Mountain fire trip alarm method and system based on Bayesian network, and medium

Abstract

The invention discloses a forest fire tripping alarm method and system based on a Bayesian network, and a medium. The method comprises the following steps: respectively obtaining coordinate information of an ignition point, geographic feature information near the ignition point, meteorological information and nearby line information; based on the coordinate information of the ignition point, the nearby geographic feature information, the meteorological information and the nearby line information, calculating the forest fire spreading speed of the target area, and based on the forest fire spreading speed, predicting the forest fire height of the target area; and inputting the mountain fire spreading speed and the mountain fire height into a preset Bayesian network to obtain a mountain fire tripping risk of the power transmission line in the target area, and issuing corresponding risk alarm information based on the mountain fire tripping risk. According to the method, the real-time meteorological condition and the line underlying surface vegetation condition are considered, and accurate alarm of power transmission line forest fire tripping can be realized.

Description

Technical Field

[0001] The present application relates to the technical field of disaster prevention and reduction, and in particular to a mountain fire trip alarm method, system and medium based on a Bayesian network. Background Art

[0002] In order to achieve optimal allocation of power resources, transmission lines inevitably pass through forest areas or wastelands and other areas prone to wildfires. In recent years, affected by extreme weather, surface environment and fire habits, wildfires around transmission lines have become more frequent. Once a wildfire occurs, the insulation strength of the air gap of the transmission line will be reduced, inducing transmission line tripping, and even endangering the safe and stable operation of the power grid.

[0003] Currently, the alarm level for power transmission line tripping caused by wildfires is mainly issued based on the distance between the wildfire location monitored by satellite and the adjacent transmission lines. However, this alarm does not take into account the real-time meteorological conditions and the vegetation conditions under the line surface. It is impossible to determine whether the intensity of the flame will cause the line to trip when the wildfire burns under the line, making it difficult to achieve accurate alarm for transmission line tripping caused by wildfires.

[0004] Application Contents

[0005] The present application provides a wildfire tripping alarm method, system and medium based on a Bayesian network, which takes into account real-time meteorological conditions and vegetation conditions on the line's underlying surface to achieve accurate alarm of wildfire tripping of transmission lines.

[0006] In a first aspect, the present application provides a mountain fire trip alarm method based on a Bayesian network, comprising:

[0007] Obtaining the coordinate information of the fire point and the geographical feature information, meteorological information and nearby line information near the fire point respectively;

[0008] Calculate the wildfire spread speed in the target area based on the coordinate information of the fire point, the nearby geographical feature information, the meteorological information and the nearby line information, and predict the wildfire height in the target area based on the wildfire spread speed;

[0009] The wildfire spreading speed and the wildfire height are input into a preset Bayesian network to obtain the wildfire tripping risk of the power transmission lines in the target area, and corresponding risk warning information is issued based on the wildfire tripping risk.

[0010] The embodiment of the present application can directly obtain the information of the fire point and the vicinity of the fire point by respectively obtaining the coordinate information of the fire point and the geographical feature information, meteorological information and nearby line information near the fire point, so as to facilitate the subsequent accurate calculation of the wildfire spread speed; the wildfire spread speed of the target area is calculated based on the coordinate information of the fire point, the nearby geographical feature information, meteorological information and nearby line information, and the wildfire height of the target area is predicted based on the wildfire spread speed, so that the wildfire spread speed and the wildfire height of the target area under the conditions of wildfire occurrence can be accurately calculated, so as to facilitate the subsequent accurate prediction of the wildfire alarm risk by the Bayesian network; finally, the wildfire spread speed and the wildfire height are input into the Bayesian network, so as to accurately predict the wildfire tripping risk of the transmission line in the target area, and the real-time meteorological conditions and the vegetation conditions of the line underlying the line can be considered to realize the accurate alarm of the wildfire tripping of the transmission line.

[0011] Furthermore, the coordinate information of the ignition points is obtained respectively, specifically:

[0012] Use meteorological satellites to preliminarily locate the initial coordinate information of the fire point, and send the initial coordinate information of the fire point to the wildfire monitoring center;

[0013] The wildfire monitoring center calls the monitoring image near the initial coordinate information to determine the coordinate information of the fire point, wherein the monitoring image is obtained from the monitoring equipment installed on the transmission line or transmission tower.

[0014] In this way, the initial coordinates of the fire point are preliminarily located by means of a meteorological satellite, and monitoring images near the initial coordinate information are called to accurately determine the coordinate information of the fire point.

[0015] Furthermore, before calculating the wildfire spreading speed in the target area based on the coordinate information of the fire point, the nearby geographical feature information, the meteorological information and the nearby line information, the method further includes:

[0016] Based on the coordinate information of the fire point, determining whether there is a power transmission line or a power transmission tower within a preset range;

[0017] If there are transmission lines or transmission towers, calculate the wildfire spread rate in the area with transmission lines or transmission towers.

[0018] In this way, by judging whether there are transmission lines or transmission towers near the fire point, if there are transmission lines or transmission towers near the fire point, there is a risk of wildfire tripping. It is necessary to further calculate the wildfire spread speed in the area with transmission lines or transmission towers to help accurately predict the wildfire tripping risk of transmission lines in the future.

[0019] Furthermore, the fire spread speed in the target area is calculated based on the coordinate information of the fire point, the nearby geographical feature information, the meteorological information and the nearby line information, specifically:

[0020] Determine the initial spread rate of the wildfire based on the meteorological information;

[0021] Obtaining a combustible configuration pattern coefficient, a slope correction coefficient, and a wind speed correction coefficient, wherein the combustible configuration pattern coefficient is determined by the nearby geographical feature information, the slope correction coefficient is determined by the slope between the coordinate information of the ignition point and the nearby line information, and the wind speed correction coefficient is determined by the meteorological information;

[0022] Based on the initial spread speed, the combustible material configuration pattern coefficient, the slope correction coefficient and the wind speed correction coefficient, the wildfire spread speed in the target area is obtained.

[0023] In this way, by comprehensively considering the coordinate information of the fire point, the nearby geographical feature information, the meteorological information and the nearby line information, the wildfire spread speed in the target area under the conditions of wildfire occurrence can be accurately calculated, which facilitates the subsequent Bayesian network to accurately predict the wildfire warning risk.

