Power distribution overhead conductor tree touch fault fire moment determination method and system

Abstract

The invention discloses a power distribution overhead conductor tree touch fault fire moment determination method and system, and the method comprises the following steps: building an experiment environment, and collecting whole-process data according to a preset sampling frequency in the experiment environment; obtaining the initial moisture content of the experimental tree, calculating the moisture evaporation heat absorption amount and the moisture evaporation mass of the tree model, updating the moisture content of the tree, and performing aligned storage according to time; fitting the relation of the conductivity, the moisture content and the temperature of the tree in the whole process to obtain a conductivity function; constructing a resistance simulation curve by using a conductivity function according to the moisture content and the temperature of the tree, and constructing a resistance actual measurement curve by using voltage and current data collected in an experimental environment; and comparing the resistance simulation curve with the resistance actual measurement curve, and judging the moment when the resistance simulation curve and the resistance actual measurement curve are obviously bifurcated as the moment when the open fire appears. The method is wide in application range and low in cost and maintenance difficulty, and high accuracy is guaranteed.

Description

Technical Field

[0001] The present invention belongs to the technical field of distribution network fault analysis, and in particular relates to a method and system for determining the fire ignition moment of a distribution overhead conductor tree-touching fault. Background Art

[0002] When medium-voltage overhead distribution lines pass through forested areas, there's a high risk of conductor contact failures caused by trees. These faults, characterized by high transition resistance and a high probability of fire, can easily trigger forest fires, resulting in serious safety hazards and property damage. For example, the 2009 Black Saturday fires in Victoria, Australia, and the 2018 wildfires in California, USA, were both confirmed to be linked to conductor contact failures. According to data from SCE's 2019 WMP program annual report, of the 302 electrical fire reports between 2015 and 2017, 92% involved medium-voltage distribution systems, with fires caused by conductor contact with trees accounting for 17% of these fires.

[0003] Current detection methods for TSF (Tree-contact Single-phase-to-ground Faults) typically classify them as common HIF (High Impedance Faults) and treat them accordingly. Existing literature analyzes the frequency domain characteristics and composite features of zero-sequence electrical quantities to achieve fault detection. For example, some studies propose using the slope of the interval curve of the zero-sequence current waveform to determine faults, while others use differential methods to establish waveform slope variation indicators and utilize the zero-sequence current fundamental frequency and harmonics to determine the occurrence of HIFs. Although these methods can identify certain types of faults, their sensitivity still falls short of the TSF detection requirements. Another approach employs artificial intelligence technology for fault identification, analyzing the distortion characteristics of HIF zero-sequence current, using ensemble learning algorithms to identify faults, or combining convolutional neural networks to analyze traveling wave signals.

[0004] The Chinese invention patent application with publication number CN118797504A discloses a staged identification method based on the evolution process of a single-phase conductor touching a tree and grounding fault. The method includes the following steps: collecting initial distribution network line data and phenomenon video of the entire process of the single-phase conductor touching a tree and grounding fault; dividing the initial distribution network line data into four fault stages and one normal stage data sets based on the time point of the phenomenon video; preprocessing the data set to obtain intermediate distribution network line data; establishing an initial fault development and evolution identification model based on a DICNN network model, and training it to obtain a final fault development and evolution identification model; setting the final fault development and evolution identification model to the distribution network line terminal, and identifying whether a tree-line contact fault occurs and the stage to which the fault develops by detecting the distribution network line data of the distribution network line.

[0005] While these methods offer certain advantages, their universal applicability in practical applications cannot be guaranteed due to the limited number of fault samples. Current research on TSF primarily focuses on fault detection and electrical signature analysis, but precise estimation of the moment of fault occurrence has received insufficient attention. Furthermore, traditional fault detection methods fail to fully consider key physical processes such as tree moisture content, water evaporation during temperature rise, and carbonization, making it difficult to accurately identify key changes before a fire breaks out. Summary of the Invention

[0006] The present invention provides a method and system for determining the fire ignition moment of a distribution overhead conductor tree-touching fault, aiming to solve the problems of the prior art such as complex system, low versatility, and inaccurate fire ignition moment prediction.

[0007] By combining the physical key parameter processes of trees, this invention establishes a dynamic fault model based on real-time monitoring. This model can accurately estimate the ignition moment of a single-phase grounding fault caused by a tree, reducing the errors of traditional temperature or current threshold methods. Compared with artificial intelligence methods, it can effectively reduce system costs, computing costs, and training costs. The fault model takes into account physical processes such as water evaporation and tree carbonization during temperature rise. By analyzing the heat transfer process of trees in stages, it can accurately identify key nodes before fire, such as the rise and turning of fault current, provide accurate fire ignition moment estimation, and provide technical support for timely fire prevention response.

