Neural-adaptive controller for an isolated power converter

The paper presents the design of a neural-adaptive controller for an isolated power converter. Accurate estimation of the unknown parameters is taken into account in both theoretical and practical domains when considering intelligent control techniques. Despite the construction of the mathematical model for the converter, some unstructured parameters have not been included in the model. The proposed resilient controller is designed to adapt to both structured and unstructured factors and provide the appropriate output. The neurons or activation functions of the controller compensate for the unstructured properties of the converter. With an increase in the number of neurons, the converter's output gains more resilience and flexibility, even in the presence of changes in the reference values. The adaptive gain is augmented proportionally to enhance the system's resilience in achieving the target output voltage. The controller's stability study is conducted using the Lyapunov Function. The derivative of the Lyapunov function's convergence yields the adaptation law, which modifies the weights associated with the converter's neurons and aligns with the reference model to achieve the desired outcome. The results are validated using the OPAL-RT real-time simulator platform.