Method for assessing occurrence and severity of power outages due to weather events

Abstract

A method for predicting the occurrence and severity of power outages is based on predicted weather hazard data for an area for a period of time together with granular information regarding the amount and locations of overhead powerlines. In addition, the extent and degree of leaf cover near the overhead powerlines factors into the determination of the likelihood of potential power outages. Weather hazard data may include a peak wind gust within the area, which may be divided into a plurality of cells with weather hazard data included for each cell. The amount of overhead powerlines present within the area or each cell is used to refine the energy reliability predictions, along with the age and design standard of the overhead powerlines. Based on these factors, critical weather hazard threshold values are adjusted and compared to corresponding predicted values for each cell to determine an energy reliability index.

Description

METHOD FOR ASSESSING OCCURRENCE AND SEVERITY OF POWER OUTAGES DUE TO WEATHER EVENTSRELATED APPLICATION DATA

[0001] This application claims the benefit of priority of U.S. Provisional Patent Application Serial No. 63 / 569,843, filed March 26, 2024, and titled “Systems and Methods For Predicting Severity of Power Outages From Weather Events,” which is incorporated by reference herein in its entirety.FIELD OF THE DISCLOSURE

[0002] The present disclosure generally relates to the field of power grids and weather forecasting. In particular, the present disclosure is directed to systems and methods for predicting severity of power outages from weather events.BACKGROUND

[0003] The majority of power outages are caused by extreme weather hazards, such as high winds, wet snow, or accumulating ice. The loss of power has negative cascading effects to critical infrastructure systems such as transportation, communication, and water systems. Electricity is critical for all essential functions and anticipating the potential duration of outages before they happen can improve preparedness and response through storm-based risk mitigation actions.SUMMARY OF THE DISCLOSURE

[0004] A method for predicting the occurrence and severity of power outages includes receiving predicted weather hazard data for a given period of time and a given geospatial area, the weather hazard data including at least a peak wind gust, a wet snow accumulation amount, or an ice accumulation amount, wherein the geospatial area is divided into a plurality of cells within the geospatial area and wherein the predicted weather hazard data includes a cell-level predicted weather hazard value for each of the plurality of cells, transforming the weather hazard data to a common raster layer, determining an asset value for each of the plurality of cells, wherein the asset value is based on a linear amount of overhead powerlines present within each of the plurality of cells, determining, for each of the plurality of cells, an asset resilience factor based on a leaf cover index, an asset age, and an asset design standard value for overhead powerlines within each of the plurality of cells, determining a weather hazard asset vulnerability factor based on the resilience factor and the asset value for each of the plurality of cells, aggregating each of the weather hazard asset vulnerability factors to determine an area weather hazard asset vulnerability factor, adjusting acritical peak wind gust threshold value, a critical wet snow accumulation threshold value, and a critical ice accumulation threshold value based on the area weather hazard asset vulnerability factor, and determining an energy reliability index for the given area for a date within the period of time based on a comparison of the peak wind gust, the wet snow accumulation amount, or the ice accumulation amount for the given area on the date with respective ones of the adjusted critical peak wind gust threshold values, the adjusted critical wet snow accumulation threshold values, or the adjusted critical ice accumulation threshold values.

[0005] Additionally or alternatively, the critical peak wind gust threshold value is adjusted based on a seasonal leaf area index.

[0006] Additionally or alternatively, the leaf cover index for each of the plurality of cells is based on the period of time and a geographic location of each of the plurality of cells.

[0007] Additionally or alternatively, the asset age is based on an average age of overhead powerlines and corresponding support structures of the overhead powerlines within each of the plurality of cells.

[0008] Additionally or alternatively, the asset design standard is based on an average asset design standard of overhead powerlines within each of the plurality of cells.

[0009] Additionally or alternatively, further including determining an average asset-weighted predicted weather hazard for the area by finding a sum of a distance of overhead powerlines within all of the plurality of cells, dividing the distance of overhead powerlines of each of the plurality of cells by the sum to obtain a cell fraction for each of the plurality of cells, multiplying the cell fraction for each of the plurality of cells by the cell-level predicted weather hazard value for that cell, and summing the result of each such multiplication.

