Bazefield Solar AI Analytics (hereby as “Solar AI”) is a premium product extension for thoroughly analysing data from solar photovoltaic power plants.


Solar AI Analytics provides an indispensable set of in-depth analytic tools that automatically break down system losses into categories to give accurate insights into the impact and possible mitigation of:

  • Curtailment
  • Degradation over time
  • Down strings
  • Energy loss due to VAr support
  • Inverter clipping
  • Inverter downtimes, categorized by duration and root cause
  • Inverter efficiency, relative to manufacturer specifications
  • Inverter partial capacity loss
  • Module thermal losses
  • Module washing
  • Night-time losses
  • Nominal plant capacity loss
  • Shading, accounting for obstructions, vegetation, and intra-array shade
  • Snow losses
  • Soiling, including zone-level assessment of different parts of each plant
  • Tracker errors, including stowing outside of wind events
  • Underperforming strings
  • Wind stowing of trackers

1. Main Functions

Solar AI calculates industry-standard key performance indicators (KPIs). It offers a detailed breakdown of system losses using state-of-the-art algorithms and models to inform asset managers, operations and maintenance (O&M) teams, and insurance underwriters on the as-built performance of photovoltaic plants. These algorithms translate system under-performance into prioritised corrective actions that help increase revenue, organise and schedule O&M more efficiently and cost-effectively, and generally ensure the long-term health of PV assets.


Based on detailed analysis and loss breakdown, Solar AI can achieve the following main functions:


  • Data quality assurance: Data quality is crucial for accurate and effective analysis. In a first step, Solar AI automatically detects and corrects missing or invalid data, ensuring the accuracy of KPI and loss breakdown analysis results.
    • Use complex and powerful automatic data filters to detect data anomalies.
    • Ensure data accuracy by including/excluding specific operating conditions of photovoltaic power plants.
    • Automatically identify the weather data anomaly.
    • Tag missing data and data anomalies.


  • Downtime event analysis: Solar AI analyzes all inverters in the photovoltaic power plant and provides a detailed analysis of the causes of losses. On this basis, the risk of photovoltaic power plant underperformance can be reduced.
    • Convert continuous inverter downtime to individual downtime events for each inverter to provide downtime analysis at the inverter level.
    • Analyze downtime events by category, such as snowfall, curtailment, or other causes, to better understand the causes of downtime.


  • Accurate and targeted operation and maintenance recommendations: Solar AI automatically determines the most efficient and economical corrective actions. The operation and maintenance team can perform targeted corrections to under-performance issues based on recommended operation and maintenance suggestions, and solve anomalies that may cause future equipment problems.
    • Provide corrective action suggestions based on the loss breakdown analysis.
    • Update the corrective action list every day to avoid missing any anomalies.
    • Lead operators to follow data-driven recommended actions.


  • Degradation analysis: Solar AI regularly and automatically calculates the system degradation rate of the photovoltaic system by utilizing various advanced algorithms to robustly and reliably derive the degradation on the DC side of the station, providing characteristic results of the operation status of the photovoltaic power plant, and providing evidence support for submitting warranty claims.


  • Accurate analysis of shadow occlusion and dust accumulation: Solar AI can create a dynamic ideal model of theoretical power production for each inverter to reveal real shading conditions , including the shadow occlusion caused by the growth of vegetation such as weeds, which can also be detected and prompted to take reasonable action. These results can provide meaningful benchmarking for the power production evaluation and performance model of system design. In addition, Solar AI will use historical data to detect the accumulation trend and pattern of dust in each inverter, and dynamically evaluate the dust loss within the entire photovoltaic power plant.


  • Wash Optimization:Based on the results of the dust algorithm, Solar AI estimates the next most cost-effective washing date. This can help to maximize the normal operation time of photovoltaic power plants and reduce washing costs.
    • Obtain the most cost-effective washing plan based on system performance and dust conditions.
    • Optimize washing plans for specific areas within the system to adapt to different designs and dust characteristics.


  • Verify operations and contract compliance: Solar AI can supervise and evaluate the overall operation and maintenance level of photovoltaic power plants.
    • Verify the planned operation and maintenance work, such as automatically detecting component washing events in photovoltaic power plants, and confirming whether effective washing has been carried out, washing dates, and whether effective washing has been carried out.
    • Track the duration of inverter downtime and string downtime events.
    • The analysis of planned operation and maintenance work, especially the analysis of the benefits of operation and maintenance activities, can help to dynamically adjust the operation and maintenance activities in the future cycle.


  • Evaluate the operation and maintenance level and indicators of the power station:
    • Specify reasonable operation and maintenance requirements based on the performance of the power station at the time of construction and the data situation of the power station.
    • Eliminate performance issues unrelated to operation and maintenance in the analysis process of power plant operation and maintenance level.
    • Through transparent analysis of downtime events and root causes, help identify the responsible parties for downtime issues.