Question 1

What products does HelioExpect offer?

Accepted Answer

HelioExpect is a renewable energy intelligence layer with three products: Short-term Forecasting (with Trading Intelligence built in), Long-term Forecasting, and Predictive Monitoring. Short-term Forecasting delivers 15-min to 14-day predictions with 98% accuracy. Long-term Forecasting covers 12-month to 25-year yield with P50/P75/P90 confidence bands for project finance. Predictive Monitoring prevents 30–50% downtime with multi-level monitoring from plant to string. Purchase products individually or together.

Question 2

Can I purchase Forecasting and Monitoring separately?

Accepted Answer

Yes, absolutely! Both products are standalone solutions that can be purchased individually based on your specific needs. However, they work exceptionally well together - using both provides complete solar intelligence by combining future predictions with real-time performance monitoring.

Question 3

What makes HelioExpect different from other solar monitoring solutions?

Accepted Answer

HelioExpect combines advanced AI/ML technology with practical operational features. Our Short-term Forecasting uses ML-corrected predictions with GraphCast integration for 98% accuracy and ships Trading Intelligence as a built-in sub-feature — upload your schedule and get back penalty + revenue calculations. Predictive Monitoring uses AI to predict failures days in advance, preventing 30–50% downtime. We offer 74+ visualization types, multi-level monitoring from plant to string, and comprehensive ESG reporting.

Question 4

What types of solar installations do you support?

Accepted Answer

We support utility-scale solar plants, commercial installations, and distributed solar portfolios. Our platform scales from single sites to 100+ site portfolios. We work with IPPs, solar asset owners, O&M providers, utilities, and grid operators managing solar assets of any size.

Question 5

How long does implementation take?

Accepted Answer

Implementation typically takes 2-4 weeks. For Forecasting, we need access to your site data and weather information. For Predictive Monitoring, optimal results require 6-12 months of historical data, though basic monitoring can start with 1 month of data. We provide full onboarding support and training.

Question 6

How accurate are your solar forecasts?

Accepted Answer

Our ML-corrected forecasting system achieves 98% accuracy. We combine multiple data sources including GraphCast AI weather models, cloud motion vectors from satellite imagery, and historical production data. The system continuously learns from actual generation data to improve predictions over time.

Question 7

What is the forecast horizon?

Accepted Answer

We provide forecasts from 15-minute intervals up to 14 days ahead. This allows for both short-term operational decisions (intraday trading, grid scheduling) and medium-term planning (maintenance scheduling, deviation settlement preparation).

Question 8

How often are forecasts updated?

Accepted Answer

Forecasts are updated every 15 minutes with the latest weather data and satellite imagery. You receive continuous updates throughout the day, ensuring you always have the most current predictions based on changing weather conditions.

Question 9

What delivery methods are available for forecasts?

Accepted Answer

Forecasts can be delivered via: 1) Interactive Dashboard with real-time updates and visualizations, 2) Automated Email reports (scheduled hourly, daily, or weekly) in CSV, Excel, or PDF formats, and 3) FTP/SFTP integration for direct server uploads in CSV, JSON, or XML formats with automated scheduling.

Question 10

How does GraphCast integration work?

Accepted Answer

We integrate Google's GraphCast AI weather model with satellite-based cloud motion vectors to predict irradiance changes. This combination provides high spatial and temporal resolution for tracking cloud movement and predicting its impact on solar generation with unprecedented accuracy.

Question 11

Can I compare forecasts with actual production?

Accepted Answer

Yes! Our platform provides real-time comparison of forecasted vs. actual generation. We automatically calculate accuracy metrics, track deviations, and monitor model performance. This comparison helps validate forecast reliability and continuously improves the ML models.

Question 12

How does forecasting help with deviation settlement?

Accepted Answer

Accurate 14-day forecasts help you better schedule energy delivery to the grid, reducing deviation penalties. The system alerts you to potential deviations and provides recommendations for optimal scheduling. Many customers report 30% reduction in deviation penalties.

Question 13

What data do you need from my site for forecasting?

Accepted Answer

We need: site location coordinates, installed capacity, panel specifications (tilt, azimuth, type), historical production data (optional but recommended for ML training), and access to your inverter data. We handle all weather data collection from our integrated sources.

Question 14

What monitoring levels does the platform support?

Accepted Answer

We provide four levels of monitoring: Plant-level (aggregate performance, total generation, plant-wide PR), Inverter-level (DC/AC power, efficiency, temperature, MPPT performance), SCB-level (string current/voltage, health scoring, connection tracking), and String-level (individual string current, voltage, power output, performance diagnostics). This multi-level approach helps pinpoint issues from high-level overviews down to specific components.

Question 15

How does predictive maintenance work?

Accepted Answer

Our AI analyzes historical equipment data to identify patterns that precede failures. Site-specific ML models learn from inverter, SCB, and string performance data, environmental conditions, and operational parameters. The system provides 0-100% health scores, failure probability estimates, days-to-failure predictions, and specific maintenance recommendations - preventing 30-50% of unplanned downtime.

