Power cable path searching method and device based on Internet platform

Abstract

The invention relates to the field of power cable path searching, in particular to a power cable path searching method and device based on an internet platform, and the method comprises the steps: a transmitter injects a signal, a receiver processes noise and errors through Kalman filtering, a built-in thermal imaging camera collects a thermal image, and an initial path is generated by means of internet platform data and an A * algorithm; a genetic algorithm optimizes a path to balance the burial depth and the construction cost, and a convolutional neural network is utilized to analyze a thermal image to position an overheat point; the device comprises a signal acquisition module, an initial path generation module, an optimization module and a fault determination module. According to the invention, signals can be accurately processed, paths can be efficiently planned, faults can be timely detected, data integration and real-time updating can be realized based on an internet platform, and the intelligence and reliability level of power cable operation and maintenance can be improved.

Description

Technical Field

[0001] The present invention relates to the field of power cable path finding, and specifically to a power cable path finding method and device based on an Internet platform. Background Art

[0002] In the field of power system operation and maintenance, accurately finding the power cable path and timely warning of faults are key links to ensure the stable operation of the power grid. With the expansion of the urban power grid scale and the increase in the laying density of underground cables, the limitations of traditional cable path finding and fault detection methods have become increasingly prominent;

[0003] Traditional methods mainly rely on electromagnetic induction equipment or manual markings, and judge the cable direction by emitting low-frequency signals and receiving feedback. This method is easily affected by environmental factors such as underground metal obstacles and electromagnetic interference, resulting in insufficient positioning accuracy. In addition, traditional solutions rely on offline drawings or local GIS systems, with lagging data updates, and cannot integrate dynamic information such as construction feedback and geological changes in real time. There are data island problems during cross-regional operation and maintenance, and the collaboration efficiency is low; in terms of fault detection, when analyzing thermal imaging maps using traditional threshold methods, only obvious temperature anomaly areas can be identified, and the sensitivity to early latent faults is insufficient;

[0004] Therefore, it is necessary to provide a power cable path finding method and device based on an Internet platform to solve the above technical problems. Summary of the Invention

[0005] To solve the technical problems raised in the above background art, the present invention provides a power cable path finding method and device based on an Internet platform.

[0006] The objectives of the present invention can be achieved through the following technical solutions:

[0007] The first aspect of the present invention provides a power cable path finding method based on an Internet platform, specifically including the following steps:

[0008] A transmitter injects a 1kHz + 10kHz signal into the target cable, and a receiver receives the signal, introduces the Kalman filter algorithm to process random noise and measurement errors, and extracts effective features;

[0009] The receiver is built-in with a thermal imaging camera to collect the cable thermal imaging map in real time;

[0010] Use the A* algorithm to generate the initial path of the cable, and use the genetic algorithm to optimize the initial path to find the optimal cable path plan;

[0011] Use a convolutional neural network to analyze the thermal imaging map of the target cable to locate the overheating point;

[0012] The present invention is further configured to connect the transmitter to the target cable, inject a 1kHz + 10kHz signal as the signal frequency into the target cable, the receiver uses a phase-locked loop to acquire the 1kHz + 10kHz signal, amplify the acquired signal through an amplifier, and adjust the gain Gamp of the amplifier according to the intensity of the received signal to ensure that the amplitude of the amplified signal is within a suitable range;

[0013] Establish a state space model, introduce the gain signal into the Kalman filter algorithm, the Kalman filter calculates the Kalman gain, updates the state estimate and the error covariance, and the state space model construction logic: Let the true state of the cable signal be Q k , representing the amplitude of the signal at time k, and the true state transition equation is expressed as Q k =Q k-1 +W k-1 , where W k-1 is the process noise; the observation equation is E k =H k E k-1 +R k-1 , where R k-1 is the measurement noise, and H k is the observation matrix; combine the true state transition equation and the observation equation to construct the state space model; the Kalman filter performs state estimation based on the state space model, and according to the state estimate at the previous moment, then use the state transition equation to predict the state estimate at the current moment, and the formula is expressed as where F k is the state transition matrix, G k is the control input matrix, and U k is the control input vector; predict the error covariance P k | k-1 at the current moment, and the formula is expressed as P k | k-1 =F k P k-1 | k-1 F k +Q k ; according to the error covariance P k | k-1 at the current moment and the observation noise covariance R k calculate the Kalman gain K k , and the Kalman gain calculation logic: K k =P k | k-1 H k (H k P k | k-1 +Rk ) -1 ; Based on the observed value E at the current moment k Update the predicted state estimate to obtain the state estimate at the current moment Update the state calculation logic: Update the error covariance at the current moment, and update the error calculation logic: P k | k =(I - K k H k )P k | k-1 , where I is the identity matrix. By continuously performing the prediction and update steps until the correction of the random noise and measurement error of the signal by the Kalman filter is completed;

