Open AccessIntegrating LQR with GRNN for LFC of energy delivery system interconnected via AC/DC tie‐lines
Author(s) -
Sharma Gulshan,
Panwar Akhilesh,
Krishnan Narayanan,
Kumawat Manoj
Publication year - 2020
Publication title -
iet energy systems integration
Language(s) - English
Resource type - Journals
ISSN - 2516-8401
DOI - 10.1049/iet-esi.2020.0021
Subject(s) - linear quadratic regulator , control theory (sociology) , controllability , observability , weighting , artificial neural network , controller (irrigation) , energy (signal processing) , computer science , control engineering , engineering , control (management) , mathematics , artificial intelligence , statistics , medicine , biology , agronomy , radiology
The present study shows an early attempt to integrate the design of discrete linear quadratic regulator (LQR) with generalised regression neural network (GRNN) to achieve the standards of load frequency control (LFC) of an energy delivery system linked via AC/DC tie‐lines for assorted operating conditions. The design of state cost weighting matrix ( Q ) has been performed for different conditions namely (i) for controllability and observability technique, (ii) for optimisation and (iii) by providing equal weightage to all the system states while designing the discrete LQR LFC. The enactment of LFC design is checked for standard load disturbance considering the different sample time with diverse structures of Q . Further, the feedback gain matrix obtained via implementing the discrete LQR for control areas of power system using different models of Q is used to meritoriously train the GRNN for assorted operating conditions. At last, the performance of the designed GRNN is evaluated for extensive operating conditions with and without non‐linearities and the obtained results are compared with multi‐layer perceptron founded artificial neural network and genetic algorithm founded LFC of an energy delivery system.