quantifying variability for grid connected photovoltaics in the tropics for microgrid application

doi: 10.1016/j.egypro.2016.11.306 Energy Procedia 103 2016 400 – 405 ScienceDirect Applied Energy Symposium and Forum, REM2016: Renewable Energy Integration with Mini/Microgrid, 19

1876-6102 © 2016 The Authors Published by Elsevier Ltd This is an open access article under the CC BY-NC-ND license

( http://creativecommons.org/licenses/by-nc-nd/4.0/ ).

Peer-review under responsibility of the scientific committee of the Applied Energy Symposium and Forum, REM2016: Renewable Energy Integration with Mini/Microgrid.

doi: 10.1016/j.egypro.2016.11.306

Energy Procedia 103 ( 2016 ) 400 – 405

ScienceDirect

Applied Energy Symposium and Forum, REM2016: Renewable Energy Integration with

Mini/Microgrid, 19-21 April 2016, Maldives Quantifying variability for grid-connected photovoltaics in the

tropics for microgrid application

a Centre of Electrical and Energy Systems(CEES), Faculty of Electrical Engineering, Universiti Teknologi Malaysia, 81310, Johor

Bahru, Johor, Malaysia

b Faculty of Electrical Engineering, Universiti Teknikal Malaysia Melaka, Hang Tuah Jaya, 76100 Durian Tunggal, Melaka

Abstract

This paper quantifies the variability of solar resource for PV systems located in the tropical climate Global horizontal irradiance data is obtained from a site in Melaka, Malaysia with one-minute-average values for a full year Clearness index and variability index are used together with a clear sky model for tropical location to classify the days according to its fluctuation profile The results show that significant amount of variability occur throughout the year for this site while days with clear skies are almost negligible PV can potentially supply part of the loads during the day but precise planning on storage and backup generators is crucial for independent microgrid operation

© 2016 The Authors Published by Elsevier Ltd

Selection and/or peer-review under responsibility of REM2016

Keywords: grid-connected photovoltaics; solar variability; microgrid; tropical climate.

1 Introduction

Countries located in the equatorial region with tropical climate enjoy high amounts of daily sunshine throughout the year as it experiences almost no seasonal variation This pose an advantage in continually harvesting and utilizing solar energy Another advantage is that the peak power produced from PV

* Corresponding author Tel.:+607-5557004; Fax:+607-5557005

E-mail address: hasimah@fke.utm.my

© 2016 The Authors Published by Elsevier Ltd This is an open access article under the CC BY-NC-ND license

( http://creativecommons.org/licenses/by-nc-nd/4.0/ ).

Peer-review under responsibility of the scientific committee of the Applied Energy Symposium and Forum,

REM2016: Renewable Energy Integration with Mini/Microgrid.

systems generally correlate with the industrial and commercial peak demand Effective and economical measure to tap this resource can greatly enhance the proliferation of grid-connected PV systems

However, variability of solar resource is a crucial element that needs to be considered Solar resource variability is mainly caused by two factors, the sun’s daily movement and cloud cover The former factor

is deterministic in the sense that it can be calculated with very high precision However, the latter is stochastic and is known to cause significant ramps in solar irradiance and PV power output [1]

Variability profiles are location dependent This is because each location is subjected to different climates and local weather conditions Even places situated within a few kilometres of each other can have different weathers at the same time due to different microclimates Therefore it is vital that studies

on the variability characteristics are conducted based on the location where PV systems will be utilised [2]

The objective of this paper is to quantify the solar variability of a single-site PV system in a tropical climate This information is useful, particularly for microgrid operation where power output smoothing through geographical dispersion of PV systems is not possible The quantification involves determining the categories of variability happening throughout a year, its severity and respective timeline

