Urban data analytics for better power grid management

• Main Author: Selvarajoo, Stefan
• Other Authors: Schläpfer Markus
• Format: Final Year Project
• Language: English
• Published: 2018
• Subjects: DRNTU::Engineering::Computer science and engineering::Mathematics of computing::Probability and statistics
• Online Access: http://hdl.handle.net/10356/73978
• Institution: Nanyang Technological University

Electric load forecasting is a field where continuous, rigorous efforts are made to improve models which predict energy consumption. Companies that require a certain amount of energy will factor in cost as a means of requiring what is necessary while energy producers are pushed toward a “smarter, big data approach” to aggregate information and generate power thereby increasing efficiency. Proposals range from rudimentary techniques involving linear regression to machine learning with various algorithms set to improve models of forecasting the electric load. With a focus toward aggregating data from cell phone users to electric load consumption, predicting energy models can prove to be method in developing a model to predict energy consumption. Firstly, Cell-phone Data Records(CDR) will be processed and mapped to grid cells from the province of Trentino located in the region of Trentino-Alto Adige to achieve a basic visualisation of the movement of people over a period of 1 day. Next, a correlation between CDR and census population is made as a baseline of people within the various municipalities. After which, a standard model will be used to estimate the population at a given time using correlated CDR. Next, with the estimated population, another correlation will be made with the actual Ampere values presented from the official data source. Again, a model will be derived by measuring the estimated population and electrical demand flowing through various municipalities. Finally, a forecast model using baseline methods of moving average and mean will be developed to show the efficiency of using CDR to estimate and forecast electricity demand of a region at a given time.