Question 16

How much historical data is needed for accurate predictions?

Accepted Answer

Minimum: 1 month of continuous data for basic predictions. Optimal: 6-12 months to capture seasonal patterns, weather variations, and equipment behavior changes. Required data includes power output, temperature, voltage, current, and environmental conditions. Data quality is crucial - gaps can reduce accuracy. For new sites, we use transfer learning from similar sites while building site-specific models.

Question 17

What are health scores and how should I interpret them?

Accepted Answer

Health scores range from 0-100% for each component. 90-100 (Excellent): Continue normal monitoring. 70-89 (Good): Normal operations. 50-69 (Fair): Schedule inspection within 30 days. 30-49 (Poor): Plan maintenance within 2 weeks, increase monitoring. 0-29 (Critical): Immediate inspection required, consider emergency maintenance. Each score includes trend analysis and specific recommendations.

Question 18

How many visualization types are available?

Accepted Answer

The platform offers 74+ visualization types including real-time charts, performance heatmaps, trend analysis graphs, comparison charts, geographic maps, string-level diagnostics, efficiency curves, and customizable dashboards. All visualizations update in real-time using WebSocket connections.

Question 19

What alert severity levels do you support?

Accepted Answer

We provide 4 severity levels: Critical (immediate action required, production-impacting), High (requires attention within hours, significant degradation), Medium (monitor closely, plan maintenance within days), and Low (informational, minor issues requiring eventual attention). You can configure custom alert rules with your own thresholds and conditions.

Question 20

Can I create custom dashboards?

Accepted Answer

Yes! Our drag-and-drop dashboard builder lets you create personalized views with any combination of our 74+ widget types. Configure granularity for each widget (15min, hourly, daily, etc.), arrange layouts, save multiple dashboard configurations, and share them with your team. Perfect for role-specific views and custom reporting needs.

Question 21

How does the platform integrate with my existing systems?

Accepted Answer

We offer comprehensive REST APIs for programmatic access, webhooks for real-time notifications to external systems (Slack, Teams, OpsGenie, PagerDuty), FTP/SFTP for automated data delivery, CSV/Excel/JSON exports for CMMS integration, and API access with authentication, rate limiting, and complete documentation.

Question 22

How often are predictions updated?

Accepted Answer

Automatic updates occur every 12 hours (6 AM and 6 PM) using the latest 12 hours of monitoring data. Manual triggers are available for critical situations or after major maintenance events. Real-time alerts notify you immediately of critical health score drops or anomaly detection. The system uses incremental learning to continuously improve predictions.

Question 23

How does the system handle multiple sites?

Accepted Answer

Each site gets dedicated ML models trained on site-specific data and equipment characteristics. Training jobs are distributed across multiple workers using parallel processing, allowing 10-20 sites to train simultaneously. Models are stored with version control for rollback and comparison. Resources are distributed efficiently with GPU/CPU allocation per job. New sites can be added without retraining existing models.

Question 24

How accurate are the failure predictions?

Accepted Answer

Typical accuracy: 85-95% for failure predictions with 6+ months of data. Factors affecting accuracy include data quality (completeness and consistency), equipment age and type, environmental conditions, and maintenance history. The system provides confidence levels for each prediction and continuously improves through feedback loops. Accuracy is higher for well-documented equipment with consistent operational patterns.

Question 25

What happens if a site doesn't have enough historical data?

Accepted Answer

We use transfer learning from similar sites with comparable equipment and environmental conditions. Generic models trained on aggregated data from multiple sites serve as fallbacks. As more data becomes available, the system automatically transitions to site-specific models. Operators can input known equipment issues and maintenance history to improve initial predictions. The system clearly indicates when predictions are based on limited data and provides confidence levels accordingly.

Question 26

Can I customize alert rules?

Accepted Answer

Yes! Administrators can create custom alert rules through the UI with 24 comprehensive rule templates, 9 operators (>, ≥, <, ≤, =, ≠, Between, Outside, In List), 9 metrics (power, voltage, current, temperature, etc.), 8 aggregations (last, avg, median, min, max, sum, delta, rate), time windows, AND/OR logic for complex rules, context filters (daytime-only, minimum load checks), and cooldown periods to prevent alert spam.

Question 27

What notification channels are supported?

Accepted Answer

Alerts can be delivered via real-time dashboard notifications, email alerts with customizable recipients, push notifications (when configured), and webhook integrations to external systems (Slack, Microsoft Teams, OpsGenie, PagerDuty). You can configure different channels for different severity levels.

Question 28

What types of reports can I generate?

Accepted Answer

The platform supports: Automated scheduled reports (daily, weekly, monthly), ESG impact reports with carbon metrics and compliance frameworks (GRI, SASB, TCFD), Performance ratio analysis, Period comparisons (week-over-week, month-over-month, year-over-year), Maintenance logs and work orders, Custom reports with selected metrics and visualizations. All reports can be exported in PDF, CSV, Excel, or JSON formats.

Question 29

How does ESG reporting work?

