Abstract
AbstractIn an islanded microgrid, while considering the complex nature of line impedance, the generalized droop control fails to share the actual real/reactive power between the distributed generation (DG) units. To overcome this power sharing issue, in this paper a new approach based on feed forward neural network (FFNN) is proposed. Also, the proposed FFNN based droop control method simultaneously controls the microgrid voltage and frequency within the limits. The proposed microgrid consists of combination of photovoltaic (PV) system and battery energy storage system (BESS) as the first DG unit and solid oxide fuel cell (SOFC) as the second DG unit. The simulation of the proposed microgrid is carried out in Matlab/Simulink environment and necessary results are compared to show the effectiveness of the proposed method.
Highlights
Since the beginning of 20th century, fossil fuels such as coal, oil and natural gas has been the main source of energy to meet out the ever increasing energy demand
Distributed generation (DG) units can be powered by both conventional power sources and renewable energy sources (RESs)
A microgrid can be defined as a low/medium voltage network consisting of various DG’s, energy storage systems (ESS) and loads that are normally connected to the utility grid through point of common coupling (PCC)
Summary
Since the beginning of 20th century, fossil fuels such as coal, oil and natural gas has been the main source of energy to meet out the ever increasing energy demand. In Ref. 13, generalized droop control technique by considering the effect of complex line impedance is proposed This technique performs well when the microgrid is operating in grid connected mode. To address the real and reactive power sharing issues for a generalized droop control and to regulate the islanded microgrid voltage and frequency, a new intelligent approach based on artificial neural network (ANN) is proposed. The proposed ANN based droop control technique performs well in controlling the microgrid voltage and frequency under all conditions (during load change as well as during varying irradiance) It performs well by sharing the actual real and reactive power demanded by the load.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have