[0024] Furthermore, the wildfire height in the target area is predicted based on the wildfire spread speed, specifically:

[0025] Based on the geographic feature information of the target area, determine the vegetation type in the target area, and obtain the unit mass calorific value and combustible load corresponding to the vegetation type;

[0026] Determine the intensity of the fire line in the target area based on the calorific value per unit mass of the combustibles, the combustible load and the wildfire spread speed;

[0027] Based on the fire line intensity and the regional correction coefficient corresponding to the target area, the wildfire height in the target area is predicted.

[0028] In this way, by considering the geographical feature information of the target area and the speed of volcanic spread, the real-time meteorological conditions and the vegetation conditions of the line underlying the surface can be considered, which will facilitate the subsequent accurate alarm of transmission line wildfire tripping through the Bayesian network.

[0029] Furthermore, the training process of the Bayesian network is specifically as follows:

[0030] Obtaining influencing factors of mountain fire warning point data and non-mountain fire warning point data respectively, discretizing each of the influencing factors, and randomly extracting mountain fire warning point samples and non-mountain fire warning point samples from the processing results to form a training set;

[0031] Based on the sample distribution of the training set, the conditional probability of each influencing factor under the fire condition and the no-fire condition is estimated;

[0032] Based on the conditional probability of each influencing factor under fire conditions and no fire conditions, the alarm probability and non-alarm probability of each influencing factor are obtained respectively;

[0033] The alarm probability and non-alarm probability of each influencing factor are normalized to determine the wildfire tripping risk of the transmission line.

[0034] In this way, by training the Bayesian network in advance, accurate warnings of power line tripping caused by wildfires can be achieved.

[0035] In a second aspect, the present application provides a mountain fire trip alarm system based on a Bayesian network, comprising: an acquisition module, a calculation module and an alarm module;

[0036] The acquisition module is used to respectively acquire the coordinate information of the fire point and the geographical feature information, meteorological information and nearby line information near the fire point;

[0037] The calculation module is used to calculate the wildfire spread speed in the target area based on the coordinate information of the fire point, the nearby geographical feature information, the meteorological information and the nearby line information, and predict the wildfire height in the target area based on the wildfire spread speed;

[0038] The alarm module is used to input the wildfire spread speed and the wildfire height into a preset Bayesian network to obtain the wildfire tripping risk of the transmission line in the target area, and issue corresponding risk alarm information based on the wildfire tripping risk.

[0039] The embodiment of the present application can directly obtain the information of the fire point and the vicinity of the fire point by respectively obtaining the coordinate information of the fire point and the geographical feature information, meteorological information and nearby line information near the fire point, so as to facilitate the subsequent accurate calculation of the wildfire spread speed; the wildfire spread speed of the target area is calculated based on the coordinate information of the fire point, the nearby geographical feature information, meteorological information and nearby line information, and the wildfire height of the target area is predicted based on the wildfire spread speed, so that the wildfire spread speed and the wildfire height of the target area under the conditions of wildfire occurrence can be accurately calculated, so as to facilitate the subsequent accurate prediction of the wildfire alarm risk by the Bayesian network; finally, the wildfire spread speed and the wildfire height are input into the Bayesian network, so as to accurately predict the wildfire tripping risk of the transmission line in the target area, and the real-time meteorological conditions and the vegetation conditions of the line underlying the line can be considered to realize the accurate alarm of the wildfire tripping of the transmission line.

[0040] Further, the acquisition module includes: a first acquisition unit and a second acquisition unit;

[0041] The first acquisition unit is used to preliminarily locate the initial coordinate information of the fire point using a meteorological satellite, and send the initial coordinate information of the fire point to a mountain fire monitoring center;

[0042] The second acquisition unit is used to call the monitoring image near the initial coordinate information through the wildfire monitoring center to determine the coordinate information of the fire point, wherein the monitoring image is obtained from the monitoring equipment installed on the transmission line or transmission tower.

[0043] Furthermore, the mountain fire trip alarm system based on the Bayesian network further includes: a judgment module;

[0044] The judgment module includes: a judgment unit and a first calculation unit;

[0045] The judgment unit is used to determine whether there is a power transmission line or a power transmission tower within a preset range based on the coordinate information of the fire point;

[0046] The first calculation unit is used to calculate the wildfire spread speed in the area with the transmission line or the transmission tower if there is a transmission line or a transmission tower.

[0047] Further, the calculation module includes: a second calculation unit, a third acquisition unit and a third calculation unit;

[0048] The second calculation unit is used to determine the initial spread speed of the wildfire based on the meteorological information;

[0049] The third acquisition unit is used to acquire a combustible configuration pattern coefficient, a slope correction coefficient and a wind speed correction coefficient, wherein the combustible configuration pattern coefficient is determined by the nearby geographical feature information, the slope correction coefficient is determined by the slope between the coordinate information of the ignition point and the nearby line information, and the wind speed correction coefficient is determined by the meteorological information;

[0050] The third calculation unit is used to obtain the wildfire spread speed in the target area based on the initial spread speed, the combustible material configuration pattern coefficient, the slope correction coefficient and the wind speed correction coefficient.

[0051] Furthermore, the calculation module further includes: a fourth acquisition unit, a fourth calculation unit and a fifth calculation unit:

[0052] The fourth acquisition unit is used to determine the vegetation type in the target area based on the geographical feature information of the target area, and obtain the unit mass calorific value and the combustible load corresponding to the vegetation type;

[0053] The fourth calculation unit is used to determine the fire line intensity of the target area based on the calorific value per unit mass of the combustible material, the combustible material load and the wildfire spread speed;

[0054] The fifth calculation unit is used to predict the height of the wildfire in the target area based on the intensity of the fire line and the regional correction coefficient corresponding to the target area.

[0055] Furthermore, the training process of the Bayesian network is specifically as follows:

[0056] Obtaining influencing factors of mountain fire warning point data and non-mountain fire warning point data respectively, discretizing each of the influencing factors, and randomly extracting mountain fire warning point samples and non-mountain fire warning point samples from the processing results to form a training set;

[0057] Based on the sample distribution of the training set, the conditional probability of each influencing factor under the fire condition and the no-fire condition is estimated;

[0058] Based on the conditional probability of each influencing factor under fire conditions and no fire conditions, the alarm probability and non-alarm probability of each influencing factor are obtained respectively;

[0059] The alarm probability and non-alarm probability of each influencing factor are normalized to determine the wildfire tripping risk of the transmission line.

