Project overview

• Ref.: SEE26-IOT-02 — Solar panel tracking & energy yield forecast.
• Measure in real time the production and condition of photovoltaic panels and environmental parameters to support forecasting and maintenance planning.

Objectives & core tasks

• Build short (1–24 h) and medium (1–7 days) horizon production forecast AI models to optimize usage and plan maintenance.
• Implement real-time measurement of panel production, panel condition and environmental sensors, and integrate these data streams for modeling.

Anomaly detection & condition monitoring

• Detect anomalies and degradations using anomaly detection algorithms and business rules applied to time-series sensor and production data.
• Define alerting rules and thresholds, implement automated flags for maintenance planning and further inspection.

Technical approach & data

• Work on time-series forecasting models (examples: LSTM, Temporal Convolutional Networks, Transformers, and tree-based regressors) and baseline statistical models; perform feature engineering from irradiance, temperature, and panel telemetry.
• Handle IoT data ingestion (MQTT/HTTP/edge collectors), data cleaning, resampling, and labeling; evaluate models with metrics such as MAE/RMSE and calibration on 1–24 h and 1–7 day horizons.

Deployment & deliverables

• Deliver trained forecasting models, anomaly detection pipelines, evaluation reports, and a prototype dashboard/visualization for production and alerts.
• Provide reproducible code, model packaging (Docker/containers), documentation for data schemas, and handover notes for operations and maintenance.

Candidate profile & required skills

• Strong skills in Python, time-series machine learning, libraries such as scikit-learn, TensorFlow or PyTorch, and experience with data pipelines (Pandas, SQL).
• Familiarity with IoT data ingestion, sensor telemetry, signal preprocessing, and basic cloud/edge deployment (Docker, REST APIs, or cloud services).

How to apply

• Apply online: https://lnkd.in/g3V53uMH .
• Use the project reference SEE26-IOT-02 in your application and in the email subject if you apply by email.