[0008] To solve the above technical problems, the present invention proposes a determination method comprising the following steps:

[0009] Build an experimental environment to simulate the entire process of a distribution overhead conductor tree fault, and collect data from the entire process at a preset sampling frequency in the experimental environment;

[0010] Obtain the initial moisture content of the experimental trees, continuously collect and align the complete process data by time during the experiment, and generate real-time data;

[0011] Calculate the water evaporation heat absorption of the tree model based on real-time data, then calculate the water evaporation mass based on the water evaporation heat absorption, use the water evaporation mass to update the tree moisture content and save it in time alignment;

[0012] Based on real-time data, aligned tree moisture content, and voltage and current data collected in the experimental environment, the relationship between tree conductivity, moisture content, and temperature throughout the entire process is fitted to obtain the conductivity function.

[0013] Using the conductivity function to construct a resistance simulation curve according to the water content and temperature of the tree, and using the voltage and current data collected in the experimental environment to construct a resistance measurement curve;

[0014] The resistance simulation curve and the resistance measurement curve are compared, and the moment when there is an obvious bifurcation between the resistance simulation curve and the resistance measurement curve is determined to be the moment when the open flame appears.

[0015] Preferably, the experimental environment includes a neutral point ungrounded simulation part, a conductor touching a tree simulation part and a data acquisition part; the neutral point ungrounded simulation part is an energized three-phase overhead conductor; the conductor touching a tree part includes trees transplanted to the test site, and one phase is drawn out from the three-phase overhead conductor and connected to the trees; the data acquisition part continuously records the phase voltage and grounding current of the entire process, records the test phenomena of the entire ignition process, and records the temperature changes of the trees until the discharge channel passes through the vegetation and a short circuit occurs.

[0016] Preferably, the full-process data includes conductor phase voltage, current passing through the experimental tree, and temperature of the experimental tree.

[0017] Preferably, the method for obtaining the initial moisture content includes experimental weighing and preset values;

[0018] The calculation formula for experimental weighing is:

[0019]

[0020] Where m s is the fresh weight of the wood block obtained from the experimental sampling, and m0 is the absolute dry mass of the wood block;

[0021] The preset values ​​are derived from the existing data of the corresponding tree species in the experimental site during the experimental season.

[0022] Preferably, the experimental process includes the tree temperature rise stage, the water evaporation stage, the carbonization development stage and the open flame flashover stage in sequence.

[0023] Preferably, the method for determining obvious bifurcation includes a difference index determination method and a visual analysis method;

[0024] The difference index is a relative index, and the calculation method is as follows:

[0025]

[0026] Where Δy t is the difference index at time t, y1(t) is the resistance value corresponding to the resistance simulation curve at time t, y2(t) is the resistance value corresponding to the resistance measured curve at time t, and ∈ is a preset parameter used to avoid the denominator being zero;

[0027] When Δy t When the preset threshold is exceeded, it is determined that the resistance simulation curve and the resistance measurement curve have obvious bifurcation.

[0028] Preferably, the relationship between the electrical conductivity, moisture content, and temperature of the trees in the entire process is fitted using the least squares method. The fitted relationship between the electrical conductivity, temperature, and moisture content is expressed as follows:

[0029] lnσ t (W,T)=7.458W+0.057T-0.050W·T-10.754

[0030] Where σ t (W,T) is the conductivity calculated relative to the temperature and moisture content at time t, where W is the tree moisture content and T is the tree temperature.

[0031] Preferably, the calculation formula of the resistance in the resistance simulation curve is:

[0032]

[0033] Where n is the level of the touch line in the tree model, l0 is the basic branch length in the tree model, that is, the trunk height, r0 is the basic branch radius, that is, the trunk radius;

[0034] The tree model is a fractal network model simplified from the experimental tree morphology. In the tree model, α and β are defined as the length ratio and radius ratio of the branches of the k+1th layer network and the kth layer network, respectively, and are expressed as follows:

[0035]

[0036] Where, l k 、l k+1 are the length of branches in the k-th layer network and the length of branches in the k+1-th layer network, r k 、r l+1 are the branch radius of the k-th layer network and the branch radius of the k+1-th layer network respectively.

[0037] Preferably, the calculation method of the water evaporation mass is:

[0038] Calculate the total heat Q generated by the current passing through the tree at time t s :

[0039]

[0040] Where U is the phase voltage of the conductor connected to the tree;

[0041] The heat dissipation and absorption of trees include convection heat dissipation, radiation heat dissipation, temperature rise heat absorption and water evaporation heat absorption. After deducting convection heat dissipation and radiation heat dissipation from the total heat output, the remainder is the total heat absorption. The product of the total heat absorption and the evaporation absorption ratio coefficient is the water evaporation heat absorption, which is expressed as follows:

[0042] Q w =(Q s -Q c -Q r )·k f

[0043] k f =(1+e 98-T ) -1

[0044] Where Q w The heat absorbed by water evaporation, Q s is the total heat production, Q c is the convective heat dissipation, Q r is the radiation heat dissipation, k f is the evaporation absorption proportional coefficient, T is the tree temperature;

[0045] Calculate the water evaporation mass based on the heat absorbed by water evaporation:

[0046]

[0047] Where Δm is the mass of water evaporated, T w is the water temperature.