[0060] In a third aspect, the present application provides a computer-readable storage medium, characterized in that the computer-readable storage medium includes a stored computer program, wherein when the computer program is running, the device where the computer-readable storage medium is located is controlled to execute a wildfire tripping alarm method based on a Bayesian network as described in the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] Figure 1 It is a flow chart of an embodiment of a mountain fire trip alarm method based on a Bayesian network provided by the present application;

[0062] Figure 2 This is a schematic diagram of the structure of the Bayesian network provided by this application;

[0063] Figure 3 It is a structural schematic diagram of an embodiment of a mountain fire trip alarm system based on a Bayesian network provided in the present application. DETAILED DESCRIPTION

[0064] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0065] It should be understood that the step numbers used in this article are only for the convenience of description and are not intended to limit the order in which the steps are executed.

[0066] It should be understood that the terms used in the present specification are only for the purpose of describing specific embodiments and are not intended to limit the present invention. As used in the present specification and the appended claims, the singular forms "a", "an" and "the" are intended to include plural forms unless the context clearly indicates otherwise.

[0067] The terms “include” and “comprising” indicate the presence of described features, integers, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or combinations thereof.

[0068] The term "and / or" means and includes any and all possible combinations of one or more of the associated listed items.

[0069] In order to achieve optimal allocation of power resources, transmission lines inevitably pass through forests and wastelands and other areas prone to wildfires. In recent years, affected by extreme climate, surface environment and fire habits, wildfires around transmission lines have become more frequent. Once a wildfire occurs, the insulation strength of the air gap of the transmission line will be reduced, which can easily cause tripping and endanger the safety of the power grid. The existing wildfire tripping alarm system is only based on satellite monitoring of the distance between the wildfire and the line, ignoring the real-time meteorological and vegetation conditions. It is difficult to accurately judge whether the intensity of the flame will cause the line to trip, so it is difficult to achieve accurate alarm.

[0070] Embodiment 1

[0071] Please refer to Figure 1 , Figure 1 It is a flow chart of an embodiment of a mountain fire trip alarm method based on a Bayesian network provided by the present application, including steps S1 to S3;

[0072] Step S1, respectively obtaining the coordinate information of the fire point and the geographical feature information, meteorological information and nearby line information near the fire point;

[0073] It is understandable that the initial coordinate information of the fire point can be preliminarily located by using meteorological satellites, and the initial coordinate information of the fire point can be sent to the mountain fire monitoring center. If monitoring equipment is installed on the transmission line or transmission tower near the fire point, the monitoring image near the initial coordinate information can be called by the mountain fire monitoring center to accurately determine the coordinate information of the fire point. Of course, the coordinate position of the fire point can also be directly and accurately determined by meteorological satellites through hot spot identification, or the position of the fire point can be determined directly by monitoring equipment installed on the transmission line or transmission tower from the position of the fire point and the distance between the position of the fire point and the transmission line or transmission tower without using meteorological satellites.

[0074] In this way, the initial coordinates of the fire point are preliminarily located by means of a meteorological satellite, and monitoring images near the initial coordinate information are called to accurately determine the coordinate information of the fire point.

[0075] After obtaining the coordinate information of the fire point, the basic database in the target area can be read to obtain the geographical feature information near the fire point, including geographic elevation data (DEM), vegetation type, land use type, NDVI vegetation index and combustible load, etc.; in addition, it is also necessary to read the meteorological station in the target area to obtain the meteorological information near the fire point, including wind speed, wind direction, temperature, relative humidity, etc.; at the same time, it is also necessary to read the transmission line ledger data to obtain the line information near the fire point, including voltage level, installation height, line ground clearance distance and line foundation tower coordinates, etc. After obtaining the geographical feature information, meteorological information and nearby line information near the fire point, it is possible to import the human factor data layer such as the historical fire point density, distance from the road, distance from the settlement, GDP and population density data in the study area, and grid all the data.

[0076] It should be noted that if the source data resolution is lower than the grid size resolution, an interpolation algorithm is used to divide it into corresponding grids; if the source data resolution is lower than the grid size resolution, an average aggregation method is used to merge them. The grid resolution can usually be a grid accuracy of 1000m×1000m, or 250m×250m or 500m×500m can be selected according to actual conditions. How to grid the grid is not the focus of this application and will not be expanded here.

[0077] After obtaining the geographical feature information, meteorological information and nearby line information near the fire point, it is necessary to determine whether there are transmission lines or transmission towers within the preset range based on the coordinate information of the fire point; if there are no transmission lines or transmission towers, it is considered that the fire point is temporarily not a threat to the power grid system, and the alarm system is exited to maintain wildfire monitoring. If there are transmission lines or transmission towers, the wildfire spread speed in the area with transmission lines or transmission towers is calculated. At this time, if the coordinate information of the fire point is (x, y), if the coordinate information of the transmission tower in the system is (x1, y1), by calculating the distance d between the fire point and the transmission tower, the relevant formula for calculating the distance is:

[0078]

[0079] in, dx=|x2-x1|,dy=|y2-y1|, when the calculated distance d between the fire point and the transmission tower is less than the preset range, it is considered that there is a transmission tower near the fire point, indicating that the fire point poses a threat to the power grid system, and it is necessary to further calculate the wildfire spread speed in the area with transmission lines or transmission towers.

[0080] It should be noted that the preset range can be freely set according to flame intensity, geographical feature information, meteorological information, etc., and this application does not impose any restrictions.

[0081] In this way, by judging whether there are transmission lines or transmission towers near the fire point, if there are transmission lines or transmission towers near the fire point, there is a risk of wildfire tripping. It is necessary to further calculate the wildfire spread speed in the area with transmission lines or transmission towers to help accurately predict the wildfire tripping risk of transmission lines in the future.

[0082] Step S2, calculating the wildfire spread speed in the target area based on the coordinate information of the fire point, the nearby geographical feature information, the meteorological information and the nearby line information, and predicting the wildfire height in the target area based on the wildfire spread speed;

[0083] It is understandable that, first, when the fire point is located near a flat land with uniformly distributed vegetation, the fire point will only be affected by meteorological information. Therefore, it is necessary to first determine the initial spread speed of the wildfire based on the acquired meteorological information such as temperature and wind speed, wherein the calculation formula for the initial spread speed R0 of the wildfire is:

[0084] R0=aT+bV+c(100-H)-d;

[0085] Where R0 is the initial spread speed of wildfire; T is the ambient temperature; V is the average wind level, which is 0 when there is no wind; H is the location elevation, and a, b, c and d are all constant coefficients (a, b, c and d can be taken as a=0.03, b=0.03, c=0.03, d=0.3).

