BMe Research Grant
The MTA-BME FASTER Lendület Research Group, led by Dr. Bálint Hartmann, aims to increase the observability of the extensive, low instrumentation low-voltage power grid. The focus of the research is on the development of algorithms and procedures that can detect abnormal events (e.g. local outages), increase the flexibility of the grid operation, and prepare it for emerging disruptive technologies (e.g. e-mobility).
Figure 2: Location of the OMSZ meteorological stations used in the research (red) and the geographic position of one of the household-scale PV generators used for validation (green)
Figure 3: Hovmöller diagrams for days with different cloud cover (clear sky, partly cloudy, overcast - bottom row from left to right) and for the whole dataset (top middle)
2. Determination of critical geographical distance
The same Hovmöller diagram gives an illustrative picture of the relationships between station measurements. However, it is not necessary to use all measurement data to estimate the solar panel at a given geographic point. Therefore, the goal of the second step is to define a critical geographic radius within which the use of measurement points based on historical data is sufficiently correlated to be used effectively. So essentially the scatter of the distributions shown in the Hovmöller diagram is shown in Figure 4 for all measurement station pairings.
Figure 4: Variation of the dispersion of the time delay distribution curve by station pairs
Figure 4 already provides clear evidence that there is a clear break in the magnitude of the standard deviations as a function of station distances. Based on this, only data within a 20 km radius of each solar producer are used.
3. Link between past and future – using historical data results to increase the spatio-temporal resolution of forecasts
The statistical relationships obtained in the previous steps were used to refine the global irradiance estimate produced by the numerical forecast. The prediction model defines values at the corners of a grid (red points), while the producer under study is located at the green point in Figure 5. For this geographic point, kriging is used to determine the expected global radiation magnitude using a linear combination of the estimated values within a 20km radius of the nearby point. This global irradiance value is finally converted to real production using a virtual solar model.
Figure 5: Illustration of numerical weather model results (red) and the position of one of the photovoltaic generators used for validation
I validated the results of the presented model with 2 months of production data from a specific small solar power plant. The presented model is simultaneously able to consider the effects of extreme cloud shifts (which most estimators are unable to do due to lack of information) and, using kriging, it is also suitable for estimation at any geographical point, i.e. for providing a regional level estimation. Based on the run results so far, the accuracy of the model significantly exceeds the accuracy of the reference forecasting models (Neural Network, ARIMA, and Persistence model).
B. Sinkovics: A napelemek termelésének előrejelzése a múlt és a jelen tükrében, Élet és Tudomány, 25, (2022)
I. Táczi, B. Sinkovics, I. Vokony, B. Hartmann: The Challenges of Low Voltage Distribution System State Estimation – An Application Oriented Review. Energies (2021), 14, 5363
B. Sinkovics, J. Kiss, B. Polgári, J. Csatár: Co‐simulation framework for calculating balancing energy needs of a microgrid with renewable energy penetration, International Journal of Energy Research, pp. 18631–18643, vol. 45 (2021)
B. Sinkovics, B. Hartmann: Analysing Effect of Solar Photovoltaic Production on Load Curves and their Forecasting, Renewable Energy and Power Quality Journal, pp. 760–765. , 6 p. (2018)
B. Sinkovics, I. Táczi, I. Vokony, B. Hartmann: A novel adaptive day-ahead load forecast method, incorporating non-metered distributed generation: a comparison of selected European countries, Mathematical Modelling of Contemporary Electricity Markets (MMCEM), pp. 62–81, Chapter 3, Elsevier (2020)
B. Hartmann, A. Kazsoki, V. Sugár, B. Sinkovics: Napsugárzás-mintázat kategorizálási módszereinek kritikai szemléletű összehasonlítása, Magyar Energetika pp. 18–24., (2020)
B. Sinkovics: Fogyasztás és termelés egyensúlya: A jövő megoldása az energiaéhségre? Élet és Tudomány, pp. 233–235, (2019)
B. Sinkovics: Fotovillamos termelésbecslő modell bemutatása térbeli-időbeli krigelés alkalmazásával, MEE Vándorgyűlés (2021)
I. Táczi, B. Sinkovics, I. Vokony, B. Hartmann: Conceptual Analysis of Distribution System State Estimation of Low Voltage Networks, International Conference on Renewable Energy and Power Quality (ICREPQ 2020)
B. Sinkovics: Computer-Based Algorithm to Predict the Change in Daily Electricity Demand, Advanced ICT Tools and Methods for Cyber-Physical Systems and Biomedical Applications, Esztergom, Hungary (2019)
B. Sinkovics, B. Hartmann: Analysing Effect of Solar Photovoltaic Production on Load Curves and their Forecasting, International Conference on Renewable Energy and Power Quality: ICREPQ'18 Vigo, Spain: Europe-an Association for the Development of Renewable Energies, Environment and Power Quality (EA4EPQ), pp. 1–6. (2018)
D. Yang, Z. Dong, T. Reindl, P. Jirutitijaroen, W. M. Walsh, Solar irradiance forecasting using spatio-temporal empirical kriging and vector autoregressive models with parameter shrinkage, Solar Energy, vol. 103, pp. 550–562, 2014
D. Yang, Z. Dong, A. Nobre, Y. S. Khoo, P. Jirutitijaroen, W. M. Walsh, Evaluation of transposition and decomposition models for converting global solar irradiance from tilted surface to horizontal in tropical regions, Solar Energy, vol. 97, pp. 369–387, 2013