[0010] Additionally or alternatively, the leaf cover index is based on leaf cover information within 30 meters of locations of the linear amount of overhead powerlines present within each of the plurality of cells.BRIEF DESCRIPTION OF THE DRAWINGS

[0011] For the purpose of illustrating the disclosure, the drawings show aspects of one or more embodiments of the disclosure. However, it should be understood that the present disclosure is not limited to the precise arrangements and instrumentalities shown in the drawings, wherein:FIG. 1 is a process diagram for determining a power outage likelihood based on factors including forecasts and asset conditions in accordance with an embodiment of the present disclosure;FIG. 2A is a map of an area divided into cells with assets shown;FIG. 2B is the map of the area of FIG. 2A with forecast wind speeds for each cell;FIG. 3 is a map of tree cover risk along powerlines at the town level, including asset-weighting for absolute tree cover risk;FIG. 4A is a map of a town with the county shown in FIG. 3 in which a geospatial aggregation of tree cover risk are given at each cell level based on tree cover and asset information for each cell;FIG. 4B shows a portion of the map of FIG. 4A with a detail view of a portion of one cell showing tree cover density along the powerlines in that cell;FIG. 4C is a detail view of a map of overhead powerlines showing sections with different levels of tree cover;FIG. 5 is a graph showing example seasonal leaf-area indexes for three locations over Julian dates, where leaf-area index is used as a proxy for tree canopy condition;FIG. 6 is a graph showing the relationship between power outage duration (hours when at least 5% of the population is without power) and intensity (the peak percentage of customers without power) for counties with at least 10,000 tracked customers across four storm events;FIG. 7 is a map showing example resilience factor multipliers on a county level used to adjust weather hazard thresholds in which resilience factor is based on a combination of tree cover, asset age, and engineering design standard;FIG. 8 is a process diagram for determining average asset-weighted weather values for an area; andFIG. 9 is a process diagram for adjusting an energy reliability index based on tree cover around pertinent overhead powerlines.DETAILED DESCRIPTION

[0012] The severity of potential power outages, such as duration, are predicted based on inputs for factors that influence the likelihood of damage to overhead powerlines under predicted weather conditions. Scenario-based or probabilistic inputs of each extreme weather hazard are used as part of the determination of a level of confidence in the outputs. Predicted weather information for various points in the future can come from a plurality of sources and be utilized at varying geospatial andtemporal resolutions. The impact of the accumulation of wet snow and ice on trees or powerlines can be estimated, such as is taught in commonly owned U.S. Pat. No. 11,143,793, “Storm outage management system”, which describes how to isolate wet snowfall accumulation using wet bulb temperatures and that the accumulation of ice accretion from freezing rain is dependent on the precipitation intensity, wet bulb temperature, and wind speed.

[0013] These and additional factors, including the extent of the presence of overhead powerlines, the condition of the powerlines and their support structures, the degree of tree cover near the powerlines, the amount of foliage present at the time, and the expected weather are combined to determine a power outage severity prediction for a given area over a given period of time.

[0014] An overview of the process for determining the likelihood of the occurrence and severity of power outages is given in FIG. 1. To generate a likelihood of occurrence and severity of a weather-based power outage for a given area, weather hazard information is obtained at step 1, and may be on a continuous grid or as raster values that can be presented on a geospatial scale as a cell value. Predicted weather class hazard information is available on an hourly time scale and can include wind information, accumulated wet snow liquid water content, and freezing rain icing accretion over some time period. Wind information can include peak wind gust, sustained wind speed, duration of high winds above some threshold, or direction at different elevations above ground level. This data is compiled at step 2.

[0015] These weather hazard inputs for a given location and time frame are transformed at step 3 into one common gridded layer with the same coordinate reference system and forecast time attribute. For example, values may be determined of predicted wind information for a forecast ensemble member scenario of high winds predicted across the area of a county at a cell level. Individual values for the smallest selected geospatial units are determined at step 4 and may show considerable spatial variability in wind speed across the area of interest. Asset information, i.e., the shape or topology of overhead powerlines, are determined at the cell level for the area of interest at step 5. This asset information is used in conjunction with the cell-level weather forecast information to improve the prediction of weather impacts on energy reliability by determining asset-based weight to give each weather value for each cell at step 6. Such information may be acquired or estimated, such as through available records, satellite images, or association with roads and structures.

