Assessing the Complementarity of Wind and Solar Energy in Kentucky

Abstract

Recently, there has been a push by countries to diversify their energy mix considering various factors. In this regard, there have been several studies conducted to assess the potential for using sources such as wind and solar to generate supplemental energy to the already present energy generation setup. In this regard, this study explores the potential of wind for the Commonwealth of Kentucky. To perform this study, wind data were sourced for eight locations across Kentucky from the publicly accessible wind speed information present at Weather Underground for the years 2020–2021 (two years). An analysis was performed concerning the seasonal, monthly, and hourly variation in the wind speed so as to identify the expected times of sufficient wind energy generation. Moreover, a comparison of the collected data was performed with data from a home-based weather station as well as a deployed wind turbine to validate the variation pattern of the publicly sourced data. Finally, in order to investigate the variation patterns of wind and solar energy sources, a comparative analysis was also performed using data from a solar power generation plant in Kentucky. A seasonal and monthly complementarity was observed between the wind and solar energy. However, when considering daily patterns, the wind was found to follow solar generation with an offset. While further research is required, this analysis indicates that it is possible to deploy wind energy power generation projects in the Commonwealth of Kentucky. The seasonal complementary behavior of wind and solar energy can be used along with battery storage in conjunction with natural gas to provide a diversified electricity generation portfolio.

1. Introduction

The power requirements of countries around the world are increasing with the growth in the human population. The population of the world is expected to increase to 8.5 billion people by 2030, and will result in an increase in world energy demand of 21% as predicted by the International Energy Agency [1]. Moreover, the electricity consumption for North America is projected to increase from 4222 TWh in 2020 to 5687 TWh in 2050 [2]. Such increases in electric power requirements have been observed in developing countries as well. Therefore, countries around the world have become interested in diversifying their energy mix to cater to growing energy needs.

Solar and wind generation systems for electricity can work as part of the grid or in a standalone arrangement and can be used in both rural as well as urban settings [3]. It therefore has been the focus of multiple studies around the world [4,5,6,7,8,9,10,11,12].

Carvajal-Romo et al. [4] studied wind and solar energy generation potential in Columbia. Çakmakçı and Huner [5] provide an analysis on the wind speed, direction, and power potential for a site in Turkey. Wind power potential analysis has been performed by Bououbeid et al. in [10] at two separate sites in Mauritania and their analysis shows promising potential of wind power generation in Mauritania. Zhang et al. [11] discuss the potential of wind in New Zealand and find that small scale wind turbines have sufficient power generation potential in the long term.

In [9], Slusarewicz and Cohan assessed wind and solar energy complementarity in Texas. They used wind and solar data from different sites across the state of Texas and found that there was appreciable complementarity on an annual and daily level. Moreover, they found complementarity between West and South Texas. Overall, they conclude that combined together, wind and solar power could be used to develop a reliable power energy mix. Another study performed in California by Mahmud et al. [8,13] studied the variability of winter and summer dominant onshore wind. They found that most of the state of California exhibits strong wind energy potential in the summer compared with the winter. This could be exploited for a more reliable grid. More work in Asia was performed by [6], who studied wind speed data considering annual and monthly variability for seven locations within the Al-Baha region of Saudi Arabia. They found that the highest wind speeds were in the months between January and March with one location showing a high in July. For a deployed wind turbine, they found that the highest performance was observed in the winter and summer time. A study conducted by Jurasz et al. [12] assessing wind and solar energy complementarity in North Africa found that there is variation in solar energy and wind, which allows for resilience to energy droughts in North Africa. Moreover, considering the North Atlantic Oscillation (NAO), which indicates atmospheric pressure change patterns in the North Atlantic, the authors found seasonal variations in the correlations for wind and solar energy with NAO with regard to the geographical region being considered. Makhubele and Folly in [14] investigated solar and wind energy complementarity in South Africa. Their analysis concluded that there is a strong daily complementarity between wind and solar energy, but a weaker relationship was observed on a yearly basis. Another study that examined variability between wind and solar energy was performed by Prol et al. [15]. In their study, the authors found that within different regions of Europe, a complementarity exists between wind and solar energy on a monthly and daily basis with a reduction occurring when considering hourly variation. In [7], the authors studied solar and wind power potential for different provinces in Nepal. They concluded that the power generating potential provided by wind and solar energy can be a substantial benefit in providing energy to remote areas not serviced by the current grid, especially in the upper mid-northern and the eastern lowlands of the country. As observed from the coverage of previous studies, while there have been numerous initiatives taking place in different parts of the world to understand the variation in the behavior of wind and solar energy, there is a gap present regarding such investigations taking place for the Commonwealth of Kentucky. This study therefore attempts to help mitigate this issue by contributing to the area.

