Search

Distributed electrical energy storage can help reduce the CO2 emissions associated with the use of electrical energy, better enabling distributed generation of energy from sources such as rooftop photovoltaic (PV) systems. However, electricity distribution systems were not designed to allow power to flow from consumers to the grid, limiting how much power can be exported from rooftop PV systems. Falling feed-in tariffs are also making it more cost-effective to store excess PV energy on site, rather than export excess energy to the grid and then import it later at a higher cost.This study determines the impact and viability of distributed electrical energy storage systems for residential consumers with rooftop PV systems. It uses PV generation and the total power being drawn by electrical appliances for thirty-eight households at Adelaide’s Lochiel Park, a housing development of about 100 homes designed to demonstrate low energy and low water consumption housing. The study simulates energy storage and calculates the impact of this on the amount and cost of imported electricity. It shows energy storage can be used to reduce the overall cost of imported energy, and that decreasing feed-in tariffs and decreasing the cost of energy storage will lead to an uptake of energy storage systems over the next few years.

In this chapter we first describe the basic concepts of synthetic generation of time series data. We examine several of the realms where it is needed. One area is downscaling from a coarse temporal resolution to a higher frequency. For example, when estimating the performance of solar cells, one-minute data is more useful than hourly, but it may not be available. Also, infilling missing data is crucial if one is forecasting output from a solar farm. On the other hand, generating any number of years of data from a limited number of measured years—known as bootstrapping—is useful for estimating risk for farms and other solar installations. In energy markets, there is scope for generating possible future trajectories for either solar radiation time series or financial instruments that depend on them. In describing these situations and others, we briefly explain the mechanisms for these computations. We end by describing in detail some specific example applications of synthetic solar irradiance tools.

We develop a new probabilistic forecasting method for global horizontal irradiation (GHI) by extending our previous bootstrap method to a case of an exponentially decaying heteroscedastic model for tracking dynamics in solar radiance. Our previous method catered for the global systematic variation in variance of solar radiation, whereas our new method also caters for the local variation in variance. We test the performance of our new probabilistic forecasting method against our old probabilistic forecasting method at three locations: Adelaide, Darwin, and Mildura. These locations are chosen to represent three distinct climates. The prediction interval coverage probability, prediction interval normalized averaged width and Winkler score results from our new probabilistic forecasting method are encouraging. Our new method performs better than our previous method at Adelaide and Mildura, regions with a higher proportion of clear-sky days, whereas our previous method performs better than our new method at Darwin, a region with a lower proportion of clear-sky days. These results suggest that the ideal probabilistic forecasting method might be climate specific.

Thesis (PhD(Mathematics))--University of South Australia, 2017. Includes bibliographical references (pages 199-226) This thesis develops methods for creating renewable energy data with realistic variations at high temporal and spatial resolutions, with sufficient accuracy for detailed system design. We address the need for solar irradiance data with high temporal resolution by developing a method for generating synthetic five-minute pairs of global horizontal irradiance (GHI) and direct normal irradiance (DNI), interpolated from hourly mean values. Our method can be applied to locations where only hourly data is available.

Rooftop solar photovoltaic panels, household electrical energy storage (batteries), home energy management, interval metering and new tariffs will change the way that households use electricity from the grid. Distributed storage can also give electricity retailers the ability to shift loads in response to changes in the wholesale price of electricity and constraints on the distribution network. What is the ideal mix of photovoltaic, storage and tariff for a customer? What is the value of these technologies to customers and to electricity retailers? Refereed/Peer-reviewed

Refereed/Peer-reviewed

The electricity generated from Concentrated Solar Thermal (CST) systems is mostly influenced by the amount of Direct Normal Irradiance (DNI). The majority of DNI values found in both historical and Typical Meteorological Year (TMY) data are hourly means. However, hourly data is too infrequent to capture the transient effects of clouds. High temporal resolution DNI is required to capture these effects on DNI and subsequent influences on CST output. Comparisons between synthetically generated five-minute DNI and observed five-minute DNI are made for Adelaide to determine the performance of this method. A mean bias deviation (MBD) of -0.4% and a normalised root mean square deviation (NRMSD) of 16.3% compares favourably with other methods described in the literature, which have a MBD of 0.3% and NRMSD of 31%. The level of variation in the synthetic compares well to the observed with standard deviations of 390.0 and 391.3 respectively. Of particular interest is how well the model performs when applied to hourly mean DNI for Woomera South Australia. A MBD of -1%, a NRMSD of 15% and standard deviations of 373.5 and 372.4, for the synthetic and observed five-minute DNI respectively, indicate the method is also performing well for Woomera. It is important to note the synthetic DNI for Woomera is derived from a model based on five-minute DNI from Adelaide. The process for generating synthetically interpolated five-minute DNI from hourly DNI is straightforward. Hourly clear sky ratios for each hourly time period are calculated to identify the clear sky index from which 12 five-minute synthetic clear sky ratios are bootstrapped (resampled with replacement) for each hour. The synthetic five-minute ratios are multiplied by the five-minute modelled perfect clear sky DNI to give synthetically interpolated five-minute DNI. This is an important result because it means the method could also be applied to other locations where only hourly DNI exists. Therefore, not only is the temporal resolution of DNI extended to five-minutes, but the spatial coverage of five-minute DNI is also extended. These results are important in determining the role solar energy will play in a 100% renewable energy system. This paper demonstrates a method where five-minute DNI data from one location is used to develop synthetic variations in the hourly data from another location. We use five-minute DNI from Adelaide, South Australia, where real five-minute data is available. Five-minute clear sky ratios (ratio of observed DNI to modelled perfect clear sky DNI) are calculated for each 5-minute time period. Hourly clear sky ratios are calculated for each hourly time period. The hourly ratio identifies a clear sky bin into which the twelve five-minute ratios corresponding to the hourly ratio are placed. The result is clear sky indexes each containing five-minute ratios that capture how five-minute DNI fluctuates for each clear sky index.8 Refereed/Peer-reviewed