[0016] At step 7, an asset-adjusted weather hazard value is derived based on the predicted weather hazard value (e.g., peak wind gusts) for a cell and the assets present in that cell. At step 8, a tree cover factor is determined for the assets in each cell based on the geographic location of the cell, the time of year, and the proximity of tree coverage to the powerlines. At step 9, asset age is determined, i.e., the age of the powerlines, which may be based on the average age within a cell, for example, and the design standard for the powerlines in the area is determined, which may also be based on an average. Based on the tree cover factor, asset age, and asset design, a resilience factor multiplier is determined at step 10.

[0017] Based on the tree cover factor, the asset age factor, and the asset design factor for each cell, the critical thresholds are adjusted for each weather hazard (e.g., wind, wet snow, accumulating ice) at step 1 1 . At step 12, reference is made to an energy reliability index table of critical thresholds for each hazard and compared to corresponding predicted hazard values for each cell of the given area, and an energy reliability value is output at step 13 for each cell. The energy reliability value may be on any scale, such as 0-5, and may be provided at various confidence levels, such as ensemble mean, 10thpercentile, 25thpercentile, 75thpercentile, and 90thpercentile. In this way, predictions of the occurrence and severity of power outages can be made that are based on the particular conditions and circumstances present with respect to assets rather than mere general weather forecasts.

[0018] FIGS. 2A-2B are a map 100 of an area of interest divided into a plurality of cells 104 (e.g., 104a-104c). Overhead powerlines 108 (e.g., 108a-108b) are overlaid on map 100, showing locations of these powerlines with respect to cells 104. This information is used to adjust an aggregated predicted weather hazard value based on the weather hazard predictions on a cell level in conjunction with the relative amount of overhead powerlines present in each cell. In FIG. 2B, predicted weather values 112 (e.g., 112a for cell 104c) are shown for each cell 104, in this example, peak wind gusts in mph. The potential of these predicted hazards to result in power outages varies based on the extent of powerlines present in a given cell. Cell 104c, for example, does not include any overhead power lines and so weather hazards in that cell will be unlikely to result in power outages. An average asset-weighted weather value can be determined for an area as outlined in FIG. 8, for example, by determining the amount of overhead powerlines present in each cell of an area, e.g., distance of overhead powerlines, and from that finding the total amount of overhead powerlines in the area. Dividing the distance of lines in a cell by the total distance of lines in the area gives the fraction of lines in each cell, and these values can be used by multiplying the fraction for each cellby the predicted weather value for that cell. Summing over all cells in the area results in an average asset-weighted weather value for the area, which provides better indicators of the likelihood of occurrence and severity of outages in the area. In the example in FIG. 2B, the average predicted wind gust for the area is 44 mph, while the asset-weighted average wind gust is 48 mph (55 * 0.21 + 54 * 0.15 + 47 * 0.22 + 46 * 0.12 + 45 * 0.08 + 43 * 0.06 + 39 * 0 + 32 * 0.02 + 38 * 0.14).

[0019] Several additional factors affect the potential resilience or durability of the electric distribution grid to power outages. These factors include tree cover along overhead lines, the age of the assets, and the design standard used for assets, which may all vary by region. Tree cover is the most spatially variable factor and can be aggregated from smaller to larger scales. FIG. 3 shows tree cover information (low, medium, high) for an area, which includes a town 205 and is shown at the level of cells (e.g., cell 204a). As shown in FIGS. 4A-4B, high-resolution tree cover information can be joined to powerline topology within each common weather information grid cell, across different classes of tree cover near (e.g., within 30 meters) and along overhead powerlines. In FIG. 4A, map 200 includes town 205, and each cell 204 (e.g., 204b) includes overhead powerline 208 information (e.g., 208a). In FIG. 4B, in which a portion of cell 204c is shown in detail, tree cover information is joined to powerline information 208 (e.g., 208b). FIG. 4C is a detail view of a map showing a section of overhead powerlines 208d with portions having different levels of tree cover nearby.Some portions have little or no tree cover, others have low (along box 211), medium (along box 213, and high (along box 215). These amounts of tree cover along different sections of lines are used to determine a tree cover factor for the overhead lines in each cell, and the tree cover factors are then aggregated to determine an area tree cover factor for the area containing the cells. Alternatively, the tree cover factors for each cell are used to determine an energy reliability index for each cell.