[0048] Another aspect of the present invention provides a system for determining the fire ignition moment of a distribution overhead conductor tree-touching fault. The system is used to implement the above-mentioned determination method, comprising:

[0049] The experimental environment module is used to simulate the entire process of a distribution overhead conductor contacting a tree fault, including the neutral point ungrounded simulation part and the conductor contacting a tree simulation part;

[0050] The parameter acquisition module further includes: a voltage and current acquisition unit for collecting phase voltage and ground current data in the experimental environment in a time-aligned manner; a temperature acquisition unit for collecting the temperature of the trees during the experiment and saving them in a time-aligned manner;

[0051] The data processing module further includes: a water evaporation calculation unit for calculating the water evaporation heat absorption of the tree model based on the real-time collected data, and thereby calculating the water evaporation mass, which is used to update the real-time water content of the tree; a conductivity fitting unit for fitting the relationship between the tree conductivity, water content and temperature based on the experimental data to generate a conductivity function;

[0052] The resistance curve generation module further includes: a resistance simulation curve construction unit for constructing a resistance simulation curve based on the conductivity function and the moisture content and temperature of the tree; a resistance measurement curve construction unit for constructing a resistance measurement curve based on the voltage and current data collected in the experimental environment;

[0053] The determination module is used to compare the resistance simulation curve with the resistance measurement curve, and determine the moment when the resistance simulation curve and the resistance measurement curve have obvious bifurcation as the moment when the open flame appears.

[0054] Compared with the prior art, the present invention has the following technical effects:

[0055] 1. In the judgment method proposed in this invention, the fault model takes into account physical processes such as water evaporation and tree carbonization during temperature rise. By analyzing the heat transfer process of trees in stages, it can accurately identify key nodes before fire, such as the rise and turning points of fault current, and provide accurate fire ignition time estimation, providing technical support for timely fire prevention response, and can better meet the needs of fire process inversion and responsibility determination.

[0056] 2. The judgment method proposed in the present invention incorporates external environmental parameters such as temperature and humidity into the model in real time and dynamically adjusts the calculation conditions, so that the method can adapt to the prediction of fault fire time in different seasons and different humidity conditions, thereby improving the applicability and prediction accuracy of the system.

[0057] 3. The proposed determination method iteratively calculates variables such as conductivity, resistance, and temperature, reflecting the changes in actual electrical characteristics during a fault. This closed-loop model allows for real-time tracking of the dynamic changes in various parameters over time, ensuring a more realistic estimate of the fire onset moment.

[0058] 4. The proposed determination method is based on physical models and experimental data, utilizing clear relationships between conductivity, moisture content, temperature, and resistance change characteristics for analysis. This method has a clear theoretical basis and is easy to verify and interpret. Compared to the black-box nature of artificial intelligence models, this method is more readily applicable to engineering practice. Furthermore, it does not rely on large-scale training data, thus providing reliable results even with limited data.

[0059] 5. The determination method proposed in the present invention can flexibly adjust parameters according to experimental conditions and tree characteristics. Compared with artificial intelligence models that need to be trained and optimized for specific application scenarios and have poor portability, the present invention has a wider range of applications.

[0060] 6. The determination method proposed in the present invention mainly relies on experimental data and simple fitting operations, does not require complex computing resources and high-performance hardware support, and can effectively reduce costs and maintenance difficulties. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] Figure 1 It is a schematic flow chart of the determination method of the present invention;

[0062] Figure 2 Schematic diagram of a real-type test platform according to an embodiment of the present invention;

[0063] Figure 3 This is a simplified diagram of a fractal tree according to an embodiment of the present invention;

[0064] Figure 4 is an iterative calculation flow chart according to an embodiment of the present invention;

[0065] Figure 5 This is a preview of the effect described in the embodiment of the present invention;

[0066] Figure 6 This is another effect preview diagram described in an embodiment of the present invention. DETAILED DESCRIPTION

[0067] In order to make the objectives, technical solutions and advantages of the present invention more clear, the technical solutions of the present invention will be clearly and completely described below in combination with specific embodiments of the present application and with reference to the accompanying drawings.