[0086] Secondly, when there is vegetation and slope near the fire point, it is necessary to obtain the combustible configuration pattern coefficient, slope correction coefficient and wind speed correction coefficient, wherein the combustible configuration pattern coefficient is determined by the nearby geographical feature information, and the combustible configuration pattern coefficient K s The specific values ​​are shown in the following table:

[0087] Table 1 Combustible material configuration pattern coefficient table K s

[0088] Vegetation Type <![CDATA[K s ]]> cedar 0.8 Thatch, weeds 1.6 Straw 0.6 Secondary forest 0.7 Coniferous forest 0.4 Flat needle leaves 0.8 Dead branches and leaves 1.2 Sedge, Birch 1.8 Pasture grassland 2 Woodlands of Korean pine, Chinese pine, Yunnan pine, etc. 1

[0089] The slope correction coefficient is determined by the slope between the coordinate information of the ignition point and the nearby line information, wherein the calculation formula of the slope correction coefficient is K θ for:

[0090]

[0091] The wind speed correction coefficient is determined by the meteorological information of wind force level V; wherein the wind speed correction coefficient K v The calculation formula is:

[0092] K v =e 0.1783V ;

[0093] Finally, based on the initial spread speed, the combustible configuration pattern coefficient, the slope correction coefficient and the wind speed correction coefficient, the wildfire spread speed in the target area is obtained, wherein the wildfire spread speed is calculated as follows:

[0094] R=R0K s K θ K v ;

[0095] Where R is the wildfire spread speed; R0 is the initial wildfire spread speed; K s K is the combustible configuration pattern coefficient; θ is the slope correction coefficient; K v is the wind speed correction factor.

[0096] In this way, by comprehensively considering the coordinate information of the fire point, the nearby geographical feature information, the meteorological information and the nearby line information, the wildfire spread speed in the target area under the conditions of wildfire occurrence can be accurately calculated, which facilitates the subsequent Bayesian network to accurately predict the wildfire warning risk.

[0097] After determining the wildfire spread speed, the wildfire height in the target area can be predicted. Specifically, first, based on the geographic feature information of the target area, determine the vegetation type in the target area, and obtain the unit mass combustible calorific value and combustible load corresponding to the vegetation type; for example, based on the geographic feature information of the target area, it can be known that there are only coniferous forests in the target area, then according to the following table, it can be known that the corresponding unit mass combustible calorific value is 19620 kJ / kg. The value range of the unit mass combustible calorific value q is shown in the following table:

[0098] Table 2 Reference range of calorific value q per unit mass of combustibles

[0099]

[0100] Secondly, based on the calorific value per unit mass of the combustibles, the combustible load and the wildfire spread speed, the fire line intensity of the target area is determined, wherein the calculation formula for the fire line intensity is:

[0101] I = qWR / 600;

[0102] In the formula, I represents the intensity of the fire line, kW / m; q represents the calorific value of the combustible material per unit mass, kJ / kg; W represents the combustible material load, t / hm2; R represents the flame spread speed, m / min.

[0103] Finally, based on the fire line intensity and the regional correction coefficient corresponding to the target area, the wildfire height of the target area is predicted, where the wildfire height H f The calculation formula is:

[0104] H f =γ(I / 250) 1 / 2 ;

[0105] In the formula, H f It represents the height of wildfire; γ represents the regional correction coefficient, usually γ=1; I represents the intensity of fire line.

[0106] In this way, by considering the geographical feature information of the target area and the speed of volcanic spread, the real-time meteorological conditions and the vegetation conditions of the line underlying the surface can be considered, which will facilitate the subsequent accurate alarm of transmission line wildfire tripping through the Bayesian network.

[0107] Step S3: input the wildfire spread speed and the wildfire height into a preset Bayesian network to obtain the wildfire tripping risk of the transmission line in the target area, and issue corresponding risk warning information based on the wildfire tripping risk.

[0108] It can be understood that after the wildfire spread speed and the wildfire height are calculated, they can be discretized to facilitate input into the Bayesian network model to predict the wildfire tripping risk of the transmission line in the target area. In addition, the altitude and the distance between the fire point and the line need to be discretized to facilitate input into the Bayesian network model. The discretization standard of the influencing factor is shown in the following table:

[0109] Table 3 Impact factor dispersion standard

[0110]

[0111] It is understandable that it is necessary to collect the fire point information of the wildfire alarm in the transmission corridor in the past five years in advance to train the Bayesian network, wherein the training process is specifically as follows: first, obtain the influencing factors of the wildfire alarm point data and the non-wildfire alarm point data respectively, discretize each of the influencing factors, and randomly extract wildfire alarm point samples and non-wildfire alarm point samples from the processing results to form a training set. Secondly, based on the sample distribution of the training set, estimate the conditional probability of each influencing factor under the condition of fire and the condition of no fire, wherein the calculation formula of the conditional probability is In the formula, sample x ij is the factor x that falls into the jth class i , n ij For sample x ij The number of ij |Y=0) and P(x ij |Y=1) are samples x under no fire and fire conditions respectively ij Again, based on the conditional probability of each influencing factor under the condition of fire and no fire, the alarm probability and non-alarm probability of each influencing factor are obtained respectively, where the calculation formula of the non-alarm probability and alarm probability of each influencing factor is: P(Y=0|x1,x2,…,x n )=∏P(x i |Y=0), P(Y=1|x1,x2,…,x n )=∏P(x i |Y=1), where P(Y=0|x1,x2,…,x n ) and P(Y=1|x1,x2,…,x n ) are the probability of no alarm and the probability of alarm respectively, x1, x2,…, x nis the influencing factor of each wildfire. Finally, the alarm probability and non-alarm probability of each influencing factor are normalized to determine the wildfire tripping risk of the transmission line. The calculation formula for the normalization is:

[0112] It should be noted that when training the Bayesian network, it is also necessary to set the wildfire occurrence time risk level and wind direction classification standard in advance, so as to train the Bayesian network model together with the influencing factors, wildfire occurrence time risk level and wind direction classification standard, so as to accurately obtain the wildfire tripping Bayesian risk probability. The structural diagram of the Bayesian network is shown in the figure. Figure 2 Among them, the classification standards for the time of occurrence of wildfire tripping risks of transmission lines and the classification standards for wind direction are shown in the following table:

[0113] Table 4 Classification standard for the occurrence time of wildfire tripping risk of transmission lines

[0114] Wildfire time 0：00-6:59 7:00-12:59 19:00-23:59 13:00-18:59 Risk Level 1 2 3 4

[0115] Table 5 Wind direction classification standards

[0116]

[0117] In this way, by training the Bayesian network in advance, accurate warnings of power line tripping caused by wildfires can be achieved.