[0020] The tree cover information is used as another factor in the resiliency determination. For example, the amount of powerline distance may be multiplied by a tree cover value (e.g., values representing low, medium, or high coverage) and summed across each grid cell to derive an adjusted measure of risk. The tree cover factor is based on a leaf area index which varies by location and time of year (see, e.g., FIG. 5), and can be used to adjust the determined energy reliability index for a given predicted event at a location. As outlined in FIG. 9, a leaf area index is determined based on a location and date for the area, which influences the extent of tree canopy. In addition, the extent of tree cover around overhead lines in the area contributes to the threat of tree damage. Therefore, for the given location, it is determined which cells are within the location and for each such cell, the total overhead powerline distance is determined. For the entire distance of overhead powerline ineach cell, the percent having no, low, medium and high tree coverage is determined (from information shown in FIG. 4C, for example), and a tree coverage factor is found for each cell. These values can be aggregated to produce an area tree coverage factor,

[0021] From the location, date, and tree coverage factor, a leaf area index is determined. If this index is below a certain threshold, it may be considered low canopy and, for a given predicted weather hazard value, the energy reliability index will be based on corresponding critical hazard thresholds for low canopy. However, if the leaf area index is above the threshold, for the given predicted weather hazard value, the energy reliability index will be based on corresponding critical hazard thresholds for high canopy. The predicted weather hazard value may also be first adjusted based on other factors, such as asset age and asset design. For example, a predicted asset-weighted wind gust of 45 mph in a location with a current leaf area index of 0.5 will use a low canopy table of adjusted critical thresholds if the leaf area index is 1 (as can be seen in Table 1) instead of the high canopy table, resulting in an energy reliability index of 2, whereas if the leaf area index was above 1, the energy reliability index would be 2 (as can be seen in Table 2).Table 1: Adjusted Critical Thresholds for Low Leaf CanopyTable 2: Adjusted Critical Thresholds for High Leaf Canopy

[0022] Power outage duration and intensity are correlated based on the nature of the physical damage to the infrastructure. Duration can be defined as some time to restore power, whereas intensity can describe the percentage of population or customers affected at the peak of the storm impacts. FIG. 6 is a graph of percent of customers out at peak (i.e., point at which the largest number of total customers are without power for a given area) versus the outage duration, which is defined in this example as the number of hours during which at least 5 precent of customers are without power. This illustrates how power outage duration and intensity are correlated, with more intense events correlated with requiring longer periods to restore power to most customers. Both absolute (time to restoration) and relative impacts (percentage of people affected) can be correlated to create classes of impacts.

[0023] Table 3 shows exemplary energy reliability index levels based on six classes related to the intensity of power outage loss and duration. The energy reliability index, as used herein, is a power outage prediction scale where higher values reflect higher risk of severe power outages. The energy reliability index relates the asset -based and resilience factor adjusted weather hazard to some level of predicted output. The predicted outputs are available as a relative (peak customers percentage) and an absolute (power outage duration) scale, with restoration times ranging from a few hours to more than five days.Table 3

[0024] Resiliency factors are used to adjust critical weather threshold values and examples are illustrated geospatially on a map of the continental US in FIG. 7. The resilience factor multiplier uses a combination of tree cover, asset age, and engineering design standard for the assets. High resilience factors are found for example along the Gulf Coast as a result of the design standards for landfalling hurricanes. Lower resilience factors in western Oregon, Northern Michigan, and much of the interior Northeast are due to higher tree cover.

