SELF-PVP: Self-Organizing Power Management for Photo-Voltaic Power Plants
Start date: 2010 Expected completion date: 2013Ph.D. Dual Degree Students:
PIs: Vítor Manuel Grade Tavares (FEUP / INESC Porto) and Shawn Blanton (CMU)
Teams: Instituto de Engenharia de Sistemas e Computadores do Porto (INESC Porto), Faculdade de Engenharia da Universidade do Porto (FEUP), Faculdade de Ciências e Tecnologia da Universidade Nova de Lisboa (FCT/UNL), Carnegie Mellon University (CMU)
Companies: Autonomia - Recursos Renováveis SA
Xuanle Ren (Electrical and Computer Engineering)
Keywords: Microelectronics, Wireless Sensors, Adaptive Systems, Post-Silicon electronics
This project presents a line of research that aims to achieve at least 15% increase in power efficiency in a photovoltaic (PV) power plant, using a novel, distributed, real time and on-line, adaptive network controller of sensors/actuators to bring optimality to the overall power output of the panels’ array. The goal is to build a self- organizing, truly distributed computational network reflecting ambience intelligent, with the ability to sense and control the operation point of each panel in a PV power plant, to accomplish a global maximum power available at any environment condition of operation. The size of the network may become massive, if one takes the example of the Amaraleja power plant, in Alentejo, Portugal, the number of nodes could go above 200,000 (one node per panel), spread in a large area (250 hectares). Moreover, we also plan to innovate all the local electronics for the sensor nodes by using post-silicon transparent technology that can be integrated in the protective glass of the photo-voltaic (PV) panels, providing a cost effective electronics integration with little impact on the PV cell and panel assembly. Due to the specificity of the sensor network (heterogeneous data, real time constraints, several communication modes – broadcast and point to point) a dedicated wireless system design will also be undertaken. Although the research here presented has a key application in mind - the power optimization in PV power plant – we envision many other implications - tackling generic global optimization with sensor networks.