Research on Measurement of Load Angle of Hydroelectric Generating Unit Based on Soft Sensing

Accurate online measurement of generator load angle is critical for stability analysis in hydroelectric power systems. The existing load angle measurement method exhibits significant measurement errors in the hydropower unit with many poles and slow speed, this paper proposes a novel soft sensing method based on Particle Swarm Optimized Radial Basis Function Neural Network (PSO-RBFNN) to address the limitations of conventional measurement approaches. The methodology establishes a nonlinear mapping between electrical parameters (terminal voltage, active/reactive power, field current) and load angle through intelligent fusion of mechanism analysis and data-driven modeling. By introducing adaptive acceleration factors and three-dimensional particle encoding strategy, the improved PSO algorithm effectively optimizes the RBFNN’s basis function centers, connection weights, and network topology. Validation on both single-machine infinite bus system and fourmachine two-area system demonstrates superior performance with mean square error below 0.0034 during transient processes (three-phase short-circuit faults) and 0.0015 under steady-state operation. The pure electrical measurement strategy enables noninvasive online monitoring without generator shutdown, showing significant engineering applicability for low-speed hydropower units with multi-pole characteristics.