Intellectual Property India Publishes Patent Application for 'GAN-Based Dual-Stream PV and Load Forecasting System With Integrated False Data Injection (FDI) Attack Detection for Smart Grids' Filed by Jis College Of Engineering

MUMBAI, India, May 1 -- Intellectual Property India has published a patent application (202631051543 A) filed by Jis College Of Engineering, Kalyani, West Bengal, on April 22, for 'gan-based dual-stream pv and load forecasting system with integrated false data injection (fdi) attack detection for smart grids.'

Inventor(s) include Partha Das; Gargi Roy; Debodyuti Updadhaya; Animesh Halder; Agnik Kundu; and Sayan Ghoai.

The application for the patent was published on May 1, under issue no. 18/2026.

According to the abstract released by the Intellectual Property India: "The present invention discloses a GAN-based dual-stream photovoltaic (PV) and load forecasting system with integrated false data injection (FDI) attack detection for smart grids. The system collects real-time PV generation data, load consumption measurements, and meteorological parameters from distributed sensors, SCADA networks, and IoT-based smart meters. A preprocessing module performs time synchronization, noise filtering, normalization, and missing data identification. A Generative Adversarial Network (GAN)-based synthetic data engine generates realistic time-series samples to compensate for missing, corrupted, or low-quality measurements, thereby improving forecasting reliability. The invention employs two parallel forecasting engines, wherein a CNN-LSTM hybrid model is used for PV power forecasting and a BiLSTM with attention mechanism is used for load demand forecasting. Additionally, the system includes a real-time FDI attack detection module combining GAN discriminator authenticity scoring, autoencoder reconstruction error evaluation, and statistical deviation analysis to detect stealthy cyberattacks. Upon detection of anomalous or tampered data, a decision fusion layer replaces suspicious measurements with GAN-reconstructed synthetic values and generates validated forecasting outputs along with cybersecurity alerts. The invention ensures accurate forecasting, enhanced data integrity, and resilient smart grid operation under normal and adversarial conditions."

Disclaimer: Curated by HT Syndication.