Accepted Answer

The system automatically calculates CO2 avoided (tons), trees equivalent, cars removed from roads, homes powered, and lives improved based on your site's energy production and regional emission factors. Reports comply with multiple disclosure frameworks (GRI, SASB, TCFD, etc.). You can customize report sections, add contextual information, and export as PDF for stakeholders or CSV for further analysis.

Question 30

What data refresh frequency does the platform support?

Accepted Answer

The platform uses WebSocket technology for real-time data updates. Monitoring data refreshes continuously as new measurements arrive from your solar assets. You can configure update intervals from 1-minute to 15-minute granularity based on your monitoring needs and data volume.

Question 31

Does the platform offer API access?

Accepted Answer

Yes! We provide comprehensive REST APIs for programmatic access to monitoring data, forecasts, alerts, predictions, and reports. Features include authentication via API keys, rate limiting for fair usage, webhook support for real-time notifications, and complete documentation with code examples in multiple languages.

Question 32

How secure is my data?

Accepted Answer

We implement enterprise-grade security including encrypted data transmission (TLS/SSL), encrypted data storage, role-based access control (4 user levels), complete activity audit trails, regular security audits, SOC 2 compliance (in progress), and data backup and disaster recovery. Your data is hosted in secure, redundant cloud infrastructure.

Question 33

Can I export my data?

Accepted Answer

Yes! You can export data at various granularities (15-min, hourly, daily, weekly, monthly) in multiple formats (CSV, Excel, JSON, PDF). Use download icons above charts or navigate to the Export section. For large datasets, the system allows batch exports to prevent timeout issues. All historical data is always accessible.

Question 34

What browsers are supported?

Accepted Answer

The platform works best on modern browsers: Google Chrome (recommended), Mozilla Firefox, Microsoft Edge, and Safari. We recommend using the latest versions for optimal performance. The platform is fully responsive and works on desktop, tablet, and mobile devices.

Question 35

How is pricing structured?

Accepted Answer

Pricing is customized based on your specific needs including number of sites, total capacity (MW), monitoring level requirements, forecast delivery frequency, and user count. We offer both products individually or as a combined package with discounted rates. Contact our sales team for a detailed quote tailored to your portfolio.

Question 36

What kind of support do you provide?

Accepted Answer

We provide comprehensive support including email support (support@helioexpect.com) with 24-hour response time, technical documentation and knowledge base, video tutorials and training materials, dedicated account manager for enterprise customers, onboarding and training sessions, regular product updates and improvements, and emergency support for critical issues.

Question 37

Do you offer training for our team?

Accepted Answer

Yes! We provide comprehensive training including live onboarding sessions for new users, role-specific training (administrators, operators, analysts), video tutorials and documentation, best practices guides, and ongoing support for questions. Training is included with all packages.

Question 38

What is your SLA (Service Level Agreement)?

Accepted Answer

We guarantee 99.95% uptime for our platform. Our SLA includes: maximum 1 hour response time for critical issues, 24-hour resolution target for critical bugs, scheduled maintenance windows (announced 7 days in advance), automatic failover and redundancy, and real-time status page for system health monitoring.

Question 39

Can I add more sites later?

Accepted Answer

Absolutely! Our platform is designed to scale. You can add new sites at any time. For Forecasting, new sites can start receiving predictions immediately. For Predictive Monitoring, we'll train site-specific models using your historical data (1-6 months). Pricing adjusts based on your total portfolio size.

Question 40

How do I get started with HelioExpect?

Accepted Answer

Contact us via our website contact form or email sales@helioexpect.com. We'll schedule a demo to understand your needs, provide a customized quote, set up your trial account, assist with data integration, provide onboarding training, and support you through go-live. The entire process typically takes 2-4 weeks.

Question 41

What information do I need to provide for setup?

Accepted Answer

For setup, we need: Site locations and installed capacity, Equipment details (inverters, SCBs, strings, panels), Access to your monitoring system/SCADA, Historical data (1-12 months, if available), User list with roles, and Integration requirements (APIs, CMMS, etc.). Our team guides you through the entire data collection process.

Question 42

How long does it take to see value?

Accepted Answer

For Forecasting: Immediate value - forecasts start on day one. For Predictive Monitoring: Basic monitoring provides value immediately with real-time dashboards and alerts. Predictive maintenance value increases over time as ML models learn (optimal at 3-6 months). Most customers report ROI within 6-12 months through reduced downtime and improved operations.

Question 43

Do you support integration with our existing systems?

Accepted Answer

Yes! We support integration with: SCADA systems (Modbus, OPC-UA, DNP3), CMMS platforms (via CSV export or API), ERP systems (via API), BI tools (Tableau, Power BI via CSV/JSON), Communication platforms (Slack, Teams, Email), and Custom systems (via REST API and webhooks). Our team assists with all integrations.

How can we help you?

General Questions

Forecasting

Predictive Monitoring

AI & Predictive Analytics

Alerts & Reporting

Technical & Integration

Pricing & Support

Getting Started

Still have questions?