[0014] The receiver built-in thermal imaging camera generates a cable thermal imaging map in real time. The port of the receiver is connected to the Internet platform, and the cable thermal imaging map is sent to the Internet platform.

[0015] The present invention is further configured to extract the map information of the target cable and the historical cable path data in the Internet platform, generate the initial path of the cable through the A* algorithm, and then use the genetic algorithm to optimize the initial path to find the optimal cable path; extract the historical cable path data in the Internet platform and the map information of the target cable. The map information includes the cable laying area and obstacle distribution data. The map is divided into discrete nodes vi, and vi represents each position node on the map. Connect any two nodes and set them as the edge set, so as to obtain the edge sets of the map, obtain the movement cost corresponding to each edge set, obtain the starting point s and the target point z of the cable laying, and then initialize the open list and the closed list. The open list is used to store the nodes to be explored. Initially, the starting point s is added to the open list. The closed list is used to store the nodes that have been explored and is initially empty; initialize the node cost. For the starting point s, set the actual cost from the starting point to itself as g(s)=0, calculate the estimated cost h(s) from the starting point to the target point, and the calculation logic is h(s)=|x s -x z |+|y s -y z |, where (x s , y s ) is the coordinate of the starting point, (x z , y z ) is the coordinate of the target point. The total cost f(s) of the starting point is calculated as: f(s)=g(s)+h(s); set a parent node pointer for each node, and the parent node pointer of the starting point s is initially set to null, indicating that there is no parent node;

[0016] When the open list is not empty, select the node (n) with the minimum total cost f(n) from the open list. f(n) represents the total cost of the path from the starting point to the target point passing through the node. Remove it from the open list and add it to the closed list. Check whether the current node (n) is the target point z. If it is the target point z, the algorithm ends, and the path from the starting point to the target point is obtained by backtracking the parent node pointer; if it is not the target point z, find all adjacent nodes m of the current node (n). Adjacent nodes are the nodes connected to the node (n) by an edge; for each adjacent node m, perform the following operations: Calculate the actual cost g temp (m) from the starting point s through the current node (n) to the adjacent node m. g temp (m) = g(n) + w nm , where w nm represents the edge weight from the node (n) to the node m; if the adjacent node m is in the closed list, it means that the node has been explored, and skip this node; if the adjacent node m is not in the open list, add it to the open list, and set g(m) = g temp (m). Calculate h(m) and then get f(m) = h(m) + g(m) to obtain the total cost of the adjacent node m. The steps of calculating h(m) are the same as those of calculating h(s). At the same time, set the node (n) as the parent node of the node m;

[0017] If the adjacent node m is in the open list, compare the sizes of g temp (m) and g(m): if g temp (m) < g(m), it means that the path through the current node (n) to the node m is better. Update g(m) = g temp (m), recalculate f(m) = h(m) + g(m), and set the node (n) as the parent node of the node m; if g temp (m) ≥ g(m), then do not update the information of the node m; when the target point z is found, starting from the target point, backtrack through the parent node pointer, and visit the parent nodes of each node in turn until returning to the starting point s. The sequence of nodes passed through is the path from the starting point to the target point. The algorithm ends and outputs the obtained initial path;

[0018] Randomly perturb the initial path through the genetic algorithm, and repeat the above steps multiple times to generate a specified number of initial populations P = {C1, C2,.....C N}., C represents the candidate path obtained after random perturbation, N represents the total number of candidate paths. For each node vi on the path, determine the cable burial depth d i at this node, set the geological condition influence factor as q i , then the burial depth calculation logic: Set the construction difficulty coefficient between adjacent nodes as r i, i + 1, then the construction difficulty cost logic: Then, according to the formula Cost = Cost depth × 0.6 + Cost difficulty × 0.4 to obtain the comprehensive cost function, where 0.6 is the fixed weight constant of the burial depth cost and 0.4 is the fixed weight constant of the construction difficulty cost;