[0025] Table 4 shows the resilience factor for several example counties based on the values of the factors and how the overall resilience factor is used to adjust critical weather threshold values for the respective counties. For example, Miami-Dade county, in Florida, has an overall resilience factor of 1.32 largely due to higher design standards, whereas Morris County, in New Jersey, has a lower resilience factor of 0.92 due largely to greater tree canopy. The resilience multiplier is used to adjust the critical weather threshold values. In the example in Table 2, the base critical threshold is for wind speed and is 45 mph (i.e., a wind speed at which, all else being equal, assets become at risk). For Miami-Dade county, the adjusted critical threshold of a peak wind gust is 59 mph because of the higher overall resilience factor, whereas Morris County has an adjusted critical wind speed value of 41 mph from the base of 45 mph. These adjusted critical thresholds may be used for the energy reliability index prediction.Table 4Clark, NY1.41.1 1.0 1.19

[0026] Table 5 provides an example of how the energy reliability index thresholds are adjusted using a resilience factor of 1.10. The thresholds for each predicted level are multiplied by 1.10 to change the scaling of impacts unique to each hazard. In this case, the aggregated resilience at some geospatial area is more resilient to extreme weather stressors and the scaling is dynamically adjusted to normalize risk. The resilience-factor adjusted values for critical thresholds are compared to the predicted asset-weighted values for the same hazard(s) to the in order to determine the energy reliability index output level.Table 5

[0027] The energy reliability index is a prediction of power outage duration and intensity across some geographic region during a given period of time. The energy reliability index outputs are provided for five prediction scenarios and can be aggregated on a daily basis showing the risk profiles through time (as shown for example in Table 6 for a county over a five-day period). Withinthe example in Table 6, there is a storm signal on April 18 with a 90% probability of a level 2 impact event and a 10% probability of a level 4 impact event.Table 6

[0028] Prediction of power outage impacts can be made for multiple scenarios. For example, for two forecast scenarios for a path of a hurricane, the potential impacts will vary based on the different possible storm tracks.

[0029] Various modifications and additions can be made without departing from the spirit and scope of this disclosure. Features of each of the various embodiments described above may be combined with features of other described embodiments as appropriate in order to provide a multiplicity of feature combinations in associated new embodiments. Furthermore, while the foregoing describes a number of separate embodiments, what has been described herein is merely illustrative of the application of the principles of the present disclosure. Additionally, although particular methods herein may be illustrated and / or described as being performed in a specific order, the ordering is highly variable within ordinary skill to achieve aspects of the present disclosure. Accordingly, this description is meant to be taken only by way of example, and not to otherwise limit the scope of this disclosure.

[0030] Exemplary embodiments have been disclosed above and illustrated in the accompanying drawings. It will be understood by those skilled in the art that various changes, omissions andadditions may be made to that which is specifically disclosed herein without departing from the spirit and scope of the present disclosure.

Claims

What is claimed is:1 . A computer-implemented method for assessing likelihood of occurrence and severity of power outages comprising: receiving predicted weather hazard data for a given period of time and a given geospatial area, the weather hazard data including at least a peak wind gust, a wet snow accumulation amount, or an ice accumulation amount, wherein the geospatial area is divided into a plurality of cells within the geospatial area and wherein the predicted weather hazard data includes a cell-level predicted weather hazard value for each of the plurality of cells; transforming the weather hazard data to a common raster layer; determining an asset value for each of the plurality of cells, wherein the asset value is based on a linear amount of overhead powerlines present within each of the plurality of cells; determining, for each of the plurality of cells, an asset resilience factor based on a leaf cover index, an asset age, and an asset design standard value for overhead powerlines within each of the plurality of cells; determining a weather hazard asset vulnerability factor based on the resilience factor and the asset value for each of the plurality of cells; aggregating each of the weather hazard asset vulnerability factors to determine an area weather hazard asset vulnerability factor; adjusting a critical peak wind gust threshold value, a critical wet snow accumulation threshold value, and a critical ice accumulation threshold value based on the area weather hazard asset vulnerability factor; and determining an energy reliability index for the given area for a date within the period of time based on a comparison of the peak wind gust, the wet snow accumulation amount, or the ice accumulation amount for the given area on the date with respective ones of the adjusted critical peak wind gust threshold values, the adjusted critical wet snow accumulation threshold values, or the adjusted critical ice accumulation threshold values.

2. The method of claim 1, wherein the critical peak wind gust threshold value is adjusted based on a seasonal leaf area index.

3. The method of claim 2, wherein the leaf cover index for each of the plurality of cells is based on the period of time and a geographic location of each of the plurality of cells.

