Universal test method and system for wind speed forecast for electric power wind disaster, and storage medium

Abstract

The invention discloses a universal test method and system for wind speed forecast for electric power wind disasters and a storage medium. The method comprises the following steps: calculating a single-point wind speed forecast general test score within a wind speed forecast attention range by using wind speed observation data and wind speed forecast data; calculating a wind speed forecast general test score value set in a neighborhood range; determining the maximum value of the universal test score of the wind speed forecast in the set; determining a disaster-causing wind speed hit or empty report condition according to the wind speed forecast general test score in the neighborhood range within a period of time; and according to the position coordinate of the maximum value of the universal test score of the wind speed forecast in a period of time, obtaining the spatial distribution characteristics of the position error. According to the method, a new system for testing the electric power and wind speed disaster-caused forecast effect is constructed, calculation is simple, rules are objective and transparent, significance is clear, and the method has the characteristics of pertinence, localization self-definition, universality, result continuity and the like. The method is used for investigating the accuracy degree of wind speed forecasting in a numerical mode and the like and has extremely high real disaster prevention decision-making value.

Description

Technical Field

[0001] The present invention relates to the technical field of electric power disaster prevention forecast verification and evaluation, and in particular to a general wind speed forecast verification method, system and storage medium for electric power wind disasters. Background Art

[0002] Wind speed forecast is of obvious importance in power wind disasters. Power facilities such as transmission lines and substations are often affected by strong winds. Especially in severe weather conditions such as typhoons and tornadoes, changes in wind speed may have a serious impact on power facilities. When the wind speed forecast exceeds the set safety threshold, measures such as power outages and facility reinforcement can be taken in time to prevent power facilities from being damaged or malfunctioning. In addition, wind speed forecasts can also help power companies predict the possibility and scope of wind disasters. Through analysis, the time, duration, and intensity of wind disasters can be predicted, so that response measures can be prepared in advance to reduce the loss of power facilities and the possibility of casualties. Therefore, wind speed forecasts play an important role in power wind disasters and can help ensure the stability and safety of power supply.

[0003] Based on this, the power sector needs meteorological wind forecasts as data support, which the power sector will obtain through self-research or external cooperation, and there will be many sources. Therefore, the power sector must understand the pros and cons of these data sources, make choices and focus in their work, and also need to have a scientific basis for trusting these data sources in disaster prevention decisions. Only through wind speed testing can we master the current wind speed forecasting technology, thereby helping power and related departments make reasonable decisions. Therefore, the inspection of these wind speed forecast data is indispensable. However, there is currently a lack of inspection standards specifically for wind disaster needs, and simple error calculations or some meteorological wind speed inspection methods are generally used. On the one hand, these methods are not very targeted, and on the other hand, there are many problems (distortion discontinuity, double penalties, etc.).

[0004] At present, there are many studies on the inspection and evaluation of wind speed forecasts in the field of meteorology. In terms of the definition and classification of wind speed inspection indicators, researchers have systematically defined and classified wind speed inspection indicators in order to better evaluate the quality and accuracy of wind speed data. Common wind speed inspection indicators include binary event inspection methods (accuracy, hit rate, TS score, ETS score, etc.), average wind speed error, standard deviation error, frequency distribution error, correlation coefficient, etc. In recent years, some new evaluation methods have begun to be applied, such as probabilistic forecast evaluation methods and spatial correlation analysis. These methods can more comprehensively evaluate the accuracy of wind speed forecasts. However, there are few studies specifically targeting power wind disasters and specifically evaluating the inspection methods for the prediction effect of strong wind speed disasters. Considering the important role of wind speed forecasts in disaster prevention and mitigation in the power industry, it is necessary to design corresponding inspection indicators specifically for this purpose.

[0005] At present, the most common wind speed inspection indicators used in the power industry are still two forms of inspection. One is usually to set a threshold, and then use a binary event inspection method (accuracy, hit rate, TS score, ETS score, etc.). This method has great limitations. The wind speed is classified by different thresholds, and it cannot be guaranteed that adjacent values ​​can be classified in the same category. If the wind speed values ​​with small numerical differences are not in the same category, it will cause distortion and discontinuity in the score. For example, if 20m / s is used as the threshold for strong wind events, assuming that strong winds occur in actual conditions, the wind speed forecast data is 19.99m / s and the score is 0, while the wind speed forecast data is 20.01m / s and the perfect score is 1. It can be said that the forecast levels of the two are almost the same, with a difference of only 0.02m / s, but the scores of the two are completely different, which is obviously unreasonable. Although 19.99m / s does not exceed the threshold and is not a hit in the traditional sense, such a prediction is obviously of certain reference value, and this also leads to the discreteness and discontinuity of the scoring curve. The second is a non-quantitative and non-discrete test method (RMSE, MAE, correlation coefficient, etc.). This test lacks sensitivity to different wind speeds. The deviations at low wind speeds and the errors near the high wind speed threshold are scored the same, but in reality, there is a huge difference in the attention paid to the two for disaster prevention needs (power disaster prevention does not actually pay attention to the errors at low wind speeds, but pays more attention to the errors near the high wind speed threshold). In addition, most of these two types of test indicators do not consider the double penalty in terms of location. For example, a certain place is forecast to have strong winds, but the actual location is far away from a certain place. In this way, not only will the certain place be falsely reported and get no points, but the actual location will also be missed and get no points, that is, a double "location penalty". Moreover, with the gradual refinement of the current forecast resolution, this double penalty situation is becoming more and more serious. Summary of the invention