[0118] By calculating the wildfire tripping Bayesian risk probability for each transmission line grid in the target area one by one, the wildfire tripping risk of the transmission line in the target area can be obtained, and corresponding risk warning information can be issued based on the wildfire tripping risk. The classification standard of the risk warning information is shown in the following table:

[0119] Table 6 Classification standards for risk warning information

[0120] Bayesian Risk (0，0.25) (0.25，0.5] (0.5，0.75] (0.75，1] Risk Level Low risk Medium risk High risk Very high risk

[0121] The embodiment of the present application can directly obtain the information of the fire point and the vicinity of the fire point by respectively obtaining the coordinate information of the fire point and the geographical feature information, meteorological information and nearby line information near the fire point, so as to facilitate the subsequent accurate calculation of the wildfire spread speed; the wildfire spread speed of the target area is calculated based on the coordinate information of the fire point, the nearby geographical feature information, meteorological information and nearby line information, and the wildfire height of the target area is predicted based on the wildfire spread speed, so that the wildfire spread speed and the wildfire height of the target area under the conditions of wildfire occurrence can be accurately calculated, so as to facilitate the subsequent accurate prediction of the wildfire alarm risk by the Bayesian network; finally, the wildfire spread speed and the wildfire height are input into the Bayesian network, so as to accurately predict the wildfire tripping risk of the transmission line in the target area, and the real-time meteorological conditions and the vegetation conditions of the line underlying the line can be considered to realize the accurate alarm of the wildfire tripping of the transmission line.

[0122] Embodiment 2

[0123] Please refer to Figure 3 , Figure 3 It is a structural diagram of an embodiment of a mountain fire trip alarm system based on a Bayesian network provided by the present application, including an acquisition module 01, a calculation module 02 and an alarm module 03;

[0124] The acquisition module 01 is used to respectively acquire the coordinate information of the fire point and the geographical feature information, meteorological information and nearby line information near the fire point;

[0125] Specifically, the acquisition module 01 includes: a first acquisition unit and a second acquisition unit; the first acquisition unit is used to use a meteorological satellite to preliminarily locate the initial coordinate information of the fire point, and send the initial coordinate information of the fire point to the wildfire monitoring center; the second acquisition unit is used to call the monitoring image near the initial coordinate information through the wildfire monitoring center to determine the coordinate information of the fire point, wherein the monitoring image is obtained from the monitoring equipment installed on the transmission line or transmission tower.

[0126] It is understandable that the initial coordinate information of the fire point can be preliminarily located by using meteorological satellites, and the initial coordinate information of the fire point can be sent to the mountain fire monitoring center. If monitoring equipment is installed on the transmission line or transmission tower near the fire point, the monitoring image near the initial coordinate information can be called by the mountain fire monitoring center to accurately determine the coordinate information of the fire point. Of course, the coordinate position of the fire point can also be directly and accurately determined by meteorological satellites through hot spot identification, or the position of the fire point can be determined directly by monitoring equipment installed on the transmission line or transmission tower from the position of the fire point and the distance between the position of the fire point and the transmission line or transmission tower without using meteorological satellites.

[0127] In this way, the initial coordinates of the fire point are preliminarily located by means of a meteorological satellite, and monitoring images near the initial coordinate information are called to accurately determine the coordinate information of the fire point.

[0128] After obtaining the coordinate information of the fire point, the basic database in the target area can be read to obtain the geographical feature information near the fire point, including geographic elevation data (DEM), vegetation type, land use type, NDVI vegetation index and combustible load, etc.; in addition, it is also necessary to read the meteorological station in the target area to obtain the meteorological information near the fire point, including wind speed, wind direction, temperature, relative humidity, etc.; at the same time, it is also necessary to read the transmission line ledger data to obtain the line information near the fire point, including voltage level, installation height, line ground clearance distance and line foundation tower coordinates, etc. After obtaining the geographical feature information, meteorological information and nearby line information near the fire point, it is possible to import the human factor data layer such as the historical fire point density, distance from the road, distance from the settlement, GDP and population density data in the study area, and grid all the data.

[0129] It should be noted that if the source data resolution is lower than the grid size resolution, an interpolation algorithm is used to divide it into corresponding grids; if the source data resolution is lower than the grid size resolution, an average aggregation method is used to merge them. The grid resolution can usually be a grid accuracy of 1000m×1000m, or 250m×250m or 500m×500m can be selected according to actual conditions. How to grid the grid is not the focus of this application and will not be expanded here.

[0130] Specifically, the wildfire tripping alarm system based on the Bayesian network also includes: a judgment module; the judgment module includes: a judgment unit and a first calculation unit; the judgment unit is used to determine whether there is a transmission line or a transmission tower within a preset range based on the coordinate information of the fire point; the first calculation unit is used to calculate the wildfire spread speed in the area with the transmission line or the transmission tower if there is a transmission line or a transmission tower.

[0131] After obtaining the geographical feature information, meteorological information and nearby line information near the fire point, it is necessary to determine whether there are transmission lines or transmission towers within the preset range based on the coordinate information of the fire point; if there are no transmission lines or transmission towers, it is considered that the fire point is temporarily not a threat to the power grid system, and the alarm system is exited to maintain wildfire monitoring. If there are transmission lines or transmission towers, the wildfire spread speed in the area with transmission lines or transmission towers is calculated. At this time, if the coordinate information of the fire point is (x, y), if the coordinate information of the transmission tower in the system is (x1, y1), by calculating the distance d between the fire point and the transmission tower, the relevant formula for calculating the distance is:

[0132]

[0133] in, dx=|x2-x1|,dy=|y2-y1|, when the calculated distance d between the fire point and the transmission tower is less than the preset range, it is considered that there is a transmission tower near the fire point, indicating that the fire point poses a threat to the power grid system, and it is necessary to further calculate the wildfire spread speed in the area with transmission lines or transmission towers.