Aims and Objectives

Following from previous efforts in different parts of the world and understanding that there is a spectrum of needs in terms of diversifying energy generation sources, this study highlights the case of wind in the Commonwealth of Kentucky. Moreover, considering that renewables such as wind and solar energy do not provide a consistent amount of power throughout the day/year, the current study’s aims and objectives are as follows.

• Analyze wind data collected from Weather Underground to explore variation patterns of wind in eight locations in Kentucky.
• Analyze the variation pattern of solar energy data acquired from a deployed solar power generation plant.
• Compare the variation patterns of wind and solar energy together to investigate any complementarity that exists between them.
• Compare the assessment of wind variation patterns from data collected from a deployed wind turbine and solar generation scheme.

Through these aims and objectives, the study will benefit power companies and government agencies planning to deploy wind- and solar-based energy systems in the state of Kentucky by providing insights into the variation of wind and solar energy in the state. The methodology of the current work is illustrated graphically in Figure 1. It can be observed that using a variety of sources of wind and solar data, analyses at different granular levels have been provided covering seasonal, monthly, hourly, and minute-wise variations of the two sources to provide an understanding of the variation pattern.

We start by providing an introduction to the current state of renewable energy generation, focusing on the wind in Section 2. Section 3 presents the data collection scheme used for gathering wind data from different locations in Kentucky. An analysis of wind speeds at the different locations is presented in Section 4. Section 5 discusses the scenario of complementary use of wind and solar energy to achieve a reliable over-the-year energy mix. Finally, the study is concluded in Section 7.

2. Energy from Wind

According to the International Energy Agency, global power generation through renewables was 30.2% in 2023 [16]. In the US, renewable energy generation is expected to more than double from 21% in 2021 to 44% in 2050, making it the fastest growing energy source [17]. This is not surprising given that renewable energy represents the majority of new power added to the generation capacity around the world [18]; it is expected to grow by 17% in 2024 [19].

The integration of renewable energy into the North American power system can produce significant economic benefits as well.

Wind energy has been touted as an important renewable energy source and is expected to be the fastest growing renewable energy source from 2020 to 2024, being the reason for two thirds of the growth in the renewable energy sector during this time [20]. As per data from the US Energy Information Agency (EIA) [21], while the US is responsible for generating 21% of the world’s wind energy generation, the utilization of wind energy for power generation by Kentucky is well below other states with similar wind characteristics. The wind speed in Kentucky seems similar to most of the east coast and west coast, whereas it is higher in the central region of the US. Even with this, utilization potential of wind for power generation in each area is high and most of the benefit is being taken from this resource. Moreover, according to the Office of Energy and Renewable Energy, Kentucky currently only produces 0.23% of its total electricity generation from solar and none from wind [22], with the majority being generated by coal followed by natural gas and hydropower as illustrated in Figure 2 [22].

In fact, according to the US Wind Turbine Database [23], there are no wind turbines located in Kentucky, which indicates to the underutilization of wind as a power generation source in the state. Therefore, there is a need to explore the potential of using wind as a complementary or supplementary energy source. With government incentives [24] recently being passed, wind-based energy generation can be a potentially viable power generation source.

3. Data Collection

With the aim of analyzing the potential for wind power generation throughout Kentucky, data collection was performed in a variety of locations while dividing Kentucky into different geological areas. Considering data availability constraints, eight locations were chosen as follows: the metro areas of Louisville and Lexington, Elizabethtown, Northern Kentucky (NKY), Eastern Kentucky (EKY), Western Kentucky (WKY), Midwestern Kentucky (MWKY), and Southern Kentucky (SKY). Weather data were scraped from the publicly available weather station website Weather Underground [25]. Weather Underground is a commercial service providing weather information in real-time. The data are sourced from hobbyists as well as commercial weather stations. The station locations from which data were sourced are shown in Figure 3.