[0006] The purpose of the present invention is to solve the deficiencies of the above-mentioned background technology and to provide a general verification method, system and storage medium for wind speed forecasting of power wind disasters.

[0007] The technical solution adopted by the present invention is: a general verification method for wind speed forecasting for power wind disasters, comprising:

[0008] The wind speed observation data and wind speed forecast data are used to calculate the single-point wind speed forecast universal test score within the wind speed prediction focus range; the wind speed forecast universal test score value set within the neighborhood is calculated; the maximum wind speed forecast universal test score within the set is determined; the disaster-causing wind speed hit or miss situation is determined based on the wind speed forecast universal test score within the neighborhood for a period of time; the spatial distribution characteristics of the position error are obtained based on the position coordinates of the maximum wind speed forecast universal test score within a period of time.

[0009] Furthermore, the wind speed prediction focus range WS is setbom ~WS top , where WS top WS is the upper limit of the wind speed forecast range; bom The lower limit of the wind speed forecast range.

[0010] Furthermore, the calculation of the single-point wind speed forecast universal verification score includes:

[0011] 1) When the wind speed observation data u <WS bom hour:

[0012] a) Wind speed forecast data x <WS bom It is not included in the inspection;

[0013] b) Wind speed forecast data x ≥ WS bom hour;

[0014]

[0015] 2) When the wind speed observation data u is located in WS bom ≤u <WS top hour:

[0016]

[0017] 3) When wind speed observation data u≥WS top hour:

[0018]

[0019] Among them, x is the wind speed forecast data; u is the wind speed observation data; Score is the general test score of wind speed forecast; w is the ratio of the false alarm defense cost divided by the accurate forecast disaster prevention value.

[0020] Furthermore, the wind speed forecast data at a certain point is set as F x , define O x Is a group of F x The wind speed observation data set of all grid points or stations {O1, O2, ..., O n}, the number of matching grid points or stations is n, according to the wind speed forecast data F x The wind speed forecast universal test score is calculated by matching with n wind speed observation data to obtain a set of wind speed forecast universal test score values ​​within the neighborhood.

[0021] Furthermore, the maximum value of the wind speed forecast universal test score Scoreneighbor in the set is determined by the following calculation formula:

[0022] Scoreneighbor=max(dw i Scorei )i=1,2,...,n

[0023] Among them, dw i is the weight coefficient of the i-th wind speed observation data; Score i is the score of the i-th wind speed observation data.

[0024] Furthermore, the number of positive values ​​of the universal test score of wind speed forecast within a neighborhood range over a period of time / the total number of samples is taken as the disaster-causing wind speed hit situation.

[0025] Furthermore, the number of negative values / total number of samples of the universal test score of wind speed forecast within a certain time neighborhood is regarded as the false alarm situation.

[0026] Furthermore, the spatial distribution characteristics of the position error are obtained by the following method:

[0027] When calculating the maximum value of multiple wind speed forecast universal test scores within a period of time, multiple longitude and latitude coordinates corresponding to the maximum value of the wind speed forecast universal test score are retained, and the average longitude and latitude deviation is obtained by calculating the average longitude and latitude as the spatial distribution characteristics of the position error.

[0028] A system for implementing the above-mentioned universal wind speed forecast verification method for power wind disasters, comprising a data acquisition module for collecting wind speed observation data and wind speed forecast data;

[0029] The scoring calculation module is used to calculate the single-point wind speed forecast universal test score within the wind speed forecast focus range using wind speed observation data and wind speed forecast data; calculate the wind speed forecast universal test score value set within the neighborhood range; and determine the maximum value of the wind speed forecast universal test score within the set;

[0030] The wind speed verification module is used to determine the disaster-causing wind speed hit or miss situation based on the universal wind speed forecast test score within a certain period of time in the neighborhood; and obtain the spatial distribution characteristics of the position error based on the position coordinates of the maximum value of the universal wind speed forecast test score within a certain period of time.