[0068] Example 1

[0069] This embodiment is a method for determining the fire moment of a distribution overhead wire touching a tree fault. Figure 1 As shown, the following steps are included from step 1 to step 6:

[0070] Step 1: Set up an experimental environment to simulate the entire process of a distribution overhead conductor contacting a tree fault. Within the experimental environment, collect data from the entire process at a preset sampling frequency. The data includes conductor phase voltage, current through the test tree, and the test tree temperature.

[0071] In order to observe the experimental phenomena and summarize the laws, a TSF real-type test platform was built. Figure 2As shown. The experimental environment includes three main parts: neutral point ungrounded simulation part, conductor touching tree simulation part and data acquisition part; the neutral point ungrounded simulation part is a three-phase overhead conductor with power, and this embodiment uses a 10kV line to build it; the conductor touching tree part includes trees transplanted to the test site, and the conductors from any phase A, B, and C of the three-phase overhead conductor are connected to the trees transplanted to the test site in advance; the data acquisition part continuously records the phase voltage and grounding current of the whole process, and records the phase voltage and grounding current of the whole process of TSF from the start of power-on. The phase voltage is used Figure 2 The medium voltage sensor records the ground current using Figure 2 The current sensor in the ignition device is used to record the entire ignition process test phenomenon with a high-definition camera, and the temperature changes of the trees are recorded with a thermal imaging thermometer until the discharge channel passes through the vegetation and a short circuit occurs.

[0072] Step 2: Obtain the initial moisture content of the experimental trees. During the experiment, continuously collect and align the complete process data by time to generate real-time data.

[0073] The method for obtaining the initial moisture content includes experimental weighing and preset values;

[0074] The calculation formula for experimental weighing is:

[0075]

[0076] Where m s is the fresh weight of the wood block obtained from the experimental sampling, and m0 is the absolute dry mass of the wood block;

[0077] The preset values ​​are derived from existing data on the corresponding tree species in the experimental site during the experimental season. For example, this can be achieved by making advance measurements of local tree species in different seasons, or by directly using research data on local tree species from relevant departments or institutions.

[0078] The test process is roughly divided into four stages: tree temperature rise stage, water evaporation stage, carbonization development stage, and open flame flashover stage. The phenomena in each stage are described below.

[0079] During the tree's temperature rise phase, at the initial stage of the fault, a visible light spot appears at the contact point, while no light spots appear elsewhere in the tree. This is because the current passing through the contact point causes a localized temperature increase. As the current continues to flow, the tree's temperature gradually rises, and the heat from the contact point spreads to the rest of the tree. During this process, the tree's electrical conductivity gradually increases with temperature, but water evaporation or carbonization has not yet occurred. The tree's temperature rise is most significant during this phase, and the rate of temperature rise can vary significantly from tree to tree.

[0080] During the evaporation phase, as the tree's temperature continues to rise, water begins to evaporate. Water vapor appears to emanate from the tree's surface, indicating that water within the tree is evaporating. This evaporation process absorbs a significant amount of heat, a process that inhibits further temperature increases. As the tree's water gradually evaporates, its moisture content decreases, causing changes in the tree's electrical conductivity. During this phase, water evaporation not only helps slow the temperature rise but also creates conditions for subsequent carbonization.

[0081] During the carbonization development stage, as the tree's temperature continues to rise, the carbonization process begins to manifest. The carbonization path begins at the contact point and gradually extends toward the ground, typically extending vertically at a much higher rate than lateral expansion. The width of the carbonization path is typically greatest at the contact point and extends downward over time. Furthermore, the expansion of the carbonization path is not always continuous; it may temporarily stagnate before carbonization resumes in previously uncarbonized areas. During this stage, the tree's electrical resistance gradually increases, and the carbonized areas appear as bright spots. The locations of these spots vary randomly and over time.

[0082] During the flashover stage, when the carbonization path further expands and reaches a certain extent, the tree begins to ignite. These flames burn along the carbonization path and may spread outward. At this point, a faint flash often appears at the end of the path, then darkens and continues to lengthen. In the final stages of the fault, the carbonization path extends toward the ground, but no wire-to-ground flashover occurs. As the flames spread, the tree bends, eventually forming a ground fire.

[0083] As shown in the four stages above, when TSF reaches the carbonization development stage, carbonization has already begun, and the carbonization path of the tree gradually expands, at which point the fire risk begins to emerge. As carbonization progresses, the tree's structure is destroyed, and the high temperatures caused by the electric current make the carbonization path increasingly obvious, gradually extending to the trunk and the ground. As the carbonization path expands, the decomposition of organic matter within the tree generates a large amount of heat, increasing the potential risk of fire.

[0084] If the carbonized area continues to expand, especially in the lower parts of trees and in drier conditions, flames can easily spread to surrounding vegetation, creating the conditions for a forest fire. Especially after carbonization reaches a certain depth, the fire can spread rapidly, endangering the surrounding ecosystem and safety. Therefore, starting in the third stage, the risk of forest fires increases significantly. To estimate the time of fire outbreak, it is necessary to model the tree temperature rise and water evaporation stages.

