Cloud transient characterization in different time steps

AIP Conference Proceedings 1850, 140016 (2017); https://doi.org/10.1063/1.4984524 1850, 140016 © 2017 Author(s). Cloud transient characterization in different time steps Cite as: AIP Conference Proceedings 1850, 140016 (2017); https://doi.org/10.1063/1.4984524 Published Online: 27 June 2017 Miguel Larrañeta, Sara Moreno-Tejera, Isidoro Lillo-Bravo, and Manuel A. Silva-Pérez ARTICLES YOU MAY BE INTERESTED IN Economical and environmental analysis of thermal and photovoltaic solar energy as source of heat for industrial processes AIP Conference Proceedings 1850, 180005 (2017); https://doi.org/10.1063/1.4984572 Increasing the temporal resolution of direct normal solar irradiance forecasted series AIP Conference Proceedings 1850, 140007 (2017); https://doi.org/10.1063/1.4984515 A novel procedure for generating solar irradiance TSYs AIP Conference Proceedings 1850, 140015 (2017); https://doi.org/10.1063/1.4984523 Cloud Transient Characterization in Different Time Steps Miguel Larrañeta1, a), Sara Moreno-Tejera 2), Isidoro Lillo-Bravo2), Manuel A. Silva-Pérez2, b) 1Andalusian Association for Research and Industrial Cooperation (AICIA). Miguel Larrañeta, Andalusian Association for Research and Industrial Cooperation, Camino de los Descubrimientos s/n. 41092, Seville, Spain. Phone/Fax number: (+34)954487237/ (+34)954487233. E-mail: [email protected] 2Department of Energy Engineering, University of Seville, Avd de los Descubrimientos s/n. 41092, Seville, Spain a)Corresponding author: [email protected] b)[email protected] Abstract. In this paper we evaluate the cloud transients by analyzing the dynamics of the direct fraction index kb for one year of solar radiation data in different time steps. We use instant 5-sec data integrated data and compare the number and percentage of occurrences of the different defined sky conditions. We find that the most common situation is a progressive transient and that the average transient lasts between one and 5 minutes. We also perform a cloud transient duration analysis observing that the denser clouds have greater persistence. INTRODUCTION The passage of clouds involves a decrease in the solar radiation that depends on the density of the clouds, implying disturbances in the production of solar thermal systems [1]. When forecasting the solar radiation, the common practice is to predict the cloud shape, base height and position and afterwards estimate its corresponding attenuation [2] generally in an hourly time step. But the cloud passages may last from seconds to hours, therefore this resolution may not be sufficient to simulate transient processes [3]. Figure 1 illustrates the same daily profile but in different resolutions, hourly, 10-min, and 1-min. (a) (b) (c) FIGURE 1. Same daily profile but in different resolutions: hourly (a), 10-min (b) and 1-min (c). DATABASE The data set used for this study corresponds to measurements of Direct Normal Insolation (DNI) for the year 2014 at the location of Seville (Spain). The measurements were taken with a sampling and storing frequency of 0.2 0 100 200 300 400 500 600 700 800 900 1000 01234567891011121314151617181920212223 DirectNormalIrradiance(W/m2) Time HourlyDNIday147GTERstation 0 100 200 300 400 500 600 700 800 900 1000 01234567891011121314151617181920212223 DirectNormalIrradiance(W/m2) Time 10‐minDNIday147GTERstation 0 100 200 300 400 500 600 700 800 900 1000 0 1 2 3 4 5 6 7 8 9 1011121314151617181920212223 DirectNormalIrradiance(W/m2) Time 1‐minDNIday147GTERstation SolarPACES 2016 AIP Conf. Proc. 1850, 140016-1–140016-7; doi: 10.1063/1.4984524 Published by AIP Publishing. 