[0031] Set the wind speed forecast focus range WS bom ~WS top , where WS top WS is the upper limit of the wind speed forecast range; bom The lower limit of the wind speed forecast range.

[0032] Calculation of the single point wind speed forecast common verification score includes:

[0033] 1) When the wind speed observation data u <WS bom hour:

[0034] a) Wind speed forecast data x <WSbom It is not included in the inspection;

[0035] b) Wind speed forecast data x ≥ WS bom hour;

[0036]

[0037]

[0038] 2) When the wind speed observation data u is located in WS bom ≤u <WS top hour:

[0039]

[0040] 3) When wind speed observation data u≥WS top hour:

[0041]

[0042] Among them, x is the wind speed forecast data; u is the wind speed observation data; Score is the general test score of wind speed forecast; w is the ratio of the false alarm defense cost divided by the accurate forecast disaster prevention value.

[0043] Set the wind speed forecast data at a certain point to F x , define O x Is a group of F x The wind speed observation data set of all grid points or stations {O1, O2, ..., O n}, the number of matching grid points or stations is n, according to the wind speed forecast data F x The wind speed forecast universal test score is calculated by matching with n wind speed observation data to obtain a set of wind speed forecast universal test score values ​​within the neighborhood.

[0044] The maximum value of the wind speed forecast universal test score Scoreneighbor in the set is determined by the following calculation formula:

[0045] Scoreneighbor=max(dw i Score i )i=1,2,...,n

[0046] Among them, dw i is the weight coefficient of the i-th wind speed observation data; Score i is the score of the i-th wind speed observation data.

[0047] The number of positive values ​​of the universal test score of wind speed forecast within a neighborhood range over a period of time / the total number of samples is taken as the disaster-causing wind speed hit situation.

[0048] The number of negative values ​​of the universal test score of wind speed forecast within a certain time neighborhood / the total number of samples is regarded as the false alarm situation.

[0049] The spatial distribution characteristics of the position error are obtained by the following method:

[0050] When calculating the maximum value of multiple wind speed forecast universal test scores within a period of time, multiple longitude and latitude coordinates corresponding to the maximum value of the wind speed forecast universal test score are retained, and the average longitude and latitude deviation is obtained by calculating the average longitude and latitude as the spatial distribution characteristics of the position error.

[0051] A computer-readable storage medium stores a computer program, wherein the computer program implements the steps of any of the above methods when executed by a processor.

[0052] The present invention is to improve the application effect of wind speed forecast in the power industry. Mainly based on the sensitivity of power wind disasters to different wind speeds (only sensitive to the range near the disaster-causing wind speed), by setting a threshold value by itself, a function is constructed to be more sensitive to the error within the wind speed range of concern. Based on the actual wind speed observation value, the hyperbolic tangent function is constructed by using the e exponential function to form a bounded, unclassified, sensitive, and smooth gradient evaluation score. And based on the neighborhood method idea, by selecting a scheme according to the specific needs of the industry to complete the forecast-observation data matching, the forecast value corresponds to a set of scoring value sets of statistical schemes, according to the needs within a reasonable range to avoid double penalties on the position, while reflecting the accuracy of the forecast position. Finally, a general comprehensive inspection index is formed, and a set of characteristic indicators of wind speed forecast deviation (hit and empty report situation, position deviation index) are given. The general comprehensive inspection index can be used for comprehensive decision-making, and can also be used to improve the objective function of forecast modeling (completely for power disasters). The characteristic indicators of wind speed forecast deviation can also be used to assist decision-making and error situation analysis, which is helpful to improve the research on forecast error.

[0053] The present invention evaluates the quality of wind speed forecasts based on the needs of power wind disaster prevention. It solves the problems of the current wind speed inspection scheme (mainly from the meteorological industry) lacking pertinence and having many unreasonable problems such as discontinuous scoring and double penalties. The biggest difference between the present invention and the prior art is that it first customizes a complete set of inspection schemes completely for the inspection needs of power wind disasters, and at the same time solves and rationalizes many problems of the prior art, as follows:

[0054] 1) Compared with the prior art, the present invention fully addresses the disaster prevention needs of the power industry for strong wind disasters, and has the characteristics of localized threshold setting and scheme selection, and is highly targeted, customizable and universally applicable.

[0055] 2) The present invention does not need to adopt an interpolation scheme to unify the wind speed observation data and the wind speed prediction data, thereby avoiding the error caused by the interpolation scheme and having better adaptability to different data.

