Use this URL to cite or link to this record in EThOS: http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.538193
Title: Nonlinear self-tuning control for power oscillation damping
Author: Arif, Jawad
Awarding Body: Imperial College London
Current Institution: Imperial College London
Date of Award: 2011
Availability of Full Text:
Access through EThOS:
Full text unavailable from EThOS. Please try the link below.
Access through Institution:
Abstract:
Power systems exhibit nonlinear behavior especially during disturbances, necessitating the application of appropriate nonlinear control techniques. Lack of availability of accurate and updated models for the whole power system adds to the challenge. Conventional damping control design approaches consider a single operating condition of the system, which are obviously simple but tend to lack performance robustness. Objective of this research work is to design a measurement based self-tuning controller, which does not rely on accurate models and deals with nonlinearities in system response. Designed controller is required to ensure settling of inter-area oscillations within 10−12s, following disturbance such as a line outage. The neural network (NN) model is illustrated for the representation of nonlinear power systems. An optimization based algorithm, Levenberg-Marquardt (LM), for online estimation of power system dynamic behavior is proposed in batch mode to improve the model estimation. Careful study shows that the LM algorithm yields better closed loop performance, compared to conventional recursive least square (RLS) approach with the pole-shifting controller (PSC) in linear framework. Exploiting the capability of LM, a special form of neural network compatible with feedback linearization technique, is applied. Validation of the performance of proposed algorithm is done through the modeling and simulating heavy loading of transmission lines, when the nonlinearities are pronounced. Nonlinear NN model in the Feedback Linearization (FLNN) form gives better estimation than the autoregressive with an external input (ARX) form. The proposed identifier (FLNN with LM algorithm) is then tested on a 4−machine, 2−area power system in conjunction with the feedback linearization controller (FBLC) under varying operating conditions. This case study indicates that the developed closed loop strategy performs better than the linear NN with PSC. Extension of FLNN with FBLC structure in a multi-variable setup is also done. LM algorithm is successfully employed with the multi-input multi-output FLNN structure in a sliding window batch mode, and FBLC controller generates multiple control signals for FACTS. Case studies on a large scale 16−machine, 5−area power system are reported for different power flow scenarios, to prove the superiority of proposed schemes: both MIMO and MISO against a conventional model based controller. A coefficient vector for FBLC is derived, and utilized online at each time instant, to enhance the damping performance of controller, transforming into a time varying controller.
Supervisor: Chaudhuri, Balarko Sponsor: National University of Sciences and Technology, Pakistan
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID: uk.bl.ethos.538193  DOI: Not available
Share: