Fast estimation of synchronizing and damping torque coefficients using an adaptive neural network

This paper presents an adaptive technique for the estimation of the synchronizing and damping torque coefficients of a synchronous machine using a linear adaptive neuron (Adaline). The proposed technique is based on estimating the synchronizing and damping torque coefficients from online measurements of the changes of the rotor angle, rotor speed, and electromagnetic torque of the synchronous machine using Adaline architecture. These coefficients can be used as indices, which provide insight into the small-signal stability of power systems. The proposed approach can quickly predict the possible unstable conditions and hence help the operator to take the correct control action beforehand. The performance of the Adaline is compared with both Kalman filter and least-square error techniques. The Adaline offers several advantages including significant reduction in computing time and storage as well as remarkable reduction in the computational complexity, which is associated with Kalman filter. The simulation results over a wide range of operating conditions show that the Adaline can be used as efficient tool for either online or offline estimation of the synchronizing and damping torque coefficients. Therefore, it is believed that Adaline is a strong candidate for online monitoring of small-signal stability and security.