[0056] 3) The present invention avoids the problems of discontinuous scoring distortion, lack of sensitivity and double penalty in position in the traditional method. The method of the present invention is continuous and distortion-free, more sensitive to high wind speed errors, avoids double penalty in position, and also reasonably reflects the accuracy of the forecast position.

[0057] 4) The present invention takes into account the trade-off between the cost of false disaster warnings and the value of accurate disaster forecasting, and has extremely strong practical decision-making value.

[0058] 5) The present invention is simple to calculate and has clear physical practical significance. The universal comprehensive inspection index can be used for comprehensive decision-making and can also be used to improve the objective function of forecast modeling (completely for power disasters).

[0059] 6) This product ultimately forms a general comprehensive inspection index, and also provides a set of characteristic indicators for wind speed forecast deviation (hit and miss conditions, direction of position deviation, etc.). The characteristic indicators can also be used to assist decision-making and error analysis, which helps improve the research on forecast errors. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] Figure 1 A flow chart of a method provided by an embodiment of the present invention;

[0061] Figure 2 This is a schematic diagram of the distribution of the general test scores for wind speed forecast when the wind speed observation data u<10m / s and the wind speed prediction data is 10-30m / s;

[0062] Figure 3 It is a schematic diagram of the distribution of the universal inspection score of wind speed forecast data when 10m / s≤wind speed observation data u<20m / s;

[0063] Figure 4 It is a schematic diagram of the distribution of the universal inspection score of wind speed forecast data when the wind speed observation data u≥20m / s;

[0064] Figure 5 It is a schematic diagram for matching forecast points and observation points;

[0065] Figure 6 A system diagram for connecting modules of the present invention. DETAILED DESCRIPTION

[0066] The present invention will be further described in detail below in conjunction with the accompanying drawings and specific embodiments to facilitate a clear understanding of the present invention, but they do not constitute a limitation on the present invention.

[0067] Example 1

[0068] like Figure 1 As shown, the present invention provides a universal wind speed forecast verification method for power wind disasters, comprising the following steps:

[0069] 1. Obtain wind speed observation data and wind speed forecast data;

[0070] According to the actual data situation, wind speed forecast data is generally grid data, while wind speed observation data may be grid data or station data. This method is universal. The method does not need to interpolate and unify between wind speed observation data and wind speed forecast data, thus avoiding the error caused by interpolation. This method is based on the wind speed forecast data, grid by grid (or station by station), and finally averages to get the overall conclusion.

[0071] 2. Setting of wind speed disaster threshold and wind speed forecast focus range;

[0072] The wind speed above which a power disaster will occur depends on the design and wind resistance of the power system. Generally speaking, the power system is designed to cope with the local maximum wind speed according to the geographical location and climatic conditions. The wind speed limits in different regions may vary. Generally speaking, when the wind speed exceeds 20 m / s (72 km / h), the risk of power disasters may increase. However, this is only a general reference value. In fact, different power systems may have different wind speed limits, and other factors such as equipment quality, maintenance status, etc. will also be considered.

[0073] This method needs to determine the wind speed disaster threshold and the wind speed forecast focus range. The upper limit of the wind speed forecast focus range (i.e., the wind speed disaster threshold) is expressed as WS top Indicates that the lower limit is WS bom This embodiment assumes that 20m / s is the wind speed disaster threshold, and the lower limit of the forecast attention range is 10m / s. That is, in addition to the strong winds exceeding 20m / s that can cause disasters, this embodiment also defines a range of 10-20m / s close to strong winds as the forecast attention range. The specific wind speed threshold should be set according to the design and operation requirements of the local power system.

[0074] 3. Calculate the general test score for single-point wind speed forecast;

[0075] For wind speed inspection of wind disasters, the most important principle is to examine the forecast while reducing the incidence of false alarms on the basis of hitting the target as much as possible, so the main focus is on the hits and false alarms of the wind speed that caused the disaster. If the actual situation does not occur, the main focus is on the impact of false alarms, and if the actual situation occurs, the main focus is on the impact of the hit. This method is a comprehensive trade-off between the two, using the e-exponential function to construct a hyperbolic tangent function (also often expressed as tanh) to avoid sudden changes in the score and make the score smooth, continuous and not discrete.