[0019] The fitness function is inversely proportional to the comprehensive cost function, expressed as Calculate the fitness Fitness(C u ) of each candidate path, and the total fitness of all chromosomes in the population. The calculation logic of the total fitness: u represents the number of the candidate path, and the probability of each candidate path being selected is p u , and the calculation logic of the selection probability: Randomly select a chromosome, generate a random number r between 0 and 1, and accumulate the selection probability p of each chromosome in turn u . When the accumulated value is greater than or equal to r, select the corresponding optimal candidate path;

[0020] Perform a mutation operation on the corresponding optimal candidate path and iteratively output the optimal cable path: Randomly select a node vj in the corresponding optimal candidate path and replace it with another legal node v in the map new , and continuously repeat the above steps, setting the fitness threshold When in the population , the algorithm ends and outputs the cable path corresponding to the optimal candidate path.

[0021] The present invention is further configured to use a CNN network architecture model, adopt a three-layer convolution-activation-pooling structure, with the number of convolution kernels in each layer being 32, 64, and 128 in sequence, and the stride being 1; the fully connected layer outputs a hot spot probability map, performs iteration through the binary cross-entropy loss function, and uses the Adam optimizer with a batch size of 32 to train for 50 rounds. The data is divided into a 70% training set, a 20% validation set, and a 10% test set, and the training is terminated when the validation set loss does not decrease for 10 consecutive rounds;

[0022] The Internet platform inputs the thermal imaging map into the convolutional layer, sets the hot spot probability threshold θ = 0.9 to locate the high-confidence region. The pooling layer outputs a pixel-level hot spot probability map, superimposes the probability heat map on the original thermal imaging map, and marks the hot spot region with a rectangular box, outputting a coordinate list.

[0023] Please refer to Figure 2 shown. The second aspect of the present invention provides a power cable path finding device based on an Internet platform, specifically as follows:

[0024] The signal acquisition module injects a 1kHz + 10kHz signal into the target cable according to the transmitter. The receiver receives the signal and introduces the Kalman filtering algorithm to process random noise and measurement errors, and extracts effective features;

[0025] The initial path generation module generates an initial path according to the A* algorithm combined with the map information of the target cable and the historical cable path data;

[0026] The optimization module optimizes the initial path according to the genetic algorithm, simulates the evolution mechanism, and outputs the optimal path plan after selection and mutation iteration;

[0027] The fault determination module locates the overheating point according to the convolutional neural network analysis of the thermal imaging map of the target cable.

[0028] Compared with the prior art, the beneficial effects of the present invention are as follows: Based on the map information and historical cable path data integrated by the Internet platform, the A* algorithm quickly generates an initial path through a heuristic function, while avoiding the problems of manual dependence and drawing lag; The genetic algorithm is introduced to construct a comprehensive cost function to comprehensively consider geological conditions and construction feasibility, and improve the path safety;

[0029] The receiver is built-in with a thermal imaging camera to collect the thermal imaging map of the cable in real time and send it to the Internet platform. The convolutional neural network is used to analyze the thermal imaging map to locate the overheating point. By using a specific CNN network architecture model, through training and optimization, the overheating point of the cable can be accurately identified, and potential cable fault hazards can be discovered in time. At the same time, by setting the overheating point probability threshold, the high-confidence area is located, and the probability heat map is superimposed on the original thermal imaging map, the overheating point area is marked with a rectangular frame and the coordinate list is output, which is convenient for the operation and maintenance personnel to quickly locate and handle faults, and effectively guarantees the safe operation of the cable. Brief Description of the Drawings

[0030] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. The following drawings are not deliberately drawn to scale in actual size, and the focus is on showing the gist of the present invention.

[0031] Figure 1 It is a block diagram of the method steps of the present invention.

[0032] Figure 2 It is a module connection diagram of the invention. Detailed Embodiments

[0033] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only partial embodiments of the present invention, rather than all embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts also belong to the scope of protection of the present invention.