[0134] It should be noted that the preset range can be freely set according to flame intensity, geographical feature information, meteorological information, etc., and this application does not impose any restrictions.

[0135] In this way, by judging whether there are transmission lines or transmission towers near the fire point, if there are transmission lines or transmission towers near the fire point, there is a risk of wildfire tripping. It is necessary to further calculate the wildfire spread speed in the area with transmission lines or transmission towers to help accurately predict the wildfire tripping risk of transmission lines in the future.

[0136] The calculation module 02 is used to calculate the wildfire spread speed in the target area based on the coordinate information of the fire point, the nearby geographical feature information, the meteorological information and the nearby line information, and predict the wildfire height in the target area based on the wildfire spread speed;

[0137] Specifically, the calculation module 02 includes: a second calculation unit, a third acquisition unit and a third calculation unit; the second calculation unit is used to determine the initial spread speed of the wildfire based on the meteorological information; the third acquisition unit is used to obtain the combustible material configuration pattern coefficient, the slope correction coefficient and the wind speed correction coefficient, wherein the combustible material configuration pattern coefficient is determined by the nearby geographical feature information, the slope correction coefficient is determined by the slope between the coordinate information of the fire point and the nearby line information, and the wind speed correction coefficient is determined by the meteorological information; the third calculation unit is used to obtain the wildfire spread speed in the target area based on the initial spread speed, the combustible material configuration pattern coefficient, the slope correction coefficient and the wind speed correction coefficient.

[0138] It is understandable that, first, when the fire point is located near a flat land with uniformly distributed vegetation, the fire point will only be affected by meteorological information. Therefore, it is necessary to first determine the initial spread speed of the wildfire based on the acquired meteorological information such as temperature and wind speed, wherein the calculation formula for the initial spread speed R0 of the wildfire is:

[0139] R0=aT+bV+c(100-H)-d;

[0140] Where R0 is the initial spread speed of wildfire; T is the ambient temperature; V is the average wind level, which is 0 when there is no wind; H is the location elevation, and a, b, c and d are all constant coefficients (a, b, c and d can be taken as a=0.03, b=0.03, c=0.03, d=0.3).

[0141] Secondly, when there is vegetation and slope near the fire point, it is necessary to obtain the combustible configuration pattern coefficient, slope correction coefficient and wind speed correction coefficient, wherein the combustible configuration pattern coefficient is determined by the nearby geographical feature information, and the combustible configuration pattern coefficient K s The specific values ​​are shown in the following table:

[0142] Table 1 Combustible material configuration pattern coefficient table K s

[0143]

[0144]

[0145] The slope correction coefficient is determined by the slope between the coordinate information of the ignition point and the nearby line information, wherein the calculation formula of the slope correction coefficient is K θ for:

[0146]

[0147] The wind speed correction coefficient is determined by the meteorological information of wind force level V; wherein the wind speed correction coefficient K v The calculation formula is:

[0148] K v =e 0.1783V ;

[0149] Finally, based on the initial spread speed, the combustible configuration pattern coefficient, the slope correction coefficient and the wind speed correction coefficient, the wildfire spread speed in the target area is obtained, wherein the wildfire spread speed is calculated as follows:

[0150] R=R0K s K θ K v ;

[0151] Where R is the wildfire spread speed; R0 is the initial wildfire spread speed; K s K is the combustible configuration pattern coefficient; θ is the slope correction coefficient; K v is the wind speed correction factor.

[0152] In this way, by comprehensively considering the coordinate information of the fire point, the nearby geographical feature information, the meteorological information and the nearby line information, the wildfire spread speed in the target area under the conditions of wildfire occurrence can be accurately calculated, which facilitates the subsequent Bayesian network to accurately predict the wildfire warning risk.

[0153] Specifically, the calculation module 02 also includes: a fourth acquisition unit, a fourth calculation unit and a fifth calculation unit: the fourth acquisition unit is used to determine the vegetation type in the target area based on the geographic feature information of the target area, and obtain the unit mass of combustible calorific value and combustible load corresponding to the vegetation type; the fourth calculation unit is used to determine the fire line intensity of the target area based on the unit mass of combustible calorific value, the combustible load and the wildfire spread speed; the fifth calculation unit is used to predict the wildfire height of the target area based on the fire line intensity and the regional correction coefficient corresponding to the target area.

[0154] After determining the wildfire spread speed, the wildfire height in the target area can be predicted. Specifically, first, based on the geographic feature information of the target area, determine the vegetation type in the target area, and obtain the unit mass combustible calorific value and combustible load corresponding to the vegetation type; for example, based on the geographic feature information of the target area, it can be known that there are only coniferous forests in the target area, then according to the following table, it can be known that the corresponding unit mass combustible calorific value is 19620 kJ / kg. The value range of the unit mass combustible calorific value q is shown in the following table:

[0155] Table 2 Reference range of calorific value q per unit mass of combustibles

[0156]

[0157] Secondly, based on the calorific value per unit mass of the combustibles, the combustible load and the wildfire spread speed, the fire line intensity of the target area is determined, wherein the calculation formula for the fire line intensity is:

[0158] I = qWR / 600;

[0159] In the formula, I represents the intensity of the fire line, kW / m; q represents the calorific value of the combustible material per unit mass, kJ / kg; W represents the combustible material load, t / hm2; R represents the flame spread speed, m / min.

[0160] Finally, based on the fire line intensity and the regional correction coefficient corresponding to the target area, the wildfire height of the target area is predicted, where the wildfire height H f The calculation formula is:

[0161] H f =γ(I / 250) 1 / 2 ;

[0162] In the formula, H f It represents the height of wildfire; γ represents the regional correction coefficient, usually γ=1; I represents the intensity of fire line.

[0163] In this way, by considering the geographical feature information of the target area and the speed of volcanic spread, the real-time meteorological conditions and the vegetation conditions of the line underlying the surface can be considered, which will facilitate the subsequent accurate alarm of transmission line wildfire tripping through the Bayesian network.