[0076] Define x as wind speed forecast data, u as wind speed observation data, Score as wind speed forecast universal test score, w as the ratio of false alarm defense cost divided by accurate forecast disaster prevention value, w is generally between 0-1, and is set by power disaster prevention technicians. Score ranges from -w to 1, 1 is a completely correct hit, 0 to 1 is no correct hit but has a certain reference value, 0 is no impact, -w means completely false alarm, -w to 0 means no false alarm, but there is a certain misleading over-forecast. When the Score is positive, it has a positive reference value, and when the Score is negative, it has a negative reference value. The final average is used to get the overall value of the forecast. The scoring calculation scheme is as follows:

[0077] 1) When the wind speed observation data u <WS bom hour:

[0078] a) Wind speed forecast data x <WS bom When Score = Nan;

[0079] When both the wind speed observation data and the wind speed forecast data are far less than the disaster-causing wind speed, although it is a correct denial of the wind disaster and it is a high-probability time in the sample, from an application perspective, this correct forecast is almost meaningless. We do not include this sample in the test, so as to highlight the information within the scope of our wind speed forecast.

[0080] b) Wind speed forecast data x ≥ WS bom hour;

[0081]

[0082] For the sake of intuitive demonstration, assume that WS top 20m / s, WS bom When the wind speed observation data u<10m / s, the correspondence between the wind speed forecast data and the wind speed forecast universal test score is as follows: Figure 2 As shown in Figure 2, it can be seen that the score transitions smoothly, from no value to a somewhat misleading over-prediction, and finally to a completely empty negative value.

[0083] 2) When the wind speed observation data u is located in WS bom ≤u <WS top hour:

[0084]

[0085] For the sake of intuitive demonstration, assume that WS top 20m / s, WS bomis 10m / s, w=0.5. When 10m / s≤wind speed observation data u<20m / s, the correspondence between wind speed forecast data and wind speed forecast universal test score is as follows Figure 3 It can also be seen that the score transitions smoothly, from no value to a somewhat misleading over-prediction, and finally to a completely empty negative value.

[0086] 3) When wind speed observation data u≥WS top hour:

[0087]

[0088] For intuitive demonstration, assuming that the wind speed observation u≥20m / s, the correspondence between the wind speed forecast data and the wind speed forecast universal test score is as follows: Figure 4 As shown in the figure, it can be seen that the score transition is smooth, from no value to a certain reference value close to the threshold, and finally to a positive value of complete hit.

[0089] 4. Calculate the set of universal verification scores for wind speed forecast within the neighborhood;

[0090] Since most of the previous general inspection scores based on wind speed forecasts did not consider the neighborhood, there is still a double penalty for position deviation, which needs to be improved by the neighborhood method. The matching of wind speed forecast data and wind speed observation data is an important part of the neighborhood method forecast inspection method. Since the current numerical model's ability to predict strong wind speeds at fixed points and quantitatively is still very limited, for the public, the presence of some errors in the forecast is tolerable. For example, a certain place is forecast to have strong winds, but the actual location is far away from a certain place. In this way, not only is the certain place empty and no points are scored, but the actual location is also missed if it is not forecast, and no points are scored, that is, "double punishment", which is obviously not in line with the actual situation. When preventing disasters in the power industry, it is often not necessary to be accurate to each inspection grid point, and there will be a certain tolerance inspection range. Therefore, the neighborhood matching method is more appropriate. The following introduces the matching method of forecast inspection points and observation points in this article.

[0091] like Figure 5 As shown in Figure 2, the forecast observation data matching method is as follows:

[0092] According to the neighborhood matching between the forecast check point and the observation point, let the wind speed forecast data of a certain point be F x , define O x Is a group of F x The wind speed observation data set of all grid points or stations {O1, O2, ..., O n}, the number of matching grid points or stations is n, that is, the wind speed forecast data F x The wind speed forecast universal test score is calculated by matching with n wind speed observation data. Figure 5It is a schematic diagram. The observation and forecast shown in the figure are both grid files. If the forecast is a grid file, it does not matter if the observation is a site file. You can still select it within the neighborhood radius R. The neighborhood radius R here can be automatically set according to the actual disaster prevention needs of the power industry. Finally, a set of general inspection score values ​​for wind speed forecasts within the neighborhood range is obtained.

[0093] 5. Determine the maximum value of the universal verification score of wind speed forecast within the ensemble;

[0094] Since one wind speed forecast data corresponds to a set of wind speed forecast universal inspection score values, and the disaster prevention technicians in the power industry actually hope that each wind speed forecast data has only one score value, it is necessary to use statistical methods to count the scoring results. Conventional statistical indicators include the maximum value, set mean or median in the score value set. In order to strengthen the consideration of the accuracy of the forecast position, this method uses the distance weighted score maximum statistical method to determine the maximum value of the wind speed forecast universal inspection score Scoreneighbor at the forecast inspection point. Among them, Score i It is the wind speed forecast general test score Score calculated by the wind speed forecast data and each observation grid point or station within the matching range. The distance weight coefficient dw ranges from 0 to 1. When the position is at the maximum distance, dw is 0. When the position is at the position of the wind speed forecast data itself, dw is 1. Other positions correspond to values ​​between 0 and 1. The dw values ​​between 0 and 1 are linearly distributed according to the inverse distance weight. This design not only considers avoiding serious double penalties, but also reasonably reflects the accuracy value of the forecast position. In actual use, the value of dw can be set according to the actual requirements for position accuracy.