[0085] Step three: Calculate the water evaporation heat absorption of the tree model based on the real-time data, and then calculate the water evaporation mass based on the water evaporation heat absorption. Use the water evaporation mass to update the tree moisture content and save it in time alignment.

[0086] This step involves modeling the tree model. When the wire touches the surface of the tree, the current flows from the wire through each branch to the trunk and into the ground. To simplify the analysis, the tree is approximated as a cylinder with uniform electrical and thermal properties, and the transition area between branches is ignored. Figure 3 As shown in the figure, trees, as typical fractal structures, exhibit similar geometric features from trunk to branches and twigs. Therefore, a fractal network model can be used to simplify tree morphology. By setting the ratio of interlayer length and radius and assuming that conductivity is related to temperature and moisture content, the conductance of each layer can be calculated based on the branching structure.

[0087] The electrical conductivity of trees is affected by temperature and moisture content, so the core parameters of the model include tree temperature, moisture content, and size. The total TSF resistance is equivalent to the combined resistance of all branches in the current path. When current flows through a tree, joules of power are generated, part of which increases the wood temperature, part of which is dissipated through convection and radiation, and the remainder of which is lost to evaporation.

[0088] When the fault voltage and environmental parameters are known, the initial tree resistance can be calculated to determine the initial current and generated joule power of the tree. Using this initial resistance, the convective heat dissipation power and radiative heat dissipation power at the initial moment can be further calculated. Figure 4 As shown in Figure 1, as iterations proceed, the current Joule heat and water evaporation power are used to update the changes in tree temperature and moisture content. These changes affect the conductivity of the tree material, thereby affecting the resistance value of each branch and the total fault resistance in the path.

[0089] Specifically, the calculation formula of the resistance in the resistance simulation curve is:

[0090]

[0091] Where n is the level of the touch line in the tree model, l0 is the basic branch length in the tree model, that is, the trunk height, r0 is the basic branch radius, that is, the trunk radius;

[0092] The tree model is a fractal network model simplified from the experimental tree morphology. In the tree model, α and β are defined as the length ratio and radius ratio of the branches of the k+1th layer network and the kth layer network, respectively, and are expressed as follows:

[0093]

[0094] Where, l k 、l k+1are the length of branches in the k-th layer network and the length of branches in the k+1-th layer network, r k 、r k+1 are the branch radius of the k-th layer network and the branch radius of the k+1-th layer network respectively.

[0095] The calculation method of the water evaporation mass is:

[0096] Calculate the total heat Q generated by the current passing through the tree at time t s :

[0097]

[0098] Where U is the phase voltage of the wire connected to the tree; it should be noted that the heat generated by the current Q calculated in the above formula is s = is the heat generated by a branch at one level in the tree model. When considering the current heat generation of the entire tree model, the heat generation of each branch at each level needs to be calculated separately and then summed to form the current heat generation of the entire tree model. Subsequent calculations of convective heat dissipation, radiative heat dissipation, water evaporation heat absorption, and tree temperature rise heat absorption all represent the heat of a single level within the tree model. The heat dissipation or absorption at each level needs to be summed to form the heat dissipation or absorption of the entire tree model.

[0099] When electric current passes through a tree, the Joule heat (i.e., the heat generated by the electric current) generated is consumed inside the tree in three ways: part of the heat is used to increase the temperature of the wet wood itself, part of it is transferred to the external environment through convection and radiation heat dissipation on the surface of the wood, and the remaining part is absorbed by the water and becomes the latent heat of evaporation of the water, which is used for the phase change of the water from liquid to gas.

[0100] The heat dissipation and absorption of trees include convection heat dissipation, radiation heat dissipation, temperature rise heat absorption and water evaporation heat absorption. After deducting convection heat dissipation and radiation heat dissipation from the total heat output, the remainder is the total heat absorption. The product of the total heat absorption and the evaporation absorption ratio coefficient is the water evaporation heat absorption, which is expressed as follows:

[0101] Q w =(Q s -Q c -Q r )·k f

[0102] Q c =∫k c S c (TT e )dt

[0103]

[0104] k f =(1+e 98-T )-1

[0105] Where Q w The heat absorbed by water evaporation, Q s is the total heat production, Q c is the convective heat dissipation, Q r is the radiation heat dissipation, k c is the branch convection heat dissipation coefficient, S c is the lateral area of ​​the branch, T is the tree temperature, T e is the ambient temperature, k r is the branch radiation heat transfer coefficient, δ0 is the Stefan-Boltzmann constant, k f is the evaporation absorption proportional coefficient.