[0164] The alarm module 03 is used to input the wildfire spreading speed and the wildfire height into a preset Bayesian network to obtain the wildfire tripping risk of the power transmission line in the target area, and issue corresponding risk alarm information based on the wildfire tripping risk.

[0165] It can be understood that after the wildfire spread speed and the wildfire height are calculated, they can be discretized to facilitate input into the Bayesian network model to predict the wildfire tripping risk of the transmission line in the target area. In addition, the altitude and the distance between the fire point and the line need to be discretized to facilitate input into the Bayesian network model. The discretization standard of the influencing factor is shown in the following table:

[0166] Table 3 Impact factor dispersion standard

[0167]

[0168] It is understandable that it is necessary to collect the fire point information of the wildfire alarm in the transmission corridor in the past five years in advance to train the Bayesian network, wherein the training process is specifically as follows: first, obtain the influencing factors of the wildfire alarm point data and the non-wildfire alarm point data respectively, discretize each of the influencing factors, and randomly extract wildfire alarm point samples and non-wildfire alarm point samples from the processing results to form a training set. Secondly, based on the sample distribution of the training set, estimate the conditional probability of each influencing factor under the condition of fire and the condition of no fire, wherein the calculation formula of the conditional probability is In the formula, sample x ij is the factor x that falls into the jth class i , n ij For sample x ij The number of ij |Y=0) and P(x ij |Y=1) are samples x under no fire and fire conditions respectively ij Again, based on the conditional probability of each influencing factor under the condition of fire and no fire, the alarm probability and non-alarm probability of each influencing factor are obtained respectively, where the calculation formula of the non-alarm probability and alarm probability of each influencing factor is: P(Y=0|x1,x2,…,x n )=∏P(x i |Y=0), P(Y=1|x1,x2,…,x n )=∏P(x i |Y=1), where P(Y=0|x1,x2,…,x n ) and P(Y=1|x1,x2,…,x n ) are the probability of no alarm and the probability of alarm respectively, x1, x2,…, x n is the influencing factor of each wildfire. Finally, the alarm probability and non-alarm probability of each influencing factor are normalized to determine the wildfire tripping risk of the transmission line. The calculation formula for the normalization is:

[0169] It should be noted that when training the Bayesian network, it is also necessary to set the wildfire occurrence time risk level and wind direction classification standard in advance, so as to train the Bayesian network model together with the influencing factors, wildfire occurrence time risk level and wind direction classification standard, so as to accurately obtain the wildfire tripping Bayesian risk probability. The structural diagram of the Bayesian network is shown in the figure. Figure 2 Among them, the classification standards for the time of occurrence of wildfire tripping risks of transmission lines and the classification standards for wind direction are shown in the following table:

[0170] Table 4 Classification standard for the occurrence time of wildfire tripping risk of transmission lines

[0171] Wildfire time 0：00-6:59 7:00-12:59 19:00-23:59 13:00-18:59 Risk Level 1 2 3 4

[0172] Table 5 Wind direction classification standards

[0173]

[0174] In this way, by training the Bayesian network in advance, accurate warnings of power line tripping caused by wildfires can be achieved.

[0175] By calculating the wildfire tripping Bayesian risk probability for each transmission line grid in the target area one by one, the wildfire tripping risk of the transmission line in the target area can be obtained, and corresponding risk warning information can be issued based on the wildfire tripping risk. The classification standard of the risk warning information is shown in the following table:

[0176] Table 6 Classification standards for risk warning information

[0177] Bayesian Risk (0，0.25) (0.25，0.5] (0.5，0.75] (0.75，1] Risk Level Low risk Medium risk High risk Very high risk

[0178] The embodiment of the present application can directly obtain the information of the fire point and the vicinity of the fire point by respectively obtaining the coordinate information of the fire point and the geographical feature information, meteorological information and nearby line information near the fire point, so as to facilitate the subsequent accurate calculation of the wildfire spread speed; the wildfire spread speed of the target area is calculated based on the coordinate information of the fire point, the nearby geographical feature information, meteorological information and nearby line information, and the wildfire height of the target area is predicted based on the wildfire spread speed, so that the wildfire spread speed and the wildfire height of the target area under the conditions of wildfire occurrence can be accurately calculated, so as to facilitate the subsequent accurate prediction of the wildfire alarm risk by the Bayesian network; finally, the wildfire spread speed and the wildfire height are input into the Bayesian network, so as to accurately predict the wildfire tripping risk of the transmission line in the target area, and the real-time meteorological conditions and the vegetation conditions of the line underlying the line can be considered to realize the accurate alarm of the wildfire tripping of the transmission line.

[0179] The specific embodiments described above further describe the purpose, technical solutions and beneficial effects of the present application in detail. It should be understood that the above description is only a specific embodiment of the present application and is not intended to limit the scope of protection of the present application. It is particularly pointed out that for those skilled in the art, any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the scope of protection of the present application.

Claims

1. A wildfire tripping alarm method based on Bayesian network, It is characterized in that include: Respectively obtain the coordinate information of the fire point and the geographical feature information, meteorological information and nearby line information near the fire point; Calculate the wildfire spread speed in the target area based on the coordinate information of the fire point, the nearby geographical feature information, the meteorological information and the nearby line information, and predict the wildfire height in the target area based on the wildfire spread speed; The wildfire spreading speed and the wildfire height are input into a preset Bayesian network to obtain the wildfire tripping risk of the power transmission lines in the target area, and corresponding risk warning information is issued based on the wildfire tripping risk.

2. A mountain fire trip alarm method based on Bayesian network according to claim 1, It is characterized in that The coordinate information of the ignition points is obtained respectively, specifically: Use meteorological satellites to preliminarily locate the initial coordinate information of the fire point, and send the initial coordinate information of the fire point to the wildfire monitoring center; The wildfire monitoring center calls the monitoring image near the initial coordinate information to determine the coordinate information of the fire point, wherein the monitoring image is obtained from the monitoring equipment installed on the transmission line or transmission tower.

3. A mountain fire trip alarm method based on Bayesian network according to claim 1, It is characterized in that Before calculating the wildfire spreading speed in the target area based on the coordinate information of the fire point, the nearby geographical feature information, the meteorological information and the nearby line information, the method further includes: Based on the coordinate information of the fire point, determining whether there is a transmission line or a transmission tower within a preset range; If there are transmission lines or transmission towers, calculate the wildfire spread rate in the area with transmission lines or transmission towers.

