A pollution risk assessment method and system based on night light driving

Abstract

The present application provides a pollution risk assessment method and system based on night light drive, including: obtaining night light data and meteorological data of the area to be assessed; using rasterization processing to obtain raster values ​​of night light data and raster values ​​of meteorological data; then converting them into the preset pollution degree of each element and the weight assignment model of each classification; converting the preset pollution degree and weight assignment model of each element into a score model; setting the numerical risk threshold corresponding to different pollution area levels according to the score model to obtain a pollution area map. In the actual application process, the method of the present application uses night light data as the first driving factor and meteorological data as the secondary driving factor. The pollution degree of each local area of ​​the insulator is judged and scored through the optimized insulator pollution degree prediction model, and finally large-scale, rapid and accurate non-contact detection of insulators with unknown pollution degree is achieved, providing a scientific basis for the determination of insulator pollution degree.

Description

Technical Field

[0001] The present application relates to the field of electric power technology, and in particular to a pollution risk assessment method and system based on nighttime light drive. Background Art

[0002] Contamination deposited on the surface of insulators can easily cause contamination flashover. Large-scale contamination flashover power outages in power transmission and transformation equipment will seriously endanger the safety of power production, cause losses to power companies, and bring significant harm to the national economy. This has brought new challenges to the early warning and prevention of contamination accidents on the external insulation of power transmission and transformation equipment.

[0003] At present, in order to assess the probability of pollution flashover in different regions, the power systems in various provinces of my country usually draw pollution zone maps manually on a regular basis and carry out cleaning work manually according to the pollution zone maps.

[0004] The drawing of pollution area maps requires a large number of manual point measurements, which is a heavy workload, especially for some areas where transmission lines have not yet been assumed. Due to factors such as complex terrain, the risk of field operations is relatively high, data acquisition is difficult, and it is difficult to accurately judge the potential risk of pollution accumulation. Summary of the invention

[0005] In order to solve the problems of heavy workload and incomplete coverage in drawing pollution area maps of power systems, the embodiment of the present application provides a pollution risk assessment method and system based on nighttime light drive, including obtaining nighttime light data and meteorological data of the area to be assessed;

[0006] The nighttime light data and the meteorological data are subjected to rasterization processing to obtain raster values ​​of the nighttime light data and raster values ​​of the meteorological data;

[0007] Using the IsoData algorithm, the raster values ​​of the lighting data and the raster values ​​of the meteorological data are converted into the preset pollution degree of each element;

[0008] Using a rough set algorithm, the grid-type values ​​of the night light data and the grid-type values ​​of the meteorological data are converted into a weight assignment model for each classification;

[0009] Using a determination system, the preset pollution degree of each element and the weight assignment model are converted into a score model;

[0010] The numerical risk thresholds corresponding to different pollution area levels are set according to the scoring model to obtain a pollution area map.

[0011] Furthermore, the meteorological data of the area to be evaluated include atmospheric PM10 values, atmospheric NO2 values, atmospheric SO2 concentration values, wind speed values, rainfall values, CO2 emission values, NOX emission values ​​and SO2 emission values.

[0012] Furthermore, the rasterization processing is specifically: using the EBK method to implement spatial interpolation to convert the night light data and the meteorological data into raster values.

[0013] Furthermore, the weight assignment model is specifically:

[0014]

[0015] POS C (D)=U C X;

[0016] Among them, POS C (D) represents the positive domain of D on C, which is used to refer to the expressive power of set D on set C; U is a non-empty finite set of objects, which is called the universal set. X represents the object set of the category, C represents the set x∈U / I(D), where I(D) represents the inseparable relation I specified by the set D.

[0017] Furthermore, the score model is specifically:

[0018]

[0019] Where rate j represents the total score of any pixel j, weight p Represents the weight value of index p, Represents the score of pixel j on index p.

