Full text
| |
| Author(s): |
Wesley Angelino de Souza
Total Authors: 1
|
| Document type: | Doctoral Thesis |
| Press: | Campinas, SP. |
| Institution: | Universidade Estadual de Campinas (UNICAMP). Faculdade de Engenharia Elétrica e de Computação |
| Defense date: | 2016-11-07 |
| Examining board members: |
Luiz Carlos Pereira da Silva;
Denis Vinicius Coury;
Esther Luna Colombini;
José Antenor Pomilio;
Paulo Fernando Ribeiro
|
| Advisor: | Luiz Carlos Pereira da Silva; Fernando Pinhabel Marafão |
| Abstract | |
This thesis presents a new methodology for power monitoring and management through smart energy meters. This methodology is based on three techniques that involve the load disaggregation, the consumption management system and the management of photovoltaic microgenerators system with battery. The load disaggregation technique performs the calculation of load indicators using the Conservative Power Theory (CPT). These indicators are calculated by the waveforms of voltage and current and with these indicators for various equipment, an dataset for the pattern recognition algorithms was created. Among many options, five classifiers algorithms were chosen: support vector machine (SVM), optimum path forest (OPF), multilayer perceptron (MLP), K-nearest neighbor (KNN) and the decision tree (DT), which were compared in terms of accuracy and computational time. At this point the KNN was presented as the best algorithm for the created dataset. The power management system (supervisory system) uses the concept of sending meter data to a database. Thus, an interface makes gets the database information and presents reports to consumer in daily, weekly, monthly and annual periods. The system also performs the forecast consumption using the time delay neural network algorithm (TDNN) and creates energy saving strategies by means of time reduction or the use at times that the price of energy is lower (considering the existence of time pricing). The mini/micro generation management system technique uses historical consumption information, weather forecast and battery power to propose strategies to reduce electricity costs. The proposed methodology was evaluated using a developed prototype, demonstrating their feasibility in an electronic energy meter, enabling the development of cognitive smart meters and giving coherence to the "smart" term of the meters (AU) | |
| FAPESP's process: | 12/19375-1 - Study of analysis and technologies techniques for the development of energy smartmeters with a focus on residential microgrids |
| Grantee: | Wesley Angelino de Souza |
| Support Opportunities: | Scholarships in Brazil - Doctorate |