In order to improve the fidelity of the data, multiple stations were considered for the most populated areas of Louisville (four), Lexington (three), and Elizabethtown (two), while for the other locations, data from only one station was collected. The data for the stations was updated regularly every five minutes. Also, in order to ensure that a uniform comparison was made, data were collected for the years 2020 and 2021 from the beginning of January to the end of December. Moreover, it should be noted that all wind speeds were measured in m/s.

4. Results and Discussion

In this section, the data collected from Weather Underground were analyzed and a discussion is presented on the results of the analysis.

4.1. Wind Speeds by Season

An important aspect to explore when considering the use of wind as a potential source of energy is the variation with respect to season. This is important as energy requirements vary with season. To provide insight into the trend of the seasonal variation of wind, Figure 4 shows the box plots for the wind speed by season for the eight stations, the circles representing outliers in terms of deviation from the mean. It can be observed from the figure that the highest speeds are observed in winter, followed by spring and fall, with the lowest wind speeds observed in summer for nearly all locations considered.

It can also be observed that from the data collected, the locations at Elizabethtown, Lexington, Louisville, and Western Kentucky (WKY) demonstrated the highest wind speeds among the stations, whereas the location considered in Eastern Kentucky (EKY) was the least windy. When looking at the mean wind speeds with respect to the location and season together, it can be observed that an appreciable difference is observed between the winter and spring for Elizabethtown, Lexington, MWKY, and SKY, whereas the difference is not as stark for EKY, WKY, and Louisville. For NKY, only a small difference is observed in the mean wind speed in winter and spring. In terms of trend with respect to season considering the mean wind speeds, for Louisville, it was observed that the mean wind speed in summer was slightly higher than in fall, whereas for all other locations, the progression in mean wind speeds followed a decreasing pattern of winter, spring, fall, and summer with winter being the highest and summer being the lowest.

In conclusion, with the mean wind speed either doubling or tripling in winter for nearly all locations (with the exception of Louisville where it increases by more than half) when compared to summer, the winter season was observed to hold the largest potential for power generation from wind.

4.2. Wind Speeds by Month

In order to compare the wind speed for all stations on a monthly basis, the plots for the mean wind speed for each month for all stations are shown in Figure 5. It can be observed that Elizabethtown is the windiest station as observed by month as well. The least windy is Eastern Kentucky. Moreover, it can be observed from the figure that the wind speed is sufficiently high for the months of January–April, November, and December. However, for the other months, a below-par speed is observed.

Moreover, for most stations, there was little variation in the wind speed for the months of June–October, which were the months of summer going into fall. The exception to this was Elizabethtown, where, for the period of data collection, the mean monthly wind speed started to pick up in October. The months with the lowest wind speeds were observed to be July or August, depending on the location being considered. The exception was NKY, where the lowest wind speed was observed in October. When looking at highs of mean monthly wind speeds, March was the month where the peak was observed for five of the eight locations with the other months being February, April, and May.

4.3. Wind Speeds by Day

In order to understand the daily variation of wind speed, for each location, the average wind speed was computed on a daily basis. From this,

Table 1. Windiest day (on average) for all locations.

Station/Location	Date-Month	Average Wind Speed (m/s)
EKY	1-April	0.91
Elizabethtown	28-March	3.1
Lexington	9-May	1.63
Louisville	25-December	1.86
MWKY	31-March	1.65
NKY	5-February	1.76
SKY	31-January	2.49
WKY	10-May	2.81

lists the dates on which the wind was the highest on average for the combined two years considered. It was found that five of the eight locations from which data were analyzed had the windiest day in the spring (end of March, April, and beginning of May), whereas three had their windiest day in the winter (end of December, January, and beginning of February).