[0020] A pollution risk assessment system based on night light driving includes:

[0021] A rasterization module, for performing rasterization processing on the nighttime light data and the meteorological data to obtain rasterized values ​​of the nighttime light data and rasterized values ​​of the meteorological data;

[0022] An IsoData algorithm module, used for converting the grid-type values ​​of the lighting data and the grid-type values ​​of the meteorological data into a preset pollution degree of each element;

[0023] A rough set algorithm module, used for converting the grid-type values ​​of the night light data and the grid-type values ​​of the meteorological data into a weight assignment model for each classification;

[0024] A determination module, used for converting the preset pollution degree of each element and the weight assignment model into a score model;

[0025] The pollution area map conversion module is used to set the numerical risk thresholds corresponding to different pollution area levels in the scoring model to obtain a pollution area map.

[0026] Furthermore, the rasterization processing is specifically: using the EBK method to implement spatial interpolation to convert the night light data and the meteorological data into raster values.

[0027] Furthermore, the weight assignment model is specifically:

[0028]

[0029] POS C (D)=U C X;

[0030] Among them, POS C (D) represents the positive domain of D on C, which is used to refer to the expressive power of set D on set C; U is a non-empty finite set of objects, which is called the universal set. X represents the object set of the category, C represents the set x∈U / I(D), where I(D) represents the inseparable relation I specified by the set D.

[0031] Furthermore, the score model is specifically:

[0032]

[0033] Where rate j represents the total score of any pixel j, weight p Represents the weight value of index p, Represents the score of pixel j on index p.

[0034] It can be seen from the above technical scheme that a pollution risk assessment method and system based on night light drive includes obtaining night light data and meteorological data of the area to be assessed; rasterizing the night light data and the meteorological data to obtain raster values ​​of the night light data and the meteorological data; using the IsoData algorithm to convert the raster values ​​of the light data and the raster values ​​of the meteorological data into preset pollution degrees of each element; using the rough set algorithm to convert the raster values ​​of the night light data and the raster values ​​of the meteorological data into weight assignment models of each classification; using a judgment system to convert the preset pollution degrees of each element and the weight assignment model into a scoring model; setting numerical risk thresholds corresponding to different pollution area levels according to the scoring model to obtain a pollution area map.

[0035] In actual application, the method of the present application uses night light data as the first driving factor and meteorological data as the secondary driving factor. The degree of contamination of each local area of ​​the insulator is judged and scored through the optimized insulator contamination degree prediction model, and finally large-scale, fast and accurate non-contact detection of insulators with unknown contamination degrees is achieved. The introduction of IsoData algorithm and rough set algorithm can well assist in determining the degree of insulator contamination. The pre-judgment detection process is simple and efficient, making the cleaning of insulators more targeted, targeted and objective, and providing a scientific basis for the determination of insulator contamination degrees. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] In order to more clearly illustrate the technical solution of the present application, the drawings required for use in the embodiments are briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0037] Figure 1 It is a flowchart of a pollution risk assessment method based on night-time light driving;

[0038] Figure 2 The figure is a flow chart of a pollution risk assessment system based on night-time light drive. DETAILED DESCRIPTION

[0039] In order to make the purpose, technical solution and advantages of the embodiments of the present application clearer, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments. The components of the embodiments of the present application described and shown in the drawings here can be arranged and designed in various different configurations.

[0040] In the description of this application, it should also be noted that, unless otherwise clearly specified and limited, the term "connection" should be understood in a broad sense, for example, it can be an electrical connection or a communication connection. For ordinary technicians in this field, the specific meanings of the above terms in this application can be understood according to specific circumstances.

[0041] In addition, the terms "including" and "for" and any variations thereof are intended to cover but not exclude inclusion, for example, a product or device comprising a list of components is not necessarily limited to those components explicitly listed but may include other components not explicitly listed or inherent to such products or devices.

[0042] In order to solve the problems of heavy workload and incomplete coverage in drawing pollution area maps of power systems, see Figure 1The figure shows a flow chart of a pollution risk assessment method based on night-time light driving. A first aspect of an embodiment of the present application provides a pollution risk assessment method based on night-time light driving, comprising steps S101 to S106.

[0043] Step S101: Obtain nighttime lighting data and meteorological data of the area to be evaluated.

[0044] In step S101, in order to obtain high-resolution, large-coverage quantitative assessment results of the cumulative risk of power pollution, this application uses nighttime light data obtained by remote sensing as the primary driving factor, and takes textual pollution source point information as the secondary driving factor, and combines atmospheric PM10 values, atmospheric NO2 values, atmospheric SO2 concentration values, wind speed values, rainfall values, CO2 emission values, NOX emission values, and SO2 emission values ​​and other meteorological parameters to establish a unified multivariate parameter data set for induced insulator pollution. Since the purpose is to reflect the degree of concentration of pollution sources and the scope of pollution source influence, it is necessary to conduct necessary in-depth processing on the nighttime light data and meteorological data of the area to be evaluated, and it is necessary to process the data of pollution source points, including kernel density analysis and distance analysis; among them, the important input parameter of kernel density analysis is the search radius. When the search radius is too large, it is easy to overestimate the concentration of the distribution of pollution point sources, and vice versa, it is easy to underestimate the concentration of the distribution of pollution point sources; in one embodiment of the present application, the formula for determining the size of the search radius is specifically:

[0045]

[0046] Where h is the search radius, m d is the median distance between all points, σ ​​is the corresponding standard deviation, and n is the number of points.

[0047] In some embodiments of the present application, after the search radius is determined, the kernel density at each coordinate can be calculated according to a formula for the size of the search radius.

[0048]

[0049] Among them, K h is the kernel, a positive function with an integral value of 1, h is the smoothing parameter (bandwidth), is the search radius.

[0050] It can be seen from the formula for determining the size of the search radius that when the point distribution is determined, the kernel density value is mainly determined by Kh and h. The selection of Kh can consider a variety of different functions, such as Gaussion, uniform, triweight, Quartic, triangular, Epanechnikov, etc.; because the kernel of the point source diffusion plume model is also a Gaussion function, the Gaussion function is used as the kernel function in the embodiment of the present application, and the kernel density analysis result can approximate the point source diffusion plume model under the uniform state of the meteorological field.

[0051] In some embodiments of the present application, distance analysis is to directly calculate the distance from any pixel to the nearest emission point; the calculation result indicates the distance between different pixels and the emission source, which is expressed as the degree of influence. In some embodiments of the present application, Euclidean distance is used, and other distances can be selected, such as Euclidean distance and Mahalanobis distance.

[0052] Step S102: rasterizing the nighttime light data and the meteorological data to obtain raster values ​​of the nighttime light data and raster values ​​of the meteorological data.

[0053] In step S102, for the night light data and the meteorological data in text form, spatial interpolation is required to be used to convert them into raster data. The EBK method is actually used to implement spatial interpolation. EBK is a ground statistical interpolation method that can automatically establish an effective Kriging model.

[0054] In some embodiments of the present application, after all input attributes are rasterized, it is necessary to use spatiotemporal registration technology to unify the coordinate system, range and resolution of all input layers into a numerical raster layer.

[0055] Step S103: using the IsoData algorithm to convert the raster values ​​of the lighting data and the raster values ​​of the meteorological data into the preset pollution degree of each element.

[0056] In step S103, since different attributes have large differences in numerical values, the distribution of numerical values ​​within the same attribute also varies greatly; if a simple normalization process is adopted, it is easy to cause excessive fluctuations in scores, thereby reducing the decisiveness of some attributes.

[0057] As shown in Table 1, this is a classification level diagram of the IsoData algorithm for a pollution risk assessment method driven by night lights. In some embodiments of the present application, the IsoData automatic clustering technology is used to classify the attribute value field. The parameters that need to be specified for the IsoData algorithm are the expected number of classifications and the number of iterations; since the classification level of the pollution area map includes 4 different levels such as be, the method used in this application divides different attributes into 3-4 categories according to their respective characteristics. When using the IsoData algorithm, the expected number of classifications is set to a minimum of 5 categories and a maximum of 6 categories, and the number of iterations is selected to be 5 times. After automatic clustering, it is necessary to further observe the classification situation and merge the 5-6 subclasses into 3-4 categories according to the class statistical size.

[0058] Table 1 is a classification level diagram of the IsoData algorithm based on a night-time light-driven pollution risk assessment method.