4. A mountain fire trip alarm method based on Bayesian network according to claim 1, It is characterized in that The calculation of the wildfire spread speed in the target area based on the coordinate information of the fire point, the nearby geographical feature information, the meteorological information and the nearby line information is specifically as follows: Determine the initial spread rate of the wildfire based on the meteorological information; Obtaining a combustible configuration pattern coefficient, a slope correction coefficient, and a wind speed correction coefficient, wherein the combustible configuration pattern coefficient is determined by the nearby geographical feature information, the slope correction coefficient is determined by the slope between the coordinate information of the ignition point and the nearby line information, and the wind speed correction coefficient is determined by the meteorological information; Based on the initial spread speed, the combustible material configuration pattern coefficient, the slope correction coefficient and the wind speed correction coefficient, the wildfire spread speed in the target area is obtained.

5. A mountain fire trip alarm method based on Bayesian network according to claim 1, It is characterized in that The prediction of the wildfire height in the target area based on the wildfire spread speed is specifically: Based on the geographic feature information of the target area, determine the vegetation type in the target area, and obtain the unit mass calorific value and the combustible load corresponding to the vegetation type; Determine the intensity of the fire line in the target area based on the calorific value per unit mass of the combustibles, the combustible load and the wildfire spread speed; Based on the intensity of the fire line and the regional correction coefficient corresponding to the target area, the height of the wildfire in the target area is predicted.

6. A mountain fire trip alarm method based on Bayesian network according to claim 1, It is characterized in that The training process of the Bayesian network is specifically as follows: Obtaining influencing factors of mountain fire warning point data and non-mountain fire warning point data respectively, discretizing each of the influencing factors, and randomly extracting mountain fire warning point samples and non-mountain fire warning point samples from the processing results to form a training set; Based on the sample distribution of the training set, the conditional probability of each influencing factor under the fire condition and the no-fire condition is estimated; Based on the conditional probability of each influencing factor under fire conditions and no fire conditions, the alarm probability and non-alarm probability of each influencing factor are obtained respectively; The alarm probability and non-alarm probability of each influencing factor are normalized to determine the wildfire tripping risk of the transmission line.

7. A mountain fire trip alarm system based on Bayesian network, It is characterized in that include: Acquisition module, calculation module and alarm module; The acquisition module is used to respectively acquire the coordinate information of the fire point and the geographical feature information, meteorological information and nearby line information near the fire point; The calculation module is used to calculate the wildfire spread speed in the target area based on the coordinate information of the fire point, the nearby geographical feature information, the meteorological information and the nearby line information, and predict the wildfire height in the target area based on the wildfire spread speed; The alarm module is used to input the wildfire spread speed and the wildfire height into a preset Bayesian network to obtain the wildfire tripping risk of the transmission line in the target area, and issue corresponding risk alarm information based on the wildfire tripping risk.

8. The mountain fire trip alarm system based on Bayesian network according to claim 7, It is characterized in that The acquisition module includes: a first acquisition unit and a second acquisition unit; The first acquisition unit is used to preliminarily locate the initial coordinate information of the fire point using a meteorological satellite, and send the initial coordinate information of the fire point to a mountain fire monitoring center; The second acquisition unit is used to call the monitoring image near the initial coordinate information through the wildfire monitoring center to determine the coordinate information of the fire point, wherein the monitoring image is obtained from the monitoring equipment installed on the transmission line or transmission tower.

9. The mountain fire trip alarm system based on Bayesian network according to claim 7, It is characterized in that Also includes: Judgment module; The judgment module includes: a judgment unit and a first calculation unit; The judgment unit is used to determine whether there is a power transmission line or a power transmission tower within a preset range based on the coordinate information of the fire point; The first calculation unit is used to calculate the wildfire spread speed in the area with the transmission line or the transmission tower if there is a transmission line or a transmission tower.

10. A mountain fire trip alarm system based on Bayesian network according to claim 7, It is characterized in that The calculation module includes: a second calculation unit, a third acquisition unit and a third calculation unit; The second calculation unit is used to determine the initial spread speed of the wildfire based on the meteorological information; The third acquisition unit is used to acquire a combustible configuration pattern coefficient, a slope correction coefficient and a wind speed correction coefficient, wherein the combustible configuration pattern coefficient is determined by the nearby geographical feature information, the slope correction coefficient is determined by the slope between the coordinate information of the ignition point and the nearby line information, and the wind speed correction coefficient is determined by the meteorological information; The third calculation unit is used to obtain the wildfire spread speed in the target area based on the initial spread speed, the combustible material configuration pattern coefficient, the slope correction coefficient and the wind speed correction coefficient.

11. The mountain fire trip alarm system based on Bayesian network according to claim 7, It is characterized in that The calculation module further includes: a fourth acquisition unit, a fourth calculation unit and a fifth calculation unit: The fourth acquisition unit is used to determine the vegetation type in the target area based on the geographical feature information of the target area, and obtain the unit mass calorific value and the combustible load corresponding to the vegetation type; The fourth calculation unit is used to determine the fire line intensity of the target area based on the calorific value per unit mass of the combustible material, the combustible material load and the wildfire spread speed; The fifth calculation unit is used to predict the height of the wildfire in the target area based on the intensity of the fire line and the regional correction coefficient corresponding to the target area.

12. The mountain fire trip alarm system based on Bayesian network according to claim 7, It is characterized in that The training process of the Bayesian network is specifically as follows: Obtaining influencing factors of mountain fire warning point data and non-mountain fire warning point data respectively, discretizing each of the influencing factors, and randomly extracting mountain fire warning point samples and non-mountain fire warning point samples from the processing results to form a training set; Based on the sample distribution of the training set, the conditional probability of each influencing factor under the fire condition and the no-fire condition is estimated; Based on the conditional probability of each influencing factor under fire conditions and no fire conditions, the alarm probability and non-alarm probability of each influencing factor are obtained respectively; The alarm probability and non-alarm probability of each influencing factor are normalized to determine the wildfire tripping risk of the transmission line.

13. A computer-readable storage medium, It is characterized in that The computer-readable storage medium includes a stored computer program, wherein when the computer program is executed, the device where the computer-readable storage medium is located is controlled to execute a mountain fire trip alarm method based on a Bayesian network as described in any one of claims 1 to 6.

Images