978-0-7354-1522-5/$30.00 140016-1 Hz with a first class Eppley NIP pyrheliometer coupled to a sun tracker Kipp&Zonen 2AP. The devices are located at the meteorological station of the Group of Thermodynamics and Renewable Energy of the University of Seville. Data has been subjected to quality-control procedures following the Baseline Surface Radiation Network (BSRN) recommendations. Besides the instant data, we use six temporal resolutions calculated as the integration of the 5-seg measurements. 1-min, 5-min, 10-min, 15-min, 30-min and 1-h to assess the cloud transients. Data recorded at the sun´s altitude lower than 6º has been removed from the database to avoid the noise that would cause the horizon obstacles. METHODOLOGY We calculate the direct fraction index (kb) in each of the selected time steps dividing the observed DNI by the clear-sky DNI. The focus of our analysis is on the dynamics of the solar radiation and not in the atmospheric turbidity, for this reason, the clear-sky DNI is calculated for each day by empirically fitting a clear sky model -in this case the AB model [4]. The sky condition (related to cloudiness or cloud type) is defined by means of the levels of attenuation of the direct solar radiation reaching the earth surface. Based on [5] we can generalize the type of clouds into 5 groups. The kb threshold selected for performing this classification are presented in table 1. TABLE 1. Classification of sky conditions. Group kb Value G1 kb ≤ 0.2 G2 0.2 < kb ≤ 0.4 G3 0.4 < kb ≤ 0.6 G4 0.6 < kb ≤ 0.8 G5 kb > 0.8 To quantify the cloud transients, we count the number of times that kb varies from a sky condition to another state defined by the threshold values specified in Table 1. We also quantify the duration of the transient by counting the time that the sky condition remains inside those threshold values. The analysis of the results is performed for the 7 posed time resolutions evaluating the differences that entails the use of DNI data in different frequencies. RESULTS For the assessment of the results we present the transition matrices [6]. Each value represent the transition from an initial sky condition (column) to a final sky condition (row). In this manner, values in the diagonal represent the persistence, values over the diagonal represent the entrance of clouds and values above the diagonal represent the exit of clouds. We divide the tables into number of occurrences, and percentage of occurrences where we omit the persistence in order to quantify only the cloud transients. Tables 2-8 present the transition matrixes for the seven evaluated resolutions. TABLE 2. Transition matrix for the instant 5-sec resolution. 5-sec Number of occurrences-Entrance of clouds Percentage of occurrences-Entrance of clouds Group Group G1 G2 G3 G4 G5 G1 G2 G3 G4 G5 Exit of clouds G1 74110789031363451246 11% 2% 1% 0% G2 8919712187352150369611% 9% 2% 1% G3 1355748283901750316442% 10% 10% 2% G4 52915117726143184 91671% 2% 10% 12% G5 209576153894661762772 0% 1% 2% 12% 140016-2 TABLE 3. Transition matrix for the integrated 1-min resolution. 1-min Number of occurrences-Entrance of clouds Percentage of occurrences-Entrance of clouds Group Group G1 G2 G3 G4 G5 G1 G2 G3 G4 G5 Exit of clouds G1 57613 1630 76 22 9 11% 1% 0% 0% G2 1666 6416 1632 100 10 12% 11% 1% 0% G3 99 1650 7167 1700 98 1% 11% 12% 1% G4 17 129 1748 11932 1789 0% 1% 12% 12% G5 11 20 101 1861 142635 0% 0% 1% 13% TABLE 4. Transition matrix for the integrated 5-min resolution. 5-min Number of occurrences-Entrance of clouds Percentage of occurrences-Entrance of clouds Group Group G1 G2 G3 G4 G5 G1 G2 G3 G4 G5 Exit of clouds G1 10686 641 228 95 52 10% 3% 1% 1% G2 643 724 396 215 123 10% 6% 3% 2% G3 246 404 877 469 239 4% 6% 7% 4% G4 164 228 481 1644 741 2% 3% 7% 11% G5 57 129 264 828 27450 1% 2% 4% 12% TABLE 5. Transition matrix for the integrated 10-min resolution. 