[0059] Variable Name Classification score Night Lights High, Higher, Medium, Low 1，0.6，0.3，0 Pollution point source nuclear density High, Higher, Medium, Low 1，0.6，0.3，0 Distance from pollution point source High, Higher, Medium, Low 1，0.6，0.3，0 Rainfall High, Medium, Low 1，0.6，0.3 Wind speed High, Medium, Low 1，0.6，0.3 NDVI High, Higher, Medium, Low 1，0.6，0.3，0 CO2 emissions High, Medium, Low 1，0.6，0.3 PM10 concentration High, Medium, Low 1，0.6，0.3 NO2 concentration High, Higher, Medium, Low 1，0.6，0.3，0 SO2 concentration High, Higher, Medium, Low 1，0.6，0.3，0

[0060] Step S104: using a rough set algorithm to convert the grid-type values ​​of the nighttime light data and the grid-type values ​​of the meteorological data into a weight assignment model for each classification.

[0061] In step S104, specifically, since different factors have different influences on pollution accumulation risk, after determining the preset pollution degree of each element within each factor, it is necessary to further specify the weight scores of different factors.

[0062] In some embodiments of this application, rough set theory is used to analyze the contribution of each input variable factor to the accumulation of pollution. Since rough set theory is a new mathematical tool for dealing with uncertainty. For attribute set D, if it is completely dependent on attribute set C, it is recorded as I(C)∈I(D). If D depends on C with degree k (0≤k≤1), it is recorded as:

[0063]

[0064] in

[0065] POS C (D)=U C X;

[0066] Among them, POS C (D) represents the positive domain of D on C, which is used to refer to the expressive power of set D on set C. The larger this value is, the stronger the ability of D to express C. U is a non-empty finite set of objects, which is called the universal set. X represents the set of objects in the category. C represents the set x∈U / I(D), where I(D) represents the inseparable relation I specified by the set D.

[0067] Since rough sets can select the most decisive variables from a large number of input variables to form a simplified set, when any variable is removed from the simplified set, the POS will decrease, so 1-POS can be used to quantify the importance of different variables.

[0068] Step S104: using a determination system to convert the preset pollution degree of each element and the weight assignment model into a score model.

[0069] In step S104, specifically, in some embodiments of the present application, an appropriate amount of e, d, c, and b samples are selected from the existing dirty area map as training sample sets. Since the probability of e appearing is relatively small, usually <3%, when the level transitions from e to b, the probability of appearing increases. Therefore, when preparing training sample sets, attention should be paid to controlling the proportions of each different sample set. The method of some embodiments of the present application uses the same sample ratio of 1:1:1:1 to collect samples of different categories. This method can ensure that the rough set model has similar adaptability to different levels when it is established.

[0070] In some embodiments of the present application, after the rough set determines the score and ranking of the importance of each variable, it can be further appropriately adjusted according to the judgment system. Taking a certain area as an example, the indicators and weights involved in the final pollution risk assessment are determined as follows: night light (0.4), pollution source nuclear density (0.3), NDVI (-0.2), pollution source distance (0.15), CO2 emissions (0.15), NO2 concentration (0.1), SO2 concentration (0.1), PM10 concentration (0.1) rainfall (-0.05) and wind speed (-0.05).

[0071] Step S105: setting numerical risk thresholds corresponding to different pollution area levels according to the scoring model to obtain a pollution area map.

[0072] In step S105, the score model is obtained according to the preset pollution degree of each element and the weight assignment model, which is specifically:

[0073]

[0074] Among them, rate j represents the total score of any pixel j, weight p Represents the weight value of index p, Represents the score of pixel j on index p.

[0075] After obtaining the numerical risk assessment results, further classification processing can be performed to directly obtain the pollution area classification results. This application partially implements two methods to complete the qualitative process of the numerical risk assessment results. The first method is to set the threshold in proportion. It can be artificially stipulated that categories e, d, c, and b account for 5%, 10%, 20%, and 65% of the total, respectively, so as to determine the thresholds of different categories. The second method is to use K-MEANS automatic clustering, requiring the classifier to distinguish 4-5 categories, and perform post-classification processing to adjust the final classification results to 4 categories, thereby corresponding to the classification standard of the pollution area map.

[0076] In the embodiment of the present application, a pollution risk assessment method based on night light drive is used to obtain and use night light data as the first driving factor, and meteorological data as the secondary driving factor. The pollution degree of each local area of ​​the insulator is judged and scored through the optimized insulator pollution degree prediction model, and finally large-scale, fast and accurate non-contact detection of insulators with unknown pollution degree is achieved. The introduction of IsoData algorithm and rough set algorithm can well assist in determining the pollution degree of insulators. The pre-judgment detection process is simple and efficient, making the cleaning work of insulators more targeted, targeted and objective, and providing a scientific basis for the determination of insulator pollution degree.

[0077] In order to realize the practical application of the above method, the second aspect of the embodiment of the present application also provides a pollution risk assessment system based on night light drive, and the pollution risk assessment system based on night light drive is used to execute the pollution risk assessment method based on night light drive described in the first aspect of the embodiment of the present application. For details not disclosed in the second aspect of the embodiment of the present application, please refer to the technical solution provided in the first aspect of the embodiment of the present application.

[0078] See also Figure 2 The figure is a flow chart of a pollution risk assessment system based on night-time light drive.

[0079] A pollution risk assessment system based on night light driving includes:

[0080] A rasterization module is used to perform rasterization processing on the night light data and the meteorological data to obtain raster values ​​of the night light data and the meteorological data; an IsoData algorithm module is used to convert the raster values ​​of the light data and the meteorological data into preset pollution degrees of each element; a rough set algorithm module is used to convert the raster values ​​of the night light data and the meteorological data into weight assignment models of each classification; a determination module is used to convert the preset pollution degrees of each element and the weight assignment model into a scoring model; a pollution area map conversion module is used to set the scoring model to numerical risk thresholds corresponding to different pollution area levels to obtain a pollution area map.

[0081] In some embodiments of the present application, the rasterization processing is specifically: using the EBK method to implement spatial interpolation to convert the night light data and the meteorological data into raster values.

[0082] In some embodiments of the present application, the weight assignment model is specifically:

[0083]

[0084] POS C (D)=U C X;

[0085] Among them, POS C (D) represents the positive domain of D on C, which is used to refer to the expressive power of set D on set C; U is a non-empty finite set of objects, which is called the universal set. X represents the object set of the category, C represents the set x∈U / I(D), where I(D) represents the inseparable relation I specified by the set D.

[0086] In some embodiments of this application, the score model is specifically:

[0087]

[0088] Where rate j represents the total score of any pixel j, weight p Represents the weight value of index p, Represents the score of pixel j on index p.

[0089] From the above technical scheme, it can be seen that a pollution risk assessment method based on night light drive is characterized in that it includes obtaining night light data and meteorological data of the area to be assessed; rasterizing the night light data and the meteorological data to obtain raster values ​​of the night light data and the meteorological data; using the IsoData algorithm to convert the raster values ​​of the light data and the raster values ​​of the meteorological data into preset pollution degrees of each element; using the rough set algorithm to convert the raster values ​​of the night light data and the raster values ​​of the meteorological data into weight assignment models of each classification; using a judgment module to convert the preset pollution degrees of each element and the weight assignment model into a scoring model; setting numerical risk thresholds corresponding to different pollution area levels according to the scoring model to obtain a pollution area map.

[0090] In actual application, the method of the present application uses night light data as the first driving factor and meteorological data as the secondary driving factor. The degree of contamination of each local area of ​​the insulator is judged and scored through the optimized insulator contamination degree prediction model, and finally large-scale, fast and accurate non-contact detection of insulators with unknown contamination degrees is achieved. The introduction of IsoData algorithm and rough set algorithm can well assist in determining the degree of insulator contamination. The pre-judgment detection process is simple and efficient, making the cleaning of insulators more targeted, targeted and objective, and providing a scientific basis for the determination of insulator contamination degrees.

[0091] The present application is described in detail above in conjunction with specific implementation methods and exemplary examples, but these descriptions cannot be understood as limiting the present application. Those skilled in the art understand that, without departing from the spirit and scope of the present application, various equivalent substitutions, modifications or improvements can be made to the technical solutions and implementation methods of the present application, all of which fall within the scope of the present application.