4. The method of claim 1, wherein the asset age is based on an average age of overhead powerlines corresponding support structures of the overhead powerlines within each of the plurality of cells.

5. The method of claim 1, wherein the asset design standard is based on an average asset design standard of overhead powerlines within each of the plurality of cells.

6. The method of claim 1, further including determining an average asset-weighted predicted weather hazard for the area by finding a sum of a distance of overhead powerlines within all of the plurality of cells, dividing the distance of overhead powerlines of each of the plurality of cells by the sum to obtain a cell fraction for each of the plurality of cells, multiplying the cell fraction for each of the plurality of cells by the cell-level predicted weather hazard value for that cell, and summing the result of each such multiplication.

7. The method of claim 1, wherein the leaf cover index is based on leaf cover information within 30 meters of locations of the linear amount of overhead powerlines present within each of the plurality of cells.

8. A computer-implemented method for assessing likelihood of occurrence and severity of power outages comprising: receiving predicted weather hazard data for an area for a period of time, the weather hazard data including at least a peak wind gust, wherein the area is divided into a plurality of cells within the area and wherein the predicted weather hazard data includes a cell-level predicted weather hazard value for each of the plurality of cells; transforming the weather hazard data to a common raster layer; determining an asset value for each of the plurality of cells, wherein the asset value is based on a linear amount of overhead powerlines present within each of the plurality of cells; determining, for each of the plurality of cells, an asset resilience factor based on a leaf cover index, an asset age, and an asset design standard value for overhead powerlines within each of the plurality of cells; adjusting a critical peak wind gust threshold value associated with each of the plurality of cells based on the leaf cover index, the asset age, and the asset design standard value for overhead powerlines within each of the plurality of cells; anddetermining an energy reliability index for each of the plurality of cells for a date within the period of time based on a comparison of the peak wind gust for each of the plurality of cells on the date with respective ones of the adjusted critical peak wind gust threshold values.

9. The method of claim 8, wherein the critical peak wind gust threshold value is adjusted based on a seasonal leaf area index.

10. The method of claim 9, wherein the leaf cover index for each of the plurality of cells is based on the period of time and a geographic location of each of the plurality of cells.11 . The method of claim 8, wherein the asset age is based on an average age of overhead powerlines and corresponding support structures of the overhead powerlines within each of the plurality of cells.

12. The method of claim 8, wherein the asset design standard is based on an average asset design standard of overhead powerlines within each of the plurality of cells.

13. The method of claim 8, further including determining an average asset- weighted predicted weather hazard for the area by finding a sum of a distance of overhead powerlines within all of the plurality of cells, dividing the distance of overhead powerlines of each of the plurality of cells by the sum to obtain a cell fraction for each of the plurality of cells, multiplying the cell fraction for each of the plurality of cells by the cell-level predicted weather hazard value for that cell, and summing the result of each such multiplication.

14. The method of claim 8, wherein the leaf cover index is based on leaf cover information within 30 meters of locations of the linear amount of overhead powerlines present within each of the plurality of cells.

15. The method of claim 8, further including aggregating each of the asset resilience factors to determine a weather hazard asset vulnerability factor for the area.

16. The method of claim 8, wherein the weather hazard data further includes a predicted wet snow accumulation amount and an ice accretion amount for each of the plurality of cells.

17. The method of claim 16, further including adjusting a critical wet snow accumulation threshold value associated with each of the plurality of cells based on the asset resiliencefactor and adjusting a critical ice accretion threshold value associated with each of the plurality of cells based on the asset resilience factor.

18. The method of claim 17, determining the energy reliability index further includes comparing the predicted wet snow accumulation amount for each of the plurality of cells on the date to the critical wet snow accumulation threshold value and comparing the predicted ice accretion amount for each of the plurality of cells on the date to the critical ice accretion threshold value.

19. The method of claim 10, wherein the leaf cover index for each of the plurality of cells is further based on levels of tree cover of overhead powerlines within each of the plurality of cells.

20. The method of claim 19, wherein the levels of tree cover are percentages based on no, low, medium, and high tree coverage over portions of a total distance of overhead powerlines in each of the plurality of cells.

Images