Name: RAYANA KRISTINA SCHNEIDER BARCELOS
Type: MSc dissertation
Publication date: 09/07/2018
Advisor:
Name |
Role |
|---|---|
| WANDERLEY CARDOSO CELESTE | Advisor * |
Examining board:
| Name |
Role |
|---|---|
| DANIEL JOSÉ CUSTODIO COURA | Internal Examiner * |
| FLAVIO GARCIA PEREIRA | External Alternate * |
| GISELE DE LORENA DINIZ CHAVES | Internal Alternate * |
| WANDERLEY CARDOSO CELESTE | Advisor * |
Summary: This work shows the result of applying characterization techniques to define load signatures in Smart Grids. The differential of this work is that the loads have the same technical data and are from the same manufacturer (loads with a high degree of similarity), making the
identification process more difficult and describing a challenging condition. The prototype is a platform with four technically identical fluorescent lamps, allowing 16 possible operation configurations, this means, from no one lamp turned on to all the lamps turned on. Two
techniques are tested to define the load signature: one with 14 simple features to represent each one of the 16 possible configurations; and another form based on the Shannon and Renyi entropy. Next, the signature sets, classified through Case-Based Reasoning
(RBC), are submitted to an optimizer aiming to find the highest possible accuracy for the identification system. The biggest hit rate obtained in this work is 77,31% and represents a good performance of the identification system, given the complexity of the problem.
These initial results will serve as a reference for new solutions to this new problem.
