Design of a hybrid fuzzy/Markov chain-based hierarchal demand-side frequency control

A developed model of the Thermostatically-based controllable loads (TCLs) is proposed for a hierarchal Demand Side Frequency Response (DSR). The model involves three layers based on Markov chain, Fuzzy Logic, and Particle Swarm Optimization (PSO) method. Layer 1 is the local Markov population of the TCLs, Layer 2 is the Fuzzy/Markov-based Frequency controller, and Layer 3 is the PSO-based Supervisory control. Layer 1 can detect a non-zero frequency deviation to start-up the process of Layer 2. Layer 2 is dynamically updating the switching probabilities of the Markov model of Layer 1 to control the power consumption of the controllable load. Layer 3 is dealing with a global optimization issue to obtain the start/end values of the switching settings as well as the amount of the droop value. The proposed design was tested in the Great Britain (GB) single bus machine as well as a multi-machines dynamic benchmark system.