Transformer winding condition detection based on the nonlinear characteristics of vibrations

Abstract

Power transformers are critical devices in power grid systems, with windings as their core components. To perform state detection on online windings, this paper establishes a nonlinear model between vibration amplitude and current based on the relationship between winding structural parameters and load variations. The numerical relationship between structural stiffness and model parameters is proved, and a winding clamping force evaluation method based on the nonlinear characteristics of vibrations is proposed. In the experiment, the relationship between current and vibration amplitude is obtained by adjusting the axial clamping force of windings and performing short-circuit load tests. The experimental results show that the nonlinear characteristics of vibrations are significant when the winding clamping force is small, and the characteristic parameter C3 accurately reflects the equivalent stiffness change of the winding structure. In practical applications, the vibration characteristics of oil tank surface corresponding to normal windings and loose windings are compared, and the results are presented in the form of two-dimensional contour plot. The results indicate that the vibration characteristics of the tank reflect the structural state of the winding near the measurement point. The above results demonstrate that the parameter C3 is a monotonic function of clamping force, and there is structural degradation in the winding when this value is greater than 0.5.

Keywords

• transformer winding
• fault diagnosis
• electric-vibration model
• nonlinear characteristics
• clamping force state
• vibration image