Photovoltaic energy storage step-up transformer

How to choose a step-up transformer in a PV plant?

In general, the selection of the step-up transformer in a PV plant is a quite complex task as several variables depending on the transformer rated power must be taken into account as: initial cost of the system, energy losses due to transformer efficiency, energy storage system efficiency and possible plant disconnections due to grid instability.

Should a transformer be rated near a PV plant peak power?

In fact, while selecting a transformer rated power close to the PV plant peak power makes theoretically possible to fully transfer the captured solar energy to the utility network, such a design criterion will in practice lead to oversize both the transformer, the inverter and the power line.

What is a grid-tied PV system without energy storage?

Before untangling more puzzling windings decisions for isolation transformers, transformers with energy storage in microgrid scenarios, or PV systems supplying both three-phase and single-phase dedicated loads, let us consider a common case: a grid-tied PV system without storage. In this scenario, the PV system is exporting power to the grid.

What is multi-step ahead PV power forecasting (pvpf)?

In this research, the multi-step ahead PV power forecasting (PVPF) problem is dealt with for predicting the next day's hourly power generation, which have different applications, such as making an energy storage policy and deciding the system marginal price by comparing the energy forecasts with the next day's energy consumption.

How does a centralized inverter work in a PV plant?

As shown in Fig. 1, in a traditional PV plant a large number of PV modules are series connected in long strings and a single centralized inverter provides the voltage inversion. Step-up transformers are required to boost the 480÷690 V inverters output voltage to the 13.8÷46 kV of the medium voltage utility network .

How is photovoltaic power generation forecasted?

Photovoltaic power generation is forecasted using deep learning. Weather observation and forecast, and solar geometry data are used as input. Three variants of the transformer networks are designed for the power forecasting. The networks were evaluated with the data of two power plants in South Korea.