedicated on-site servers or specialized sensors are required. Cortex runs in the cloud, processing telemetry streamed from on-site Synapse Edge Controllers or integrated via secure APIs from existing compatible plant SCADA systems and dataloggers.
What inverter operational issues can Cortex help detect early?
Cortex monitors electrical parameters to detect early signs of abnormal inverter thermal derating, string voltage imbalances, blown string fuses, and unexpected power curtailment, allowing maintenance teams to schedule targeted inspections before faults result in complete inverter trips.
What data is required from our solar plant to set up Cortex?
Initial setup requires the plant Single Line Diagram (SLD), inverter and PV module datasheets, site GPS coordinates, module tilt and azimuth orientation, and 6 to 12 months of historical generation logs where available for baseline calibration.
How does Cortex integrate with the wider Enercog platform?
Cortex acts as the cloud analytics engine of the Enercog platform. It ingests field measurements from the Synapse Edge Controller and delivers forecasting curves, soiling notifications, and asset health alerts directly to the Clarity Control Centre and Solar SCADA / RMS interfaces.
Next Step
Turn Plant Telemetry into Actionable Operating Intelligence
Discuss your plant architecture, historical generation data, and SLDC forecasting requirements with our engineering team to review analytics fit.