2 Methodology

2.1 Data

The measured Global Horizontal Irradiance (GHI) values used in this paper are obtained from a weather station located on the rooftop of the Faculty of Electrical Engineering, Universiti Teknikal Malaysia Melaka (2.32ºN, 102.3ºE) The weather station is equipped with two pairs of Kipp and Zonen’s CMP 11 pyranometers located approximately 40 meters apart that measure horizontal and plane of array (POA) irradiance (15º tilt, facing South) Data from one of the GHI pyranometer is selected GHI is chosen over POA irradiance since the former is more generic To convert to equivalent POA values, transposition models can be applied with sufficient accuracy The data is taken every second and stored in

a main and backup server For this paper, the data are averaged every minute and the duration is from 1 January 2015 to 31 December 2015, a full year to account for seasonal variation Only three days are omitted from this dataset (11, 14 and 17 June) due to missing and incomplete data

2.2 Solar resource evaluation

In order to quantify the solar resource, two indices are used to classify the type of days that occur the

daily Clearness Index, CI [3] and daily Variability Index, VI from Stein, Hansen and Reno [4] CI is

defined as the ratio of the measured irradiance at the earth surface with the horizontal extraterrestrial radiation reaching the outer surface of the atmosphere and can be written as:

) ( cos ) (

0 t t

E

I

G

CI

i o

meas

with I 0= 1367 W/m2,the solar constant, E 0the earth eccentricity factor and Tis the incidence angle which is latitude dependent In essence, it is a ratio of irradiance with losses and ideal irradiance Typical values range from 0.2 during overcast condition to 0.7 during clear skies

VI is a relatively new index that is introduced as a simple measure of irradiance variability over a time

period and can be written mathematically as

¦

¦





'





'





n

n

t CSI

CSI

t GHI

GHI VI

2

2 2 1 2

2 2 1

(2)

It is essentially a ratio of the characteristic of measured GHI plotted against time over the characteristics of calculated CSI for the same time interval Typical VI values range from 1 to 30, with higher values implying higher amount of fluctuations Combining the use of CI and VI help understand the characteristic nature of variability of a location

The clear sky model used here is based on Yang, Walsh and Jirutitijaroen [5] which is specifically developed for Singapore based on the Adnot model and is approximated for this location Mathematically, it is written as:

I E

G c sc Tz 1 3585 0 00135u90T

0 cos 8298

3 Results and Discussion

3.1 Solar resource variability

Fig 1 Clearness Index (CI) vs Variability Index (VI) Fig.1 shows the relationship between the variability index with the corresponding clearness index for each day, represented by the orange dots It can be readily seen that there is no linear correlation between the two indices Ideally, we would want to have a high number of days with clear skies and low variability However, this is rarely obtained Throughout the whole year, it only occurs for two days, highlighted within the green circle A majority of the data falls between VI values of 5 and 25 and CI values of 0.3 to 0.7 It can also be observed that for eight days, the minute-by-minute variability is very high with VI values exceeding 25, as denoted by the red circle The trend from the scatter plot suggests that the effect of transient cloud movement plays a sizable role in introducing variability Overcast sky conditions were experienced for a total of eight days, most occurring between August to mid-October

Fig 2 show sampling profiles for each day type The horizontal axis representing the timeline between sunrise and sunset is depicted in minutes Data is recorded for a total of 720 minutes or 12 hours daily because of the length of daytime is approximately the same throughout the year, with a maximum of 16 minutes from the longest and shortest sunshine duration

(a) (b) (c)

Fig 2 (a) Sample of clear day (14/03/15); (b) Sample of high variability day (30/03/15); (c) Sample of overcast day (06/08/15)

3.2 Variability severity

For days with significant variability, what we are most concerned with is the variability magnitude, time of occurrence, and duration In order to show how the analysis is done, the day with high variability shown in Fig 2(b) is taken as an example To record the rate of change between each minute, the value of GHI for one minute is subtracted against its preceding minute The result is shown in Fig 3

(a) (b)

Fig.3 (a) Actual fluctuation profile (b) Absolute fluctuation profile Fig.3 (a) and (b) are complimentary as both show the same fluctuation profile To identify the ramping

up and down, the actual fluctuation profile can be used, whereas to determine the maximum magnitude of fluctuation, absolute values can be utilized These graphs provide the information needed to address our concerns From these graphs, the severity and timeline of variability can be readily assessed For example, the highest fluctuation in terms of magnitude occurred between minutes 360 and 420, which correspond

to a time between 1 pm to 2 pm The highest minute-to-minute ramp rate recorded is 804 W/m2

