Abstract
The surge in energy demand needs to be met with environmentally pleasant resources to reduce the production of greenhouse gases. Solar Photovoltaic (PV) power is a widespread choice as it is accessible in plenty and is comparatively inexpensive. However, the large-scale penetration of intermittent PV power causes multiple variabilities in the grid such as frequency issues and voltage deviations. To counteract these instabilities, Battery Energy Storage System (BESS) is integrated into the grid as it reduces the PV fluctuations and promotes optimal operation. Nevertheless, storage systems are expensive, and thus smoothing filters are coupled with the BESS for cost reduction and power smoothing. Formerly, traditional filters such as Low Pass Filters (LPF), Moving Average (MA), and Moving Median (MM) filters have been proposed for power smoothing. However, these filters have inadequate power tracking capabilities particularly with longer window sizes (W.S) and time constants, which subsequently depreciates the storage system performance. To compensate for the delayed power tracking, larger energy storage systems are required which in turn adds to the overall operational costs. This paper proposes the Moving Regression (MR) filter combined with state of charge (SoC) feedback control for solar PV variability reduction, reduced time delay, decreased battery charging/discharging power, and ramp rate. Simulation results attest that the MR filter achieves better solar power smoothing without increasing the BESS capacity. Also, the performance of the MR filter is less affected with the increase in window sizes. Additionally, the execution of the introduced smoothing filter was found to be superior when assessed against the LPF, MA, MM, Savitsky-Golay (SG), and the Gaussian filter. In comparison to the SG (W.S = 53), the MR (W.S = 45) filter reduces the battery charging/discharging power by approximately 30.48% and as opposed to the LPF (T.C = 48), the peak SoC is reduced by 19.1%.
Highlights
Over the years there has been an increase in the earth’s population which is directly proportional to the energy used as well
The proposed model has significantly better overall performance when compared to other current relevant smoothing methodologies
An original approach is proposed based on the moving regression filter combined with the state of charge feedback control to flatten the highly varying solar power output and to decrease the ramp rate to an appropriate degree
Summary
Over the years there has been an increase in the earth’s population which is directly proportional to the energy used as well. In order to reduce the battery size and the time delay caused by an MA filter, [55] proposes a halved window moving average algorithm coupled with an integral control method to flatten the power, and to limit the alteration in the battery SoC that would have been stemmed from a regular moving window procedure with the BESS. This paper contributes by proposing a Moving Regression (MR) filter combined with SoC feedback control and BESS for smoothing of solar PV variabilities, reduction in power lag, ramp rate, charging/discharging power and BESS optimization. A performance based comparison between the MR and the conventionally used smoothing techniques has been conducted, simulation results confirm that the MR filter has superior performance efficiency in power smoothing and tracking, followed by its exceptional battery charging/discharging power and SoC management capabilities.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
