No abstract

No abstract

This article addresses a critical issue on the robust unit commitment (RUC), which is the construction of an accurate and reliable uncertainty set for solar photovoltaic (PV) systems. The classic robust unit commitment (CRUC) provides an inaccurate and unreliable model for the highly intermittent solar generations. However, the authors' proposed uncertainty set utilises the forecast models that rely on the clear-sky and overcast solar forecasts, which are more accurate and reliable. The uncertainty set for solar generations is constructed based on the type of day, levels of daily uncertainty index (DUI) and daily energy index (DEI), and the uncertainty level of solar ramps. The test results on the IEEE 118-bus test system demonstrate that (i) using DEI and DUI reliably and efficiently manages and reduces the cost for highly uncertain and overcast day compared to the CRUC and (ii) the RUC cost is much sensitive to the selected level of DUI and DEI than the uncertainty level of the solar ramps.

Variability in the output of photovoltaic (PV) power plants continues to hinder their interconnection to power systems. Consequently, available PV plant data should be carefully analysed so that appropriate strategies can be developed to mitigate effects associated with variability in PV output. This paper constructively illustrates the characterisation of the variability and evaluates performance of the 1.2MW PV plant at The University of Queensland. The characterisation is based on some of the recently proposed indices and utilises I-minute resolution data captured over a 2I-month period. Keywords-Variability; photovoltaic power plant; indices;I.[NTRODUCTION Photovoltaic (PV) capacity is increasing on utility systems, As a result, potential impacts of PV output variability caused by transient clouds concerns utility planners and grid operators, Utilities and control system operators need to adapt day ahead scheduling, load following, and second to minute's regulation to accommodate this variability while at the same time maintaining existing standards of reliability, So far, there are no standard indices to characterise variability and performance of PV power plants.The existing standard for reporting individual power plant performance, [EEE Std. 762-2006, defines reliability, availability, and productivity indices to report conventional controllable power plant performance [[]. The standard assumes fuel is available and indicates to what extent the rest of the plant performs relative to its rating and availability. In contrast to conventional plants, for PV plants, performance indices must consider variability of fuel, input. This contrast between traditional dispatchable generation and emerging variable PVs, has limited incorporating solar plants into a conventional generation fleet. Therefore, indices compatible with the traditional generation are needed allowing easy comparison between new generation options. When indices are defined and adopted into standards such as IEEE, appraisals of bringing solar PV into a traditional generation fleet can be more reliable. Authors in [2] have proposed indices with generalisation capabilities to evaluate the performance and quantify output variability for any given PV plant.The objective of this paper is to provide insight into the performance of The University of Queensland (UQ) PV system. The data available at two UQ PV sites is used to evaluate the PV performance and output variability indices introduced in [2]. Daily variability and daily clearness indices are utilised to categorise days based on the level of variability. Performance of the two sites is evaluated in terms of the capacity factor and the energy performance factor. The season-wise variability corresponding to the one-minute power ramp rate is characterised in terms of probability distribution functions. The three-sigma rule is used as a confidence interval for PV ramp rates because it covers 99.73% of the PV output changes. Based on this analysis, guidelines are developed to establish the required lev...