4.4. Wind Speeds by Hour

In order to gauge the trend in the wind speed variation, Figure 6 illustrates the hourly wind speeds for the locations considered. It can be observed that a clear diurnal pattern is present in the variation of mean hourly wind speed throughout the day. Wind picks up speed in the morning and quiets down in the evening with the highest wind speeds being observed between the hours of 10:00 and 18:00. Overall, wind speeds are higher during the daytime as compared to nights. When comparing the difference between the peak and the trough of the daily mean wind speed, it was observed for all eight locations that wind speed at the peak time increased by more than half when compared to the hour of the lowest mean wind speed. Such a trend was observed for all eight station locations considered in this work.

In order to better understand the wind speed,

Table 2. Windiest hour (on average) for all locations.

Station/Location	Hour of the Day (00–24)	Average Wind Speed (m/s)
EKY	14:00	0.54
Elizabethtown	13:00	2.01
Lexington	14:00	1.03
Louisville	14:00	1.23
MWKY	12:00	0.93
NKY	14:00	0.76
SKY	13:00	1.15
WKY	13:00	0.54

lists the windiest hour on average for all considered locations. It was found that the windiest hour for all stations was found to be in the afternoon. Out of them, four stations had 14:00 h as the windiest hour, three had the most wind on average at 13:00 h with one station/location having the most wind on average at noon.

4.5. Wind Data Comparison

In this section, the publicly collected wind data from Weather Underground were compared with two different data sources. The first is a home-based weather station in a suburb in Louisville, Kentucky and the other is a wind turbine in Louisville, which also provided wind measurements. This comparison is presented to highlight the patterns of observed wind speed variation.

4.5.1. Home-Based Weather Station Data

In order to compare the wind pattern from the data analyzed, data from a home-based weather station were acquired for the time period between 1 July 2021 and 15 March 2023. This was then resampled by month and hour. The plot for monthly wind speeds has been illustrated in Figure 7. Moreover, the plot for hourly wind speeds has been illustrated in Figure 8.

From Figure 7 and Figure 8 it can be observed that the pattern for wind speed variations is very similar for the collected data and the data acquired from the home-based weather station. As observed in Figure 5 and Figure 7, the peak mean monthly wind speed occurs in March for the self-collected data from Louisville and the home-based weather station in Louisville. Similarly, the months of July, August, and September are seen to show the lowest wind speed with little change across them. A nearly identical trend can also be observed between both data when considering hourly wind speed variations as well on a daily basis; the increase between the peak of the day and trough for the home-based weather station is more than half as what is seen in the self-collected data. Lastly, the fact that the data from the home-based weather station contain wind speed information outside the time frame of the self-collected data (1 January 2020 to 31 December 2021) attests to the dependability of the wind as an energy-producing resource on a monthly and hourly basis.

4.5.2. Wind Turbine Data

In order to further compare the wind patterns from the data analyzed, data from a wind turbine in Louisville were acquired. The data contain wind speed measurements for twelve days between 14 February 2024 and 26 February 2024. In order to understand the pattern better, only the collected data from Weather Underground for the period between 14 February and 26 February but between the years 2020 and 2021 have been utilized. The plot for hourly wind speeds is illustrated in Figure 9. It should be noted that the wind speeds have been normalized to 0 and 1 to attain a fair comparison.

It can be observed from Figure 9 that there is a slight offset in terms of the pick-up of wind during the day in the wind turbine data as compared to the data collected both from Weather Underground as well as the data acquired from the home-based weather station.

5. Solar Energy

One disadvantage that energy sources like wind have is fluctuation in their availability. However, power generation by wind can be complemented with solar power generation to provide a more dependable source of energy. In order to provide an assessment of this possibility, solar power generation data from the LGE E.W. Brown Solar Facility [26] are used. The data were acquired for the period between 1 January 2020 and 31 December 2021.

5.1. Solar by Season

The seasonal total energy generation by the facility is shown in Figure 10. It can be observed that the most energy is generated around summer time when the solar irradiation is the highest and the days are the longest. Solar power production drastically decreases during the winter when the sun’s angle is at its lowest. This behavior is the opposite when compared to the prevalence of wind (Figure 4) in terms of speed; the peak season of wind speed is observed during winter. Moreover, it can also be observed from the figure that the amount of energy generated for spring and summer is very similar. It can also be observed that for the spring, a high of solar energy is observed and was the same for wind (Figure 4). In this regard, there is a seasonal overlap in terms of the power generation potential when considering wind and solar as possible sources of energy.

