Powered by
© Fyper VOF
Conference Websites
11:40   Session 2: Power System Design
Chair: Stephane Brisset
11:40
20 mins
FITTING MEASURED PHOTOVOLTAIC CURRENT-VOLTAGE CURVES: AN EFFECTIVE BENCHMARK FOR THE OPTIMIZATION TECHNIQUES
Antonino Laudani, Francesco Riganti Fulginei, Gabriele Maria Lozito, Alessandro Salvini, Salvatore Coco
Abstract: The one-diode equivalent circuit of photovoltaic (PV) panels is widely utilized as in scientific research, as in industrial applications, even if its identification from experimental data is a very difficult task in which five parameter must be detected. Several proposals are presented in literature to face this inverse problem by using modern heuristics. On the other hand, the high complexity of the five-parameter model identification allow to see this problem as a serious candidate to be a new effective optimization benchmark for testing performance and robustness of optimization algorithms. Obviously, in this case, we need to know where the global optimum is located, whereas in literatures just local optima are proposed. However, recently, a reduced form of the five-parameter model has been proposed. This last approach has permitted to reach an accuracy never reached before in the identification by any optimization method. Thus, by assuming as global optimum the solution returned by the reduced form approach, we propose the identification of the five-parameter model as a new benchmark, to be used for testing the different performance of optimization algorithms in the solution of inverse problems.
12:00
20 mins
OPTIMAL SIZING INTEGRATING POWER MANAGEMENT FOR A MICROGRID WITH STORAGE
Remy Rigo-Marianni, Bruno Sareni, Xavier Roboam
Abstract: The paper presents an optimal design process for a microgrid with an industrial load associated to a photovoltaic power plant and a storage unit based on high speed flywheel. The power management is simplified with a fast linear programming algorithm which allows simulating the whole system over a long period of time integrating both operating and sizing loops. The overall design procedure is led by the Efficient Global Optimization that interpolates the objective function with a kriging technique. Then, the optimal sizes for both storage and PV production are found in a reduced number of objective function evaluations.
12:20
20 mins
MILP FORMULATION OF A PEAK LOAD REDUCTION PROBLEM THROUGH LOAD SHIFTING IN LOW VOLTAGE NETWORKS
Clementine Benoit, Aurelien Mercier, Frédéric Wurtz, Yvon Besanger
Abstract: In this paper, we present a MILP formulation of a peak load reduction through load shifting problem. A low voltage network having 53 loads is considered over 29 days, with a 5 minutes step time. Each load can be shifted at most once a day, and a payback effect has been considered.