[0095] Scoreneighbor=max(dw i Score i )i=1,2,...,n;

[0096] Of course, if the actual disaster prevention needs of the power industry may not have such strict requirements on location, the location weight coefficient may not be used here. Only the survey range is counted.

[0097] Scoreneighbor=max(Score i )i=1,2,...,nScore。

[0098] 6. Set characteristic indicators;

[0099] This method calculates the universal inspection score of wind speed forecast for the wind speed forecast data that needs to be statistically analyzed at each grid point (station) within a certain period of time, and finally obtains the overall value of the forecast for a certain period of time that needs to be examined after averaging, forming a universal comprehensive inspection index.

[0100] At the same time, a set of characteristic indicators of wind speed forecast deviation can also be given during the calculation process. Different from the overall inspection score which is a comprehensive indicator, the characteristic indicators reflect the characteristics of forecast errors in various aspects.

[0101] The details are as follows:

[0102] ①. The number of positive values / total number of samples of the universal test score of wind speed forecast within a certain period of time in the neighborhood is regarded as the disaster-causing wind speed hit situation, and the number of negative values / total number of samples of the universal test score of wind speed forecast within a certain period of time in the neighborhood is regarded as the false alarm situation.

[0103] ②. According to the position coordinates of the maximum value of the universal inspection score of the wind speed forecast within a period of time, the spatial distribution characteristics of the position error are obtained. Specifically, it includes: when calculating the maximum values ​​of the universal inspection scores of wind speed forecasts for multiple times within a period of time, retaining the corresponding multiple longitude and latitude coordinates of the maximum value of the universal inspection score of the wind speed forecast, and calculating the average longitude and latitude to obtain the average longitude and latitude deviation as the spatial distribution characteristics of the position error. It can also be displayed in a two-dimensional plane using a thermal scatter plot.

[0104] These characteristic indicators can be used for comprehensive decision-making, and can also be used to improve the objective function of forecast modeling (completely for power disasters). Characteristic indicators such as wind speed forecast deviation can also be used to assist decision-making and error analysis, which will help improve the research on forecast errors.

[0105] This method eventually forms a universal comprehensive inspection index. At the same time, a set of characteristic indicators of wind speed forecast deviation can be given in the calculation process. Different from the overall inspection score, which is a comprehensive indicator, the characteristic indicator reflects the characteristics of the forecast error from various aspects. The positive and negative values ​​of the Score can be averaged and counted to obtain the disaster-causing wind speed hits and misses. The score value set can also be obtained, and the location coordinates of the maximum value can be recorded, so that the spatial distribution characteristics of the location error can be counted. The universal comprehensive inspection index can be used for comprehensive decision-making, and can also be used to improve the objective function of forecast modeling (completely for power disasters), and characteristic indicators such as wind speed forecast deviation can also be used for auxiliary decision-making and error analysis, which is helpful to improve the research on forecast errors.

[0106] Example 2

[0107] like Figure 6 As shown, a system for implementing the universal verification method of wind speed forecast for power wind disasters includes:

[0108] Data acquisition module, used to collect wind speed observation data and wind speed forecast data;

[0109] The scoring calculation module is used to calculate the single-point wind speed forecast universal test score within the wind speed forecast focus range using wind speed observation data and wind speed forecast data; calculate the wind speed forecast universal test score value set within the neighborhood range; and determine the maximum value of the wind speed forecast universal test score within the set;

[0110] The wind speed verification module is used to determine the disaster-causing wind speed hit or miss situation based on the universal wind speed forecast test score within a certain period of time in the neighborhood; and obtain the spatial distribution characteristics of the position error based on the position coordinates of the maximum value of the universal wind speed forecast test score within a certain period of time.

[0111] Set the wind speed forecast focus range WS bom ~WS top , where WS top WS is the upper limit of the wind speed forecast range; bom The lower limit of the wind speed forecast range.

[0112] Calculation of the single point wind speed forecast common verification score includes:

[0113] 1) When the wind speed observation data u <WS bom hour:

[0114] a) Wind speed forecast data x <WS bom It is not included in the inspection;

[0115] b) Wind speed forecast data x ≥ WS bom hour;

[0116]

[0117] 2) When the wind speed observation data u is located in WS bom ≤u <WS top hour:

[0118]

[0119] 3) When wind speed observation data u≥WS top hour:

[0120]

[0121] Among them, x is the wind speed forecast data; u is the wind speed observation data; Score is the general test score of wind speed forecast; w is the ratio of the false alarm defense cost divided by the accurate forecast disaster prevention value.