5.2. Solar-Wind by Month

In order to further explore the relationship between using solar and wind as energy generation sources, wind speed data collected from Weather Underground across all stations together have been averaged with respect to month and normalized. A similar aggregation for solar power is also performed, the aim being to observe the pattern of variation of the two sources together. The plots are shown in Figure 11. It can be observed that the peak of wind energy occurs in the month of March and from there, starts decreasing, reaching the lowest in August and exhibiting low values in the three months of July, August, and September. On the contrary, the peak of solar energy occurs in the month of July with relatively high values being observed for the months between May and August. In conclusion, when looking at monthly variations, both wind and solar energy demonstrate a complementarity with respect to their peaks over the year, which can be exploited for power generation. Solar energy is observed to dominate the months forming the spring and summer, whereas wind energy starts to pick up in the months of the fall going into winter when solar energy begins to die down. Lastly, there is an overlap in the spring and the fall months when the transition between the seasons takes place with the months of October and November seeing lows for both energy sources.

5.3. Solar-Wind by Hour

For a more granular analysis, the wind and solar data were resampled for every hour and then normalized. This is carried out to investigate the comparative patterns during the day. The plots are illustrated in Figure 12. As expected, the peak of the solar energy is observed to be around mid-day when the sun is at the highest point during the day and therefore solar irradiation is the highest. The highest wind speed (averaged across all locations of data) occurs a few hours later in the afternoon. A high of solar energy is followed by high wind speed. This is due to convection, where warmth from the sun creates an upward movement of warm air. Dense or cooler air from the surroundings then moves into this low-pressure area to fill in the void and therefore gives rise to an increase in wind speed. Following this, as solar energy dies down as the sun sets, wind speed also decreases.

This difference of a few hours between the highs of solar and wind energy can be exploited by power generation companies in optimizing planning and deployment of power generating resources in their distribution networks, e.g., as solar starts to decline, deployed wind power generation capabilities will continue power generation to supplement the drop in solar power generation, resulting in a dependable supply of power to consumers.

6. Experimental Data

In this section, experimental data for the month of March 2024 were used in a facility in Kentucky that had both wind turbine and solar power generation capabilities. These data are very useful as they can be used to gain insights into the wind–solar relationship as well as variations occurring at the same location. The wind turbine installed was an NPC-100-C-28 90kW turbine with a total height of 50 m and a hub height of 36 m. Two types of solar installations were installed at the site: one is a fixed tilt rack and the other is a set with a solar tracker that follows the sun. In order to perform a fair comparison, all three types of sources are compared after normalizing their productions. Moreover, in order to understand the behavior of the three source types, an analysis of the data has been presented on both an hourly as well as on a minute basis.

6.1. Wind and Solar by Hour

Figure 13 illustrates the variation of wind and solar energy by the hour with both the fixed as well as solar tracking illustrated. As mentioned previously, the data for each source have been normalized so that an assessment can be performed of the pattern of variation and the relationship between them.

It can be observed from the figure that the peaks of both solar installations and wind are offset by a few hours, with solar peaking around late morning and mid-day hours and wind peaking around late afternoon. This is similar behavior to what is observed in Section 5.1. Also, as previously observed, there is an overlap between the two sources when the solar generation starts to reduce and wind picks up.

It can also be observed that of the three types of sources illustrated, wind demonstrates the most fluctuations with the two solar types demonstrating appreciable stability. Moreover, when comparing the two solar installations, one can observe that the power generation time of the solar power setup with tracking is considerably wider when compared to the fixed tilt rack installation. This adds support for the use of solar tracking in solar power generating stations to maximize power generation capabilities.

6.2. Wind and Solar by Minute

The variation of wind and solar energy sources by minute is shown in Figure 14 for the complete month of March 2024. This is performed to understand wind variation at a fine resolution over the entire month as well as to observe behavior over different days.