Claims

1. A pollution risk assessment method based on nighttime light driving, characterized in that: include: Obtain nighttime light data and meteorological data for the area to be assessed; Get the size of the search radius of the area to be evaluated. The formula is: Among them, h is the search radius, m d is the median distance between all points, σ ​​is the corresponding standard deviation, and n is the number of points; The kernel density at each coordinate is calculated based on the formula of the search radius: Among them, K h is the kernel, a positive function with an integral value of 1, h is the smoothing parameter, is the search radius; The nighttime light data and the meteorological data are subjected to rasterization processing to obtain raster values ​​of the nighttime light data and raster values ​​of the meteorological data; Using the IsoData algorithm, the raster values ​​of the lighting data and the raster values ​​of the meteorological data are converted into the preset pollution degree of each element; Using a rough set algorithm, the grid-type values ​​of the night light data and the grid-type values ​​of the meteorological data are converted into a weight assignment model for each classification; The weight assignment model is specifically: POS C (D)=U C X; Among them, POS C (D) represents the positive domain of D on C, which is used to refer to the expressive power of set D on set C; U is a non-empty finite set of objects, which is called the universal set; X represents the object set of the category, C represents the set x∈U / I(D), where I(D) represents the inseparable relation I specified by the set D; Using a determination system, the preset pollution degree of each element and the weight assignment model are converted into a score model; The numerical risk thresholds corresponding to different pollution area levels are set according to the scoring model to obtain a pollution area map.

2. According to claim 1, a pollution risk assessment method based on nighttime light driving is characterized in that: The meteorological data of the area to be evaluated include atmospheric PM10 value, atmospheric NO2 value, atmospheric SO2 concentration value, wind speed value, rainfall value, CO2 emission value, NOX emission value and SO2 emission value.

3. According to claim 1, a pollution risk assessment method based on nighttime light driving is characterized in that: The rasterization process specifically includes: using the EBK method to implement spatial interpolation to convert the nighttime light data and the meteorological data into raster values.

4. The pollution risk assessment method based on nighttime light driving according to claim 1 is characterized in that: The score model is specifically: Where rate j represents the total score of any pixel j, weight p Represents the weight value of index p, Represents the score of pixel j on index p.

5. A pollution risk assessment system based on nighttime light drive, characterized in that: include: A data acquisition module is used to obtain nighttime light data and meteorological data of the area to be evaluated; Get the size of the search radius of the area to be evaluated. The formula is: Among them, h is the search radius, m d is the median distance between all points, σ ​​is the corresponding standard deviation, and n is the number of points; The kernel density at each coordinate is calculated based on the formula of the search radius: Among them, K h is the kernel, a positive function with an integral value of 1, h is the smoothing parameter, is the search radius; A rasterization module, for performing rasterization processing on the nighttime light data and the meteorological data to obtain rasterized values ​​of the nighttime light data and rasterized values ​​of the meteorological data; An IsoData algorithm module, used for converting the grid-type values ​​of the lighting data and the grid-type values ​​of the meteorological data into a preset pollution degree of each element; A rough set algorithm module, used for converting the grid-type values ​​of the night light data and the grid-type values ​​of the meteorological data into a weight assignment model for each classification; The weight assignment model is specifically: POS C (D)=U C X; Among them, POS C (D) represents the positive domain of D on C, which is used to refer to the expressive power of set D on set C; U is a non-empty finite set of objects, which is called the universal set; X represents the object set of the category, C represents the set x∈U / I(D), where I(D) represents the inseparable relation I specified by the set D; A determination module, used for converting the preset pollution degree of each element and the weight assignment model into a score model; The pollution area map conversion module is used to set the numerical risk thresholds corresponding to different pollution area levels in the scoring model to obtain a pollution area map.

6. A pollution risk assessment system based on nighttime light driving according to claim 5, characterized in that: The rasterization processing specifically includes: using the EBK method to implement spatial interpolation to convert the night light data and the meteorological data into raster values.

7. The pollution risk assessment system based on nighttime light driving according to claim 5 is characterized in that: The score model is specifically: Where rate j represents the total score of any pixel j, weight p Represents the weight value of index p, Represents the score of pixel j on index p.

Images