Skip to Main Content
The electric power system is a non-linear dynamic system, with continuous condition changes during normal operation. Therefore, it is required to have a special control strategy to cover different operating points. One of the proposed techniques to solve such problem is the use of a multiple model adaptive control (MMAC), allowing that the controller updating to represent the actual system properly. The use of a MMAC strategy is based on a physical system output measurement for system model estimation. So any noise from the output sensor measurement leads to a significant increase in the residual (the error in the output signal between the physical system and the designed model),. So the weight computation block uses this residual to calculate the weights for all controllers in the controller bank. The highest residual model means that the model is far from the true system. Therefore, it is important to remove the noise from the residual calculations. In this work the new observer is installed to provide the full state estimation and noiseless output. So the effect of measurement noise is avoided and is independent of the type of noise. The system is tested by adding unknown noise to the system output measurement. The design of the local controller is based on different linear control techniques to achieve both local and global system stabilities. The damping improvements are verified using the Matlab software.