Impacts of different data averaging times on statistical analysis of distributed domestic photovoltaic systems

• Published in: Solar energy Vol. 84; no. 3; pp. 492 - 500
• Main Authors: Widén, Joakim, Wäckelgård, Ewa, Paatero, Jukka, Lund, Peter
• Format: Journal Article
• Language: English
• Published: Kidlington Elsevier Ltd 01-03-2010; Elsevier; Pergamon Press Inc
• Subjects: Applied sciences; AUGMENTATION; DISPERSED STORAGE AND GENERATION; DISTRIBUTION; Distribution grid; Domestic electricity demand; ELECTRIC POTENTIAL; ELECTRIC POWER; Electricity; Energy; Equipments, installations and applications; Exact sciences and technology; HOUSEHOLDS; Impact analysis; Low voltage; Natural energy; Photovoltaic cells; Photovoltaic conversion; PHOTOVOLTAIC POWER SUPPLIES; Photovoltaics; POWER DEMAND; Power flow; POWER SYSTEMS; POWER TRANSMISSION AND DISTRIBUTION; RANDOMNESS; SIMULATION; Solar energy; Statistical analysis; STATISTICS; TECHNOLOGY; TEKNIKVETENSKAP; Time averaging; TIME RESOLUTION; VARIATIONS; Low voltage; Time averaging; Power flow; Distribution grid; Photovoltaics; Domestic electricity demand; High resolution; Statistical analysis; Statistical distribution; Photovoltaic system; Distributed system; Statistics; Time resolution; Simulation; Output voltage; Time analysis
• ISSN: 0038-092X; 1471-1257; 1471-1257
• DOI: 10.1016/j.solener.2010.01.011

The trend of increasing application of distributed generation with solar photovoltaics (PV-DG) suggests that a widespread integration in existing low-voltage (LV) grids is possible in the future. With massive integration in LV grids, a major concern is the possible negative impacts of excess power injection from on-site generation. For power-flow simulations of such grid impacts, an important consideration is the time resolution of demand and generation data. This paper investigates the impact of time averaging on high-resolution data series of domestic electricity demand and PV-DG output and on voltages in a simulated LV grid. Effects of 10-minutely and hourly averaging on descriptive statistics and duration curves were determined. Although time averaging has a considerable impact on statistical properties of the demand in individual households, the impact is smaller on aggregate demand, already smoothed from random coincidence, and on PV-DG output. Consequently, the statistical distribution of simulated grid voltages was also robust against time averaging. The overall judgement is that statistical investigation of voltage variations in the presence of PV-DG does not require higher resolution than hourly.