[0122] Set the wind speed forecast data at a certain point to F x , define O x Is a group of F x The wind speed observation data set of all grid points or stations {O1, O2, ..., On}, the number of matching grid points or stations is n, according to the wind speed forecast data F x The wind speed forecast universal test score is calculated by matching with n wind speed observation data to obtain a set of wind speed forecast universal test score values ​​within the neighborhood.

[0123] The maximum value of the wind speed forecast universal test score Scoreneighbor in the set is determined by the following calculation formula:

[0124] Scoreneighbor=max(dw i Score i )i=1,2,...,n:

[0125] Among them, dw i is the weight coefficient of the i-th wind speed observation data; Score i is the score of the i-th wind speed observation data.

[0126] The number of positive values ​​of the universal test score of wind speed forecast within a neighborhood range over a period of time / the total number of samples is taken as the disaster-causing wind speed hit situation.

[0127] The number of negative values ​​of the universal test score of wind speed forecast within a certain time neighborhood / the total number of samples is regarded as the false alarm situation.

[0128] The spatial distribution characteristics of the position error are obtained by the following method:

[0129] When calculating the maximum value of multiple wind speed forecast universal test scores within a period of time, multiple longitude and latitude coordinates corresponding to the maximum value of the wind speed forecast universal test score are retained, and the average longitude and latitude deviation is obtained by calculating the average longitude and latitude as the spatial distribution characteristics of the position error.

[0130] Example 3

[0131] The present invention also provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the above energy efficiency improvement method and embodiment are implemented. Alternatively, when the computer program is executed by a processor, the functions of each module / unit in the above energy efficiency improvement system are implemented.

[0132] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0133] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0134] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0135] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0136] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit its protection scope. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the field should understand that after reading the present invention, those skilled in the art can still make various changes, modifications or equivalent substitutions to the specific implementation methods of the invention, but these changes, modifications or equivalent substitutions are all within the protection scope of the pending claims of the invention.

[0137] The contents not described in detail in this specification belong to the prior art known to professional and technical personnel in this field.

Claims

1. A general verification method for wind speed forecasting for power wind disasters, characterized by: include Calculate the single-point wind speed forecast universal test score within the wind speed forecast focus range using wind speed observation data and wind speed forecast data; Calculate the wind speed forecast universal test score value set within the neighborhood range; Determine the maximum value of the universal verification score of wind speed forecast in the ensemble; The hit or miss situation of the disaster-causing wind speed is determined based on the universal verification score of the wind speed forecast within a neighborhood range over a period of time; the spatial distribution characteristics of the position error are obtained based on the position coordinates of the maximum value of the universal verification score of the wind speed forecast over a period of time.

2. A universal wind speed forecast verification method for power wind disasters according to claim 1, characterized in that: Set the wind speed forecast focus range WS bom ~WS top , where WS top WS is the upper limit of the wind speed forecast range; bom The lower limit of the wind speed forecast range.

3. A universal verification method for wind speed forecast for power wind disasters according to claim 2, characterized in that: Calculation of the single point wind speed forecast common verification score includes: 1) When the wind speed observation data u <WS bom hour: a) Wind speed forecast data x <WS bom It is not included in the inspection; b) Wind speed forecast data x ≥ WS bom hour; 2) When the wind speed observation data u is located in WS bom ≤u <WS top hour: 3) When wind speed observation data u≥WS top hour: Among them, x is the wind speed forecast data; u is the wind speed observation data; Score is the general test score of wind speed forecast; w is the ratio of the false alarm defense cost divided by the accurate forecast disaster prevention value.

4. A universal wind speed forecast verification method for power wind disasters according to claim 3, characterized in that: Set the wind speed forecast data at a certain point to F x , define O x Is a group of F x Taking point as the center, the set of wind speed observation data of all grid points or stations in the circle radially expanded with radius R is {O1, O2, ..., On}, the number of matching grid points or stations is n, and the universal test score of wind speed forecast is calculated by matching the wind speed forecast data Fx with the n wind speed observation data to obtain the set of universal test score values ​​of wind speed forecast in the neighborhood.

5. A universal verification method for wind speed forecast for power wind disasters according to claim 4, characterized in that: The maximum value of the wind speed forecast universal test score Scoreneighbor in the set is determined by the following calculation formula: Scoreneighbormax(dwi·Scorei)i.,n Among them, dw i is the weight coefficient of the i-th wind speed observation data; Score i is the score of the i-th wind speed observation data.