It can be observed that both solar generation setups provide a predictable and periodic pattern of energy generation. The peaks occur during daytime which are followed by troughs occurring as the day ends. This keeps repeating with the daily cycle of solar irradiance. It can also be observed that there are days of low solar production; this is especially apparent on 26 March when storms hit the area around the site and the solar production dropped drastically due to the cloud cover of the storm. When comparing the two solar generation schemes, it can be observed that as observed in the hourly granular data, the duration of power generation by the solar generation setup with tracking is larger than the generation time offered.

When considering wind, a much larger fluctuation is observed when compared to solar generation. While a cyclic pattern can more or less be observed over the month with peaks and troughs occurring on a daily basis, these are not as pronounced as they are for solar energy in terms of dips in energy generation during the night. Wind is seen to be able to produce some energy during night time contrary to solar. Lastly, when looking at complementary behavior, it can be observed that solar and wind do exhibit complementary behavior in terms of there being less solar and more wind daily at certain times, especially on cloudy days such as March 26.

To help provide insight into grid integration issues when combining wind and solar sources into the power grid as well as to understand the fluctuations over short time intervals, Figure 15 illustrates the three generation sources on a minute basis for the hour with the highest values on average for March 2024. The values are expressed as a percentage of the maximum value attained in the hour with the highest values on average.

It can be observed that both the solar sources are the most stable with very little variation occurring over the hour. On the contrary, there is considerable variation in the wind which drops down to nearly 70% of its hourly maximum at its lowest. When looking at the means and standard deviation of the sources, the mean of the wind stood at 85.8 with both solar generation sources having means around 99.5. As expected, wind had the highest standard deviation among the three sources for the hour standing at 7.34 whereas the standard deviation for the fixed solar stood at 0.17 with the solar with tracking at 0.25. While the standard deviation of the solar generation system with tracking is higher than its fixed rack counterpart, the mean value of the former over the hour was observed to be slightly higher due to the fact that it followed the path of the sun.

While considering fluctuation for short time intervals for both sources, it is pertinent to mention that solar irradiation fluctuation over short time intervals can be caused by a variety of reasons, such as moving cloud cover, aerosols in the atmosphere, etc., in addition to other such phenomenon. Similarly, wind is dependent on conditions in the atmosphere, differences in terrain (even over relatively short distances), obstacles, and local weather, which can give rise to fluctuations as well. These variations in the generation potential of both sources might necessitate the use of additional arrangements in terms of battery storage systems, demand side management, and accurate forecasting techniques to make full use of the short and long term complementarity of wind and solar sources for power generation.

7. Conclusions

The current work provides an exploratory study into the wind patterns for the Commonwealth of Kentucky. To perform the analysis, data were sourced from the publicly accessible Weather Underground portal for eight different locations across Kentucky, covering the three urban areas of Louisville, Lexington, and Elizabethtown, as well as the regions of Eastern Kentucky, Southern Kentucky, Western Kentucky, and Midwestern Kentucky. An analysis was then performed with different granularities, including seasonal, monthly, and hourly to investigate the pattern of variation at the different locations considered. It was found that the wind speeds were the highest in the winter, followed by spring, while they were the lowest in summer. When looking at the month, all locations had the highest wind speeds either in the winter or spring months. Considering wind speeds on an hourly basis, it was observed that the wind speed picked up around morning time and decreased in the evening. Moreover, the pattern of the collected data was compared with data acquired from a home-based weather station as well as a deployed wind turbine. It was observed that the pattern of wind variation was very similar from all data sources.

Furthermore, in order to assess the complementarity of wind with solar, data were acquired from a solar power plant in Kentucky. The energy generated by the solar power plant was taken as an indicator of the solar power availability. A seasonal and month-wise comparison revealed that wind and solar energy are complementary to each other in terms of each having peaks and troughs at opposite times of the year. The daily variations of wind and solar energy were also analyzed to ascertain the daily variations between them. It was observed that the wind was produced at an offset of the solar pattern. Finally, to further validate the analysis performed, experimental data acquired from a site consisting of wind and solar power generation installations were explored, which corroborated the findings of the study performed.

Based on the study here from a variety of locations across Kentucky, our analysis of wind data across Kentucky indicates that wind is a potential source of low-carbon electricity generation to seasonally complement solar and integrate with natural gas combined cycle plants for a more diversified and resilient electricity portfolio.