[0106] Calculate the water evaporation mass based on the heat absorbed by water evaporation:

[0107]

[0108] Where Δm is the mass of water evaporated, T w is the water temperature.

[0109] The new tree moisture content becomes W-ΔW, where:

[0110]

[0111] That is, the ratio of the water evaporation mass obtained by the above formula to the absolute dry mass of the wood block obtained in the experiment is taken as the change in the water content of the tree.

[0112] The water evaporation mass and moisture content change calculated in this step are also the values ​​of a branch at one level in the tree model. When considering the water evaporation mass and moisture content change of the entire tree model, it is necessary to calculate one branch at each level separately and summarize the water evaporation mass and moisture content change of each level as the water evaporation mass and moisture content change of the entire tree model.

[0113] Step 4: Based on the real-time data, the aligned tree moisture content, and the voltage and current data collected in the experimental environment, the relationship between the tree conductivity, moisture content, and temperature in the entire process is fitted to obtain the conductivity function.

[0114] The relationship between the conductivity, moisture content, and temperature of trees in the entire process is fitted using the least squares method. The fitted relationship between conductivity, temperature, and moisture content is expressed as follows:

[0115] lnσ t (W,T)=7.458W+0.057T-0.050W·T-10.754

[0116] Where σ t(W,T) is the conductivity calculated relative to the temperature and moisture content at time t, where W is the tree moisture content and T is the tree temperature.

[0117] Step 5: Use the conductivity function to construct a resistance simulation curve according to the water content and temperature of the tree, and use the voltage and current data collected in the experimental environment to construct a resistance measurement curve.

[0118] Step 6: Compare the resistance simulation curve and the resistance measurement curve, and determine the moment when there is an obvious bifurcation between the resistance simulation curve and the resistance measurement curve as the moment when the open flame appears.

[0119] The method for determining obvious bifurcation includes a difference index determination method and a visual analysis method;

[0120] The difference index is a relative index, and the calculation method is as follows:

[0121]

[0122] Where Δy t is the difference index at time t, y1(t) is the resistance value corresponding to the resistance simulation curve at time t, y2(t) is the resistance value corresponding to the resistance measured curve at time t, ∈ is a preset parameter used to avoid the denominator being zero; when Δy t When the preset threshold is exceeded, it is determined that the resistance simulation curve and the resistance measurement curve have obvious bifurcation.

[0123] The visual analysis rule maps the resistance simulation curve and the resistance measured curve to the same coordinate system, and outputs the coordinate system image in real time for manual intervention and judgment, such as Figure 5 、 6 As shown in the figure, after manual intervention, when the resistance curve develops to the moment corresponding to the circled part in the coordinate system, it is easy to determine that the resistance simulation curve and the resistance measured curve have obvious bifurcation.

[0124] In some other embodiments of the present invention, the method for determining obvious bifurcation also adopts statistical analysis method, statistical detection method, and time point positioning method.

[0125] Statistical analysis method is used to conduct statistical analysis on the changes in difference indicators to determine the moment and significance of bifurcation. For example, the sliding mean or variance of the difference indicator is calculated and whether it increases significantly is determined. A threshold is set (such as k times the previous difference mean). When the difference indicator exceeds the threshold for the first time, it is considered that bifurcation has occurred.

[0126] The statistical detection method uses statistical methods to determine whether the bifurcation is significant. The time series is divided into the early period (better fitting) and the late period, the mean of the difference index is calculated respectively, and the difference between the early and late periods is compared through the t-test to see whether it is significant, or the sequence change detection (CUSUM) is used to identify the change points of the mean or distribution in the time series.

[0127] The time point positioning method uses dynamic programming methods (such as Pruned Exact Linear Time, PELT) to find the global optimal change point, or uses Bayesian Change Point Detection (BCPD) to obtain the change point.

[0128] This embodiment also sets up two sets of verification experiments. The preset values ​​of verification experiment 1 are: the trunk height is set to 1.52m, the trunk radius is set to 0.051m, the initial moisture content is 0.941, the initial temperature is 31.9℃, the tree specific heat capacity is 1408J / kg*K, the tree density is 705kg / m3, and the specific heat capacity of water is 4200J / kg*K. Compared with the actual whole tree test data, the following results are obtained: Figure 5 The effect shown. Figure 5 In the coordinate system, the blue line is the measured resistance curve, and its resistance value is calculated by the voltage and current data collected by the voltage sensor and the current sensor; the orange line is the simulated resistance curve, and its resistance value is the R calculated by the above formula. t The corresponding moment of the real scene on the right side of the picture is the circled moment in the coordinate system.