6. A universal verification method for wind speed forecast for power wind disasters according to claim 1, characterized in that: The number of positive values ​​of the universal test score of wind speed forecast within a neighborhood range over a period of time / the total number of samples is taken as the disaster-causing wind speed hit situation.

7. The universal wind speed forecast verification method for power wind disasters according to claim 1 is characterized by: The number of negative values ​​of the universal test score of wind speed forecast within a certain time neighborhood / the total number of samples is regarded as the false alarm situation.

8. The universal wind speed forecast verification method for power wind disasters according to claim 1 is characterized by: The spatial distribution characteristics of the position error are obtained by the following method: When calculating the maximum value of multiple wind speed forecast universal test scores within a period of time, multiple longitude and latitude coordinates corresponding to the maximum value of the wind speed forecast universal test score are retained, and the average longitude and latitude deviation is obtained by calculating the average longitude and latitude as the spatial distribution characteristics of the position error.

9. A system for implementing the universal verification method for wind speed forecast for power wind disasters as described in any one of claims 1 to 8, characterized in that: include Data acquisition module, used to collect wind speed observation data and wind speed forecast data; The scoring calculation module is used to calculate the single-point wind speed forecast universal test score within the wind speed forecast focus range using wind speed observation data and wind speed forecast data; calculate the wind speed forecast universal test score value set within the neighborhood range; Determine the maximum value of the universal verification score of wind speed forecast in the ensemble; The wind speed verification module is used to determine the disaster-causing wind speed hit or miss situation based on the universal wind speed forecast test score within a certain period of time in the neighborhood; and to obtain the spatial distribution characteristics of the position error based on the position coordinates of the maximum value of the universal wind speed forecast test score within a certain period of time.

10. The system of a universal verification method for wind speed forecast for power wind disasters according to claim 9, characterized in that: Set the wind speed forecast focus range WS bom ~WS top , where WS top WS is the upper limit of the wind speed forecast range; bom The lower limit of the wind speed forecast range.

11. The system of the universal verification method for wind speed forecast for power wind disaster according to claim 10, characterized in that: Calculation of the single point wind speed forecast common verification score includes: 1) When the wind speed observation data u <WS bom hour: a) Wind speed forecast data x <WS bom It is not included in the inspection; b) Wind speed forecast data x ≥ WS bom hour; 2) When the wind speed observation data u is located in WS bom ≤u <WS top hour: 3) When wind speed observation data u≥WS top hour: Among them, x is the wind speed forecast data; u is the wind speed observation data; Score is the general test score of wind speed forecast; w is the ratio of the false alarm defense cost divided by the accurate forecast disaster prevention value.

12. The system of a universal verification method for wind speed forecast for power wind disasters according to claim 11, characterized in that: Set the wind speed forecast data at a certain point to F x , define O x Is a group of F x Taking point as the center, the set of wind speed observation data of all grid points or stations in the circle radially expanded with radius R is {O1, O2, ..., On}, the number of matching grid points or stations is n, and the universal test score of wind speed forecast is calculated by matching the wind speed forecast data Fx with the n wind speed observation data to obtain the set of universal test score values ​​of wind speed forecast in the neighborhood.

13. The system of a universal verification method for wind speed forecast for power wind disasters according to claim 12, characterized in that: The maximum value of the wind speed forecast universal test score Scoreneighbor in the set is determined by the following calculation formula: Scoreneighbormax(dwi·SCorei)i:,n Among them, dw i is the weight coefficient of the i-th wind speed observation data; Score i is the score of the i-th wind speed observation data.

14. The system of a universal verification method for wind speed forecast for power wind disasters according to claim 9, characterized in that: The number of positive values ​​of the universal test score of wind speed forecast within a neighborhood range over a period of time / the total number of samples is taken as the disaster-causing wind speed hit situation.

15. The system of a universal verification method for wind speed forecast for power wind disasters according to claim 9, characterized in that: The number of negative values ​​of the universal test score of wind speed forecast within a certain time neighborhood / the total number of samples is regarded as the false alarm situation.

16. The system of a universal verification method for wind speed forecast for power wind disasters according to claim 9, characterized in that: The spatial distribution characteristics of the position error are obtained by the following method: When calculating the maximum value of multiple wind speed forecast universal test scores within a period of time, multiple longitude and latitude coordinates corresponding to the maximum value of the wind speed forecast universal test score are retained, and the average longitude and latitude deviation is obtained by calculating the average longitude and latitude as the spatial distribution characteristics of the position error.

17. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 8 are implemented.

Images