10-min Number of occurrences-Entrance of clouds Percentage of occurrences-Entrance of clouds Group Group G1 G2 G3 G4 G5 G1 G2 G3 G4 G5 Exit of clouds G1 5104 350 190 85 47 8% 4% 2% 1% G2 415 277 202 131 73 10% 5% 3% 2% G3 216 226 377 257 173 5% 5% 6% 4% G4 100 142 299 708 436 2% 3% 7% 10% G5 88 109 191 515 13296 2% 3% 4% 12% 140016-3 TABLE 6. Transition matrix for the integrated 15-min resolution. 15-min Number of occurrences-Entrance of clouds Percentage of occurrences-Entrance of clouds Group Group G1 G2 G3 G4 G5 G1 G2 G3 G4 G5 Exit of clouds G1 3309 254 117 81 38 8% 4% 2% 1% G2 327 163 154 85 56 10% 5% 3% 2% G3 179 175 240 176 123 6% 5% 5% 4% G4 96 122 234 441 306 3% 4% 7% 9% G5 63 90 158 414 8622 2% 3% 5% 13% TABLE 7. Transition matrix for the integrated 30-min resolution. 30-min Number of occurrences-Entrance of clouds Percentage of occurrences-Entrance of clouds Group Group G1 G2 G3 G4 G5 G1 G2 G3 G4 G5 Exit of clouds G1 1568 144 64 34 24 7% 3% 2% 1% G2 206 89 71 59 40 10% 4% 3% 2% G3 156 77 123 107 55 8% 4% 5% 3% G4 106 90 121 194 159 5% 4% 6% 8% G5 33 80 131 249 4027 2% 4% 7% 12% TABLE 8. Transition matrix for the integrated hourly resolution. 1-h Number of occurrences-Entrance of clouds Percentage of occurrences-Entrance of clouds Group Group G1 G2 G3 G4 G5 G1 G2 G3 G4 G5 Exit of clouds G1 688 70 49 18 11 6% 4% 2% 1% G2 100 44 39 42 22 8% 3% 4% 2% G3 79 51 64 57 48 7% 4% 5% 4% G4 57 41 67 90 101 5% 3% 6% 8% G5 80 42 90 128 1923 7% 4% 8% 11% Certain symmetry between values above and below the diagonal is observed. The values closer to the diagonal are the most repeated, i.e., transitions are progressive. As the time step increases, this trend is reduced. In figure 2 we present the number of occurrences of progressive transitions versus de time step. 140016-4 FIGURE 2. Number of occurrences of the progressive transients depending on the time step. We can observe that there is a turning point in the number of occurrences of progressive transients (PTs) between one minute and five minutes. This same conclusion can be observed on table 9. Aiming to compute the entrance of the clouds, we count the number of occurrences that the sky condition changes from clear ((kb > 0.8 - G5) to any of the other defined states. TABLE 9. Transition from G5 to the rest of the groups. Number of occurrences Percentage of occurrences Group TOTAL Group G1 G2 G3 G4 G1 G2 G3 G4 Time Step 5-seg 246 696 1644 9167 11753 2% 6% 14% 78% 1-min 9 10 98 1789 1906 0% 1% 5% 94% 5-min 52 123 239 741 1155 5% 11% 21% 64% 10-min 47 73 173 436 729 6% 10% 24% 60% 15-min 38 56 123 306 523 7% 11% 24% 59% 30-min 24 40 55 159 278 9% 14% 20% 57% 1h 11 22 48 101 182 6% 12% 26% 55% Table 9 shows that as we increase the time step there is a shift towards a reduction in the percentage of progressive transitions and an increase in the steeper transitions. This can be explained because in a larger time step, more situations are integrated into a single value, averaging the result of several phenomena. We also quantify the duration of the transients. For that purpose, we count the time that the transient remains in each group. The starting point is once again the clear sky condition ((kb > 0.8 - G5). For this analysis we use the integrated data into the six selected resolutions. Figures 3-6 present the number and percentage of occurrences of the transient duration for each group. 5‐seg 1‐min 5‐min 10‐min 15‐min 30‐min 1‐h 0 10000 20000 30000 40000 50000 60000 70000 0 500 1000 1500 2000 2500 3000 3500 4000 NumberofOccurrences Time(s) Progresivetransientsindifferenttimesteps 140016-5 (a) (b) FIGURE 3. Number of occurrences (a) and percentage of occurrences (b) of the transient duration from G5 to G1. (a) (b) FIGURE 4. Number of occurrences (a) and percentage of occurrences (b) of the transient duration from G5 to G2. (a) (b) FIGURE 5. Number of occurrences (a) and percentage of occurrences (b) of the transient duration from G5 to G3. 