[0129] The preset values ​​for the second verification experiment are: trunk height of 2.05m, trunk radius of 0.032m, initial moisture content of 0.713, initial temperature of 35.1℃, tree specific heat capacity of 1495J / kg*K, tree density of 1050kg / m3, and water specific heat capacity of 4200J / kg*K. Comparing with the actual whole tree test data, we get the following: Figure 6 The effect is shown. Figure 6 In the coordinate system, the blue line is the measured resistance curve, and its resistance value is calculated by the voltage and current data collected by the voltage sensor and the current sensor; the orange line is the simulated resistance curve, and its resistance value is the R calculated by the above formula. t The corresponding moment of the real scene on the right side of the picture is the circled moment in the coordinate system.

[0130] from Figure 5 、 6 It can be seen that the method described in this embodiment can be used to accurately estimate the time of fire ignition, effectively identify the time of fire occurrence, help analyze the source of the fault and the fire mechanism, clarify the responsibility for the fire, and provide a scientific basis for subsequent fire prevention measures and equipment protection strategies.

[0131] Example 2

[0132] This embodiment is a system for determining the fire ignition moment of a distribution overhead conductor tree-touching fault. The system is used to implement the determination method described in the first embodiment, including:

[0133] The experimental environment module is used to simulate the entire process of a distribution overhead conductor contacting a tree fault, including the neutral point ungrounded simulation part and the conductor contacting a tree simulation part;

[0134] The parameter acquisition module further includes: a voltage and current acquisition unit for collecting phase voltage and ground current data in the experimental environment in a time-aligned manner; a temperature acquisition unit for collecting the temperature of the trees during the experiment and saving them in a time-aligned manner;

[0135] The data processing module further includes: a water evaporation calculation unit for calculating the water evaporation heat absorption of the tree model based on the real-time collected data, and thereby calculating the water evaporation mass, which is used to update the real-time water content of the tree; a conductivity fitting unit for fitting the relationship between the tree conductivity, water content and temperature based on the experimental data to generate a conductivity function;

[0136] The resistance curve generation module further includes: a resistance simulation curve construction unit for constructing a resistance simulation curve based on the conductivity function and the moisture content and temperature of the tree; a resistance measurement curve construction unit for constructing a resistance measurement curve based on the voltage and current data collected in the experimental environment;

[0137] The determination module is used to compare the resistance simulation curve with the resistance measurement curve, and determine the moment when the resistance simulation curve and the resistance measurement curve have obvious bifurcation as the moment when the open flame appears.

[0138] The above description is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this field, several variations and improvements can be made without departing from the creative concept of the present invention, which all fall within the scope of protection of the present invention.

Claims

1. A method for determining the fire moment of a distribution overhead wire tree-touching fault, characterized in that: The following steps are involved: Build an experimental environment to simulate the entire process of a distribution overhead conductor tree-touching fault, and collect data from the entire process at a preset sampling frequency in the experimental environment; Obtain the initial moisture content of the experimental trees, continuously collect and align the complete process data by time during the experiment, and generate real-time data; Calculate the water evaporation heat absorption of the tree model according to the real-time data, calculate the water evaporation mass according to the water evaporation heat absorption, use the water evaporation mass to update the tree moisture content and save it in time alignment; According to the real-time data, the aligned tree moisture content and the voltage and current data collected in the experimental environment, the relationship between the tree conductivity, moisture content and temperature in the whole process is fitted to obtain the conductivity function; Using the conductivity function to construct a resistance simulation curve according to the water content and temperature of the tree, and using the voltage and current data collected in the experimental environment to construct a resistance measurement curve; The resistance simulation curve and the resistance measurement curve are compared, and the moment when there is an obvious bifurcation between the resistance simulation curve and the resistance measurement curve is determined as the moment when the open flame appears.

2. A method for determining the fire moment of a distribution overhead wire tree-touching fault according to claim 1, characterized in that: The experimental environment includes a neutral point ungrounded simulation part, a wire touching a tree simulation part and a data acquisition part; the neutral point ungrounded simulation part is a three-phase overhead wire with power; the wire touching a tree part includes trees transplanted to the test site, and one phase is led out from the three-phase overhead wire and connected to the tree; the data acquisition part continuously records the phase voltage and grounding current of the entire process, records the test phenomena of the entire ignition process, and records the temperature changes of the trees until the discharge channel penetrates the vegetation and a short circuit occurs.

3. A method for determining the fire moment of a distribution overhead wire tree-touching fault according to claim 1, characterized in that: The full process data includes the conductor phase voltage, the current passing through the experimental trees, and the temperature of the experimental trees.

4. A method for determining the fire moment of a distribution overhead wire tree-touching fault according to claim 1, characterized in that: The method for obtaining the initial moisture content includes experimental weighing and preset values; The calculation formula for experimental weighing is: In the formula, m s is the fresh weight of the wood block obtained by experimental sampling, and m0 is the absolute dry mass of the wood block; The preset values ​​are derived from the existing data of the corresponding tree species in the experimental site during the experimental season.

