Search

In this study, first of all, for the one-layered winding form of the rotating electrical machines, self-induction coefficient of slot leakage has been calculated according to varying forms of slot. Given the differences in the rotating electrical machines, a dataset has been created with the slot size measurements required for each slot form as well as rotor connection forms (with squirrel cage with wound rotor) representing input and with the slot equivalent conductivity representing the output. Then, a dataset has been used to train an ANN forecasting model. Coefficient of slot leakage self-induction so calculated was compared with coefficient of predicted slot leakage self-induction. With the trained ANN model, slot equivalent conductivity of a rotating electrical machine with any size for a possible slot form has been predicted with an accuracy rate of 99.998%.

The paper presents further development steps of recently proposed model of an adaptive multi-agent-based e-social care system by introducing in it an intelligent power controller to supporting people in a Smart Ecological and Social Apartment, the SESA. The schematic diagrams of the controller are proposed for adaptive control of motors and LED lights as well as for measurements of EDA, HR, and FT signals of inhabitancies in the SESA. The proposed in this paper multi-agent system consists of two groups of agents, first, the agents for human’s physiological data collection, analysis and emotional state detection, and, second, the agents for an intelligent power control which takes environmental and hardware parameters and, with the use of embedded agents, provides adaptive control of LED lights, temperature as well as of power consuming by motors used in the SESA in accordance with the human state defined by the first group of agents.