Title Pastato elektros energijos suvartojimo adaptyvaus modelio sukūrimas ir tyrimas /
Translation of Title Development and research of adaptive model for building electricity consumption.
Authors Ignatavičius, Stanislav
Full Text Download
Pages 104
Keywords [eng] electricity consumption ; statistical forecasting models ; autoregressive forecasting models ; exponential smoothing forecasting model ; artificial neural network
Abstract [eng] The final master thesis - "Development and research of adaptive model for building electricity consumption". In the literature analysis part of this project the overview of statistical and artificial intelligence forecasting models are presented, describing their suitability to forecast the electricity consumption of a little-known individual building. Also, the theoretical part provides an overview of the Fuzzy logic regulator as a adaptability tool for forecasting model to update the values of the optimal forecasting model parameters. The methodology part of this project contains information about the selected individual building, for which electricity consumption forecasting models are created. This section also provides an overview of the methodology used to create autoregressive prediction models as well as an overview of the algorithms which are being used to create forecasting models. The methodology section also provides a brief overview of the used Matlab software and the Matlab extensions packages. In the analytical part of this final project the creation and research of the statistical and artificial intelligence forecasting models are presented. Created forecasting models are compared in order to determine the optimal forecasting model for chosen individual building. Also, in the analytical part, an Fuzzy logic regulator is created as a tool for adapting the parameters of the optimal forecasting model. Analysis of the Fuzzy logic regulator is performed to determine an optimal configuration of the regulator. A comparison is made between an optimal prediction model with parameter update and an optimal prediction model without parameter update.
Dissertation Institution Kauno technologijos universitetas.
Type Master thesis
Language Lithuanian
Publication date 2019