5. A method for determining the fire moment of a distribution overhead wire tree-touching fault according to claim 1, characterized in that: The experimental process includes the tree temperature rise stage, water evaporation stage, carbonization development stage and open flame flashover stage in sequence.

6. A method for determining the fire moment of a distribution overhead wire tree-touching fault according to claim 1, characterized in that: The obvious bifurcation determination method includes a difference index determination method and a visual analysis method; The difference index is a relative index, and the calculation method is as follows: In the formula, Δy t is the difference index at time t, y1(t) is the resistance value corresponding to the resistance simulation curve at time t, y2(t) is the resistance value corresponding to the resistance measured curve at time t, ∈ is a preset parameter used to avoid the denominator being zero; When Δy t When the preset threshold is exceeded, it is determined that the resistance simulation curve and the resistance measurement curve have obvious bifurcation.

7. A method for determining the fire moment of a distribution overhead wire tree-touching fault according to claim 1, characterized in that: The relationship between the conductivity, moisture content and temperature of the trees in the whole process is fitted by the least square method. The relationship between the conductivity, temperature and moisture content after fitting is expressed as follows: lnσ t (W,T)=7.458W+0.057T-0.050W·T-10.754 In the formula, σ t (W,T) is the conductivity calculated relative to temperature and moisture content at time t, W is the moisture content of the tree, and T is the temperature of the tree.

8. A method for determining the fire moment of a distribution overhead wire tree-touching fault according to claim 7, characterized in that: The calculation formula of the resistance in the resistance simulation curve is: Where n is the level of the touch line in the tree model, l0 is the basic branch length in the tree model, that is, the trunk height, r0 is the basic branch radius, that is, the trunk radius; The tree model is a fractal network model simplified from the experimental tree morphology. In the tree model, α and β are defined as the length ratio and radius ratio of the branches of the k+1th layer network and the kth layer network, respectively, and are expressed as follows: In the formula, l k , l k+1 are the branch lengths of the k-th layer network and the k+1-th layer network, r k 、r k+1 They are the branch radius of the k-th layer network and the branch radius of the k+1-th layer network respectively.

9. A method for determining the fire moment of a distribution overhead wire tree-touching fault according to claim 8, characterized in that: The calculation method of the water evaporation mass is: Calculate the total heat Q generated by the current passing through the tree at time t s : Where U is the phase voltage of the conductor connected to the tree; The heat dissipation and heat absorption of trees include convection heat dissipation, radiation heat dissipation, temperature rise heat absorption and water evaporation heat absorption. After deducting convection heat dissipation and radiation heat dissipation from the total heat output, the remaining part is the total heat absorption. The product of the total heat absorption and the evaporation absorption ratio coefficient is the water evaporation heat absorption, which is expressed as follows: Q w =(Q s -Q c -Q r )·k f k f =(1+e 98-T ) -1 In the formula, Q w The heat absorbed by water evaporation, Q s is the total heat production, Q c is the convective heat dissipation, Q r is the radiation heat dissipation, k f is the evaporation absorption proportional coefficient, T is the tree temperature; Calculate the water evaporation mass based on the heat absorbed by water evaporation: In the formula, Δm is the mass of water evaporation, T w For the water temperature.

10. A system for determining the fire moment of a distribution overhead wire hitting a tree fault, characterized in that: The system is used to implement the determination method according to any one of claims 1 to 9, comprising: The experimental environment module is used to simulate the entire process of the distribution overhead conductor touching the tree fault, including the neutral point ungrounded simulation part and the conductor touching the tree simulation part; The parameter acquisition module further includes: a voltage and current acquisition unit, which is used to collect phase voltage and ground current data in the experimental environment in a time-aligned manner; a temperature acquisition unit, which is used to collect the temperature of the trees during the experiment and save them in a time-aligned manner; The data processing module further includes: a water evaporation calculation unit, which is used to calculate the water evaporation heat absorption of the tree model according to the real-time collected data, and thereby calculate the water evaporation mass, which is used to update the real-time water content of the tree; a conductivity fitting unit, which is used to fit the relationship between the conductivity, water content and temperature of the tree according to the experimental data, and generate a conductivity function; The resistance curve generation module further includes: a resistance simulation curve construction unit, which is used to construct a resistance simulation curve based on the conductivity function and the water content and temperature of the tree; a resistance measurement curve construction unit, which is used to construct a resistance measurement curve based on the voltage and current data collected in the experimental environment; The determination module is used to compare the resistance simulation curve with the resistance measurement curve, and determine the moment when the resistance simulation curve and the resistance measurement curve have obvious bifurcation as the moment when the open flame appears.

Images