System Identification and Modeling Methods for Piezoelectric Proportional Valve in Flow Control Applications

Piezoelectric proportional valves, serving as key actuators in micro-thrust systems, are widely used in cold gas thrusters and high-precision flow control applications. However, due to significant nonlinearities such as hysteresis, creep, and dead zones, conventional analytical modeling methods fail to accurately capture the system dynamics. This study conducts system identification and phenomenological modeling of piezoelectric proportional valves based on experimental input-voltage and output-flow data. First, a linear ARX model is established and its limitations under strong nonlinear conditions are analyzed. Then, the Bouc-Wen hysteresis model is introduced to characterize the hysteresis behavior, and an extended B-W model incorporating a linear component and creep compensation is proposed to enhance model accuracy. Genetic algorithms are employed for parameter identification, achieving a model fitting accuracy above 93%. The proposed modeling approach lays a solid theoretical foundation for precise control of piezoelectric proportional valve systems and provides critical support for the design of subsequent nonlinear compensation strategies.