Abstract
Accurate short-term wind speed forecasting is a cornerstone of reliable grid integration for wind energy systems. Classical machine learning models, despite their maturity, face inherent limitations in handling the high-dimensional, non-linear, and non-stationary characteristics of SCADA-acquired meteorological data from Indian wind farms. This paper proposes a Hybrid Variational Quantum Circuit (HVQC) framework that synergizes classical deep learning with parameterized quantum circuits to enhance forecasting fidelity. The proposed HVQC architecture employs a Conv1D-LSTM Classical Encoding Module (CEM) to extract temporal context from 24-step multivariate windows of 10-minute resolution SCADA data, feeds the encoded representation through an 8-qubit, 4-layer parameterized Variational Quantum Circuit using angle encoding and Ry/Rz rotational gates with circular CNOT entanglement, and maps quantum measurement expectation values ⟨Z⟩ to wind speed forecasts through a dense Classical Regres