Fig.4 Fluctuation magnitude rearranged in descending order Fig 4 is a rearranged variability magnitude for each data point from the highest to the lowest value and this is done for each day Several important features can be extracted from this type of plot First, the

variability duration throughout the day can be determined by looking at the slope of the line Next, we can identify the maximum fluctuation magnitude (P1) We can also identify the amount of time the fluctuation rate exceeds a certain value In this figure, it can be seen that the total variability exceeded 400 W/m2for a total of 60 minutes throughout this day (P2) and exceeds 200 W/m2for a total of 140 minutes (P3)

Fig 5 Sample monthly fluctuation magnitude for March 2015 Fig 5 shows a sample of compiled fluctuation magnitude for one month This method follow similar works done in [6] but the analysis here uses values in minutes rather than percentage as the result is more intuitive and easy to understand The lines in the graph represent the variation in fluctuation nature for each day Days with low variability will have a short tail and approaches the horizontal axis quickly On the other hand, days with high variability has a long tail that can go up to more than 300 minutes of 5 hours cumulatively throughout a whole day Substantial amount of variability can occur for more than 2 hours on a day with high variability

4 Conclusion

The purpose of the current study was to quantify the solar variability for a single-site in a tropical climate In summary, based on the site’s data, tropical climates can potentially experience significant variability throughout the year The number of days with clear skies with low variability is negligible The magnitude and duration of the fluctuation is also relatively high Through minute-by minute analysis, these parameters can be identified accurately These fluctuations can impact the stability severely during autonomous PV microgrid system operation Therefore, careful consideration is required to incorporate energy storage and backup generators for a sustainable microgrid operation

Acknowledgements

The main author would like to express gratitude to the Ministry of Higher Education Malaysia, Universiti Teknologi Malaysia and Universiti Teknikal Malaysia Melaka for supporting his

research work This research is partially funded by the Research Acculturation Grant Scheme RAGS/1/2015/TK0/FKE/02/B00093

References

[1] A Mills, “Understanding variability and uncertainty of photovoltaics for integration with the electric power system,”

Lawrence Berkeley National Laboratory, 2010.

[2] C Trueblood, S Coley, T Key, L Rogers, A Ellis, C Hansen, and E Philpot, “PV measures up for fleet duty: Data from a

tennessee plant are used to illustrate metrics that characterize plant performance,” Power and Energy Magazine, IEEE, vol

11, no 2, pp 33–44, 2013.

[3] A Woyte, R Belmans, and J Nijs, “Fluctuations in instantaneous clearness index: Analysis and statistics,” Solar Energy,

vol 81, no 2, pp 195–206, 2007.

[4] J S Stein, C W Hansen, and M J Reno, “The variability index: a new and novel metric for quantifying irradiance and PV

output variability,” in World Renewable Energy Forum, Denver, CO, 2012.

[5] D Yang, W M Walsh, and P Jirutitijaroen, “Estimation and applications of clear sky global horizontal irradiance at the

equator,” Journal of Solar Energy Engineering, vol 136, no 3, p 034505, 2014.

[6] M Lave, M J Reno, and R J Broderick, “Characterizing local high-frequency solar variability and its impact to distribution

studies,” Solar Energy, vol 118, pp 327–337, 2015.

Biography

Hasimah Abdul Rahman (PhD) is an associate professor at Universiti Teknologi Malaysia and the Deputy Director of the Centre of Electrical and Energy Systems (CEES) Her research interests are on renewable energy with emphasis on photovoltaic systems and grid integration

Fig Clearness Index (CI) vs Variability Index (VI) Fig.1 shows the relationship between the variability index with the corresponding clearness index for. .. utilized These graphs provide the information needed to address our concerns From these graphs, the severity and timeline of variability can be readily assessed For example, the highest fluctuation in. .. observed that for eight days, the minute-by-minute variability is very high with VI values exceeding 25, as denoted by the red circle The trend from the scatter plot suggests that the effect of