Below is a flow chart of the analysis models I created. The basic steps are the same for both types of suitability analysis with a few key differences. The methods depicted by this flow chart are described and discussed below.
In order to identify and evaluate locations that are suitable for wind energy generation I developed these analysis models. To maximize the validity of comparison between the results for suitability for the two different types of wind energy exploitation I am investigating, commercial scale grid-connected wind power and off-grid small scale wind power, the same spatial analysis steps are used for both analyses. This limits the confusion that could be created by comparing the products of two completely different models. The difference between the two models is that the criteria for some of the analysis functions have been changed to reflect the different requirements and benefits of the different types of wind power generation. Also the transmission lines data set is only used in the Commercial Scale analysis model.
I will begin by describing the structure of the model, which is shared in both analyses. The model combines two basic analysis model types, binary and ranking. First a variety of physical features are analyzed with a binary model to identify locations as Suitable or Unsuitable for wind power generation. Then the locations that have been identified as Suitable are used to extract classified wind data so that the output map depicts the areas that are suitable for each type of wind power generation ranked by available wind energy.
Since my intention in creating this model is to allow the analysis steps to be quickly applied to areas of Nicaragua (and later other countries) outside of my initial investigation area I chose to use the nationwide data sets I have compiled as my inputs. These comprised four Shapefiles from SWERA representing cities, rivers, wind energy and transmission lines as well as the Digital Elevation Model raster file I created from the SRTM data. First the Select function is applied to the cities file to select Bluefields as the center of interest. Then a 100 km buffer is generated around Bluefields and this Shapefile is used to clip and extract all of the input layers to isolate the area to be analyzed. The advantage of using this method is that the subsequent analysis steps can be applied to any portion of Nicaragua simply be changing either the city identified by the Select function or by changing the distance in the Buffer function.
From this point, analysis steps were taken to identify all locations as suitable or unsuitable for a wind installation. Three dimensional raster surface analysis functions were used to calculate the slope and aspect of the area of interest. These results were reclassified to assign a value of 1 for Suitable and 0 for Unsuitable based on the criteria I chose. For slope I classified a slope greater than 30 degrees as unsuitable due to the difficulty of erecting towers that is caused by steep slopes. For aspect, I identified aspects between 157.5 and 337.5 degrees as well as flat land as suitable and all other aspects as unsuitable. This decision is based on the wind rose below which I obtained from blueEnergy's wind resource study in Bluefields. I decided that the fairly clear difference between the relatively large quantities of wind energy coming from these directions and the fairly limited amount of wind coming from other directions justified the use of a binary rather than a ranking model here. This is an area with very little installed wind capacity and the intention of this model is to identify areas that are particularly promising for future development, not to maximize the amount of land identified as Suitable. It is important to note that the classification ranges for aspect are only valid for the area around Bluefields. If the model is applied to an area that does not share the same prevailing wind directions as Bluefields, the classification of aspect will need to be changed accordingly.
Source: blueEnergy
Next, a buffer of 50 meters is created around all rivers then converted to raster and reclassified so that areas within the buffer are assigned the value of 0 for Unsuitable. This is intended to exclude areas along river banks that may be prone to flooding or instability, may have particularly dense vegetation coverage, and are frequently used by humans and wildlife.
Since the remaining inputs are analyzed differently depending on the type of wind power installation being examined I will now discuss the differing site requirements. A key consideration in siting commercial-scale wind power sites is the fact that it is very expensive to transport electricity in areas that do not have existing electric transmission lines. For this reason it is necessary to locate wind farms near existing transmission lines. To take this into account, the model creates a two kilometer buffer around transmission lines, converts the resulting features to raster and then reclassifies the raster such that only locations within the buffer area are identified as Suitable. Since Nicaragua has limited investment capital it is unlikely that a wind farm developer would be willing to construct transmission lines any longer than two kilometers to reach the existing transmission lines. Another feature of commercial wind sites is that they involve large turbines that can be deemed unsightly or potential conflict with other human activities. For this reason the model creates 500 meter buffers around all communities, converts the buffers to raster and reclassifies the raster such that locations within the buffers are identified as Unsuitable. A third consideration of commercial-scale wind power is that relatively high wind speeds are required because high levels of electric output are needed to justify and payback the large investment involved in construction. The model classifies wind power density in watts per square meter in the following manner:
This classification essential creates suitability rankings that are equal to one less than the wind power class as defined by the Energy Information Administration. This imposes a minimum wind power density of 200 watts per square meter, which is roughly equal to an average wind speed of 5.6 meters per second at 50 meters, for suitability for commercial wind generation.
The analysis model for analyzing site suitability for small off-grid wind turbines for electrification purposes handles these three factors differently. The first major difference is that proximity to transmission lines is not a factor in this model. I initially intended to use the 2 kilometer transmission line buffer to exclude areas in proximity to the line but since I do not know how many of the communities near transmission lines actually receive electricity from them, I decided I should eliminate this potential source of error. This is discussed further in the Future section. Secondly, since small-scale turbines such as the ones that blueEnergy is installing are intended to provide electricity to communities, it is necessary that the turbines be located near communities. Again the cities data is buffered to 500 meters and converted to raster but the result is classified such that only locations within the buffer are Suitable. This greatly reduces the amount of suitable land but identifies locations where this technology can benefit people. Thirdly, the amount of wind energy required for a blueEnergy turbine to be practical is much lower than the amount required for a wind farm because these turbines are designed for a low cut-in speed since it is more important to ensure that the turbines generate some electricity every day rather than maximizing their output for windy days. For this reason the small wind model omits the Unsuitable classification and instead uses the values from the Wind Power Class field of the table below as the ranking value.
Source:Energy Information Administration
With these analysis steps completed according to which model is being used, the remaining steps are the same. The binary identification of locations as Suitable or Unsuitable is accomplished by using the Times function to multiply the classified rasters for aspect, slope, rivers, transmission lines, and cities. This means that cells that have a value of zero in any of the input rasters will have the value of zero in the output raster. The Extract by Attributes tool is then used to create a new raster that contains only the cells with a value of one, these are the Suitable locations. The Extract by Mask tool is then used to extract the cells from the reclassified wind data raster that are located in suitable locations. This produces the output maps of these models which rank the suitable areas according to available wind energy.