0 10 20 30 40 50 60 1-min 5-min 10-min 15-min 30-min 1h NUMBER OF OCCURRENCES TIME STEP GROUP 1 TRANSIENT DURATION ≥5-min ≥10-min ≥30-min ≥1h ≥ 2 h 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 1-min 5-min 10-min 15-min 30-min 1h PERCENTAGE OF OCCURRENCES TIME STEP GROUP 1 TRANSIENT DURATION ≥5-min ≥10-min ≥30-min ≥1h ≥ 2 h 0 20 40 60 80 100 120 140 1-min 5-min 10-min 15-min 30-min 1h NUMBER OF OCCURRENCES TIME STEP GROUP 2 TRANSIENT DURATION ≥5-min ≥10-min ≥30-min ≥1h ≥ 2 h 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 1-min 5-min 10-min 15-min 30-min 1h PERCENTAGE OF OCCURRENCES TIME STEP GROUP 2 TRANSIENT DURATION ≥5-min ≥10-min ≥30-min ≥1h ≥ 2 h 0 50 100 150 200 250 300 1-min 5-min 10-min 15-min 30-min 1h NUMBER OF OCCURRENCES TIME STEP GROUP 3 TRANSIENT DURATION ≥5-min ≥10-min ≥30-min ≥1h ≥ 2 h 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 1-min 5-min 10-min 15-min 30-min 1h PERCENTAGE OF OCCURRENCES TIME STEP GROUP 3 TRANSIENT DURATION ≥5-min ≥10-min ≥30-min ≥1h ≥ 2 h 140016-6 (a) (b) FIGURE 6. Number of occurrences (a) and percentage of occurrences (b) of the transient duration from G5 to G4. It is noted that the transients that last longer in all resolutions are the ones defined in group 1. SUMMARY AND CONCLUSIONS From the transient analysis in different time steps we have observed that the most common occurrence is a progressive transient. And that the use of large time steps entails a great loss of information. We define an average value of entrance and exit of clouds of 3 minutes by identifying a turning point in the overall number of occurrences of the progressive transients between the 1-min and 5-min resolutions and by observing that most of the cloud passages last for more than five minutes. Future works will require a more complex definition of the cloud types with the help of sky-camera images that could lead to a better characterization of the transients. REFERENCES 1. Edward W. Law, Abhnil A. Prasad, Merlinde Kay, Robert A..Taylor. 2014 Direct normal irradiance forecasting and its application to concentrated solar thermal output forecasting – A review. Solar Energy 108, 287–307 2. Perez, R., Moore, K., Wilcox, S., Renne, D., Zelenka, A., 2007. Forecasting solar radiation – preliminary evaluation of an approach based upon the national forecast database. Solar Energy 81, 809–812. 3. Meyer, R., Beyer, H.G., Fanslau, J., Geuder, N., Hammer, A., Hirsch, T., Hoyer-Click, C., Schmidt, N., Schwandt, M., 2009. Towards standardization of CSP yield assessments. In: Proceedings of the SolarPACES Conference, Berlin (Germany). 4. Silva-Pérez, M.A., 2002. Estimación del recurso solar para sistemas termosolares de concentración. Ph. D. Thesis. University of Seville. 5. M. Martínez-Chico, F.J. Batlles, J.L. Bosch., 2011. Cloud classification in a Mediterranean location using radiation data and sky images. Energy 36, 4055-4062. 6. R.J. Aguiar, M. Collares-Pereira, J.P. Conde., 1998. Simple procedure for generating sequences of daily radiation values using a library of Markov transition matrices. Solar Energy, Volume 40, Issue 3, 1988, Pag. 269-279. 0 200 400 600 800 1000 1200 1400 1600 1800 1-min 5-min 10-min 15-min 30-min 1h NUMBER OF OCCURRENCES TIME STEP GROUP 4 TRANSIENT DURATION ≥5-min ≥10-min ≥30-min ≥1h ≥ 2 h 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 1-min 5-min 10-min 15-min 30-min 1h PERCENTAGE OF OCCURRENCES TIME STEP GROUP 4 TRANSIENT DURATION ≥5-min ≥10-min ≥30-min ≥1h ≥ 2 h 140016-7