Système inertiel de stockage d'énergie couplé au générateur photovoltaïque et piloté par un simulateur temps réel

摘要

The subject is part of the strategy to increase the penetration of renewable energy in power systems, particularly those that are poorly interconnected, such as island grids. A limit of penetration of intermittent energy by 30% in instantaneous power in these electrical grids was set by a French law. To help overcome this limitation, a solution is to couple the sources of decentralized and intermittent generation with energy storage systems. In this thesis, we are interested in flywheel energy storage systems (FESS) that converts electrical energy in kinetic energy form and vice versa. FESS have a number of cycles charge / discharge large compared with electrochemical batteries and can be used to smooth the photovoltaic power generation. The fluctuation of photovoltaic instantaneous power is indeed weakly predictable over time and it cannot be controlled, including its production fall. PV production can decrease up to 80% of its maximum power in 30 seconds, and so destabilize the grid. The island grids, such as that of Corsica, are not interconnected to the mainland power grid. The non - interconnected grids are more fragile and less stable. Thus, the massive development of photovoltaic power plants can cause fluctuations in the frequency and voltage. The flywheel has the advantage of having a low response time (a few hundred milliseconds). However, it has a lower energy capacity. The benefits of FESS are used by managing it in real time with an appropriate computer. A flywheel with a power of 15 kVA and an energy capacity of 112 Wh was characterized and tested at INES Chambery using a real time grid simulator (RTLab®), a real-time computer (dSPACE®) and a PV power plant. The storage system is composed by an asynchronous electrical machine and a cylindrical steel flywheel. The Matlab Simulink / software is used to implement the control laws necessary for its control. In this thesis, the test bench is presented and the results of ancillary services (power smoothing, frequency and voltage regulation). Three power smoothing methods are discussed and evaluated (smoothing with a transfer function, with a slope limiter function and a method not using any smoothing function). The third method uses neither a transfer function, nor a function that limits the slope variations, requires fewer parameters, and is more optimal and more robust. A flywheel with another electrical machine technology (the switched reluctance machine) has also been characterized. This is an Uninterruptible Power Supply (UPS) on which parameters such as self-discharge and efficiencies (charging mode, discharging mode and standby mode) were measured.