[0034] Please refer to Figure 1 As shown, the first aspect of the present invention provides a method for finding the path of a power cable based on an Internet platform, which specifically includes the following steps:

[0035] The transmitter injects a 1kHz + 10kHz signal into the target cable, and the receiver receives the signal, introduces the Kalman filtering algorithm to process random noise and measurement errors, and extracts effective features;

[0036] The receiver is built-in with a thermal imaging camera to collect the thermal imaging diagram of the cable in real time;

[0037] Use the A* algorithm to generate the initial path of the cable, and use the genetic algorithm to optimize the initial path to find the optimal cable path plan;

[0038] Use a convolutional neural network to analyze the thermal imaging diagram of the target cable to locate the overheating point;

[0039] The present invention is further configured to connect the transmitter to the target cable, then inject a 1kHz + 10kHz signal as the signal frequency into the target cable, the receiver uses a phase-locked loop to acquire the 1kHz + 10kHz signal, and the acquired signal is amplified by an amplifier. The gain Gamp of the amplifier is adjusted according to the intensity of the received signal to ensure that the amplitude of the amplified signal is within a suitable range;

[0040] Establish a state space model, and introduce the amplified signal into the Kalman filtering algorithm. The Kalman filter calculates the Kalman gain, updates the state estimate value and the error covariance. The construction logic of the state space model: Let the true state of the cable signal be Q k , representing the amplitude of the signal at time k, the true state transition equation is expressed as Q k = Q k-1 + W k-1 , where W k-1 is the process noise; the observation equation is E k = H k E k-1 + R k-1 , where R k-1 is the measurement noise, and H k is the observation matrix; Combine the true state transition equation and the observation equation to construct a state space model; The Kalman filter performs state estimation based on the state space model, according to the state estimate value at the previous moment Then, use the state transition equation to predict the state estimate value at the current moment The formula is expressed as where F k is the state transition matrix, G k is the control input matrix, and U k is the control input vector; predict the error covariance P k | k-1 at the current moment, and the formula is expressed as P k | k-1 = F k P k-1 | k-1 F k + Q k ; According to the error covariance P k | k-1 at the current moment and the observation noise covariance R k calculate the Kalman gain K k , and the calculation logic of the Kalman gain: K k = P k | k-1 H k (H k P k | k-1 + R k ) -1 , and the Kalman gain determines the weight of the observation value when updating the state estimate value; update the predicted state estimate value k according to the observation value E at the current moment to obtain the state estimate value at the current moment. The update state calculation logic is as follows: Update the error covariance at the current moment, and the update error calculation logic: P k | k = (I - K k H k )P k | k-1, where I is the identity matrix. By continuously performing prediction and update steps until the correction of the random noise and measurement error of the signal by the Kalman filter is completed. It should be noted that during the cable signal acquisition process, the signals received by the receiver are often interfered by various random noises, such as stray signals in the electromagnetic environment and thermal noise inside electronic devices. These noises will distort the signals and affect the subsequent extraction and analysis of cable status information. The Kalman filter algorithm describes the dynamic changes of the signal through a state space model, and can optimally estimate the true state of the signal based on the historical information and current observations of the signal. In the prediction step, the algorithm predicts the next state of the signal based on the state transition matrix. In the update step, combining the current observation value and the predicted value, the Kalman gain is calculated and the predicted value is corrected to suppress the influence of random noise and improve the signal quality;

[0041] The receiver is built-in with a thermal imaging camera to generate a cable thermal imaging map in real time. The port of the receiver is connected to the Internet platform, and the cable thermal imaging map is sent to the Internet platform.

[0042] The present invention is further configured to extract the map information of the target cable and the historical cable path data in the Internet platform, generate an initial path of the cable through the A* algorithm, and then use the genetic algorithm to optimize the initial path to find the optimal cable path; extract the historical cable path data in the Internet platform and the map information of the target cable. The map information includes the cable laying area and obstacle distribution data. The map is divided into discrete nodes vi, and vi represents each position node on the map. Connect any two nodes and set them as the edge set, so as to obtain each edge set of the map, obtain the movement cost corresponding to each edge set, and the movement cost includes factors such as distance, terrain complexity, and obstacles; obtain the starting point s and the target point z of the cable laying, and then initialize the open list and the closed list. The open list is used to store the nodes to be explored. Initially, the starting point s is added to the open list. The closed list is used to store the nodes that have been explored and is initially empty; initialize the node cost. For the starting point s, set the actual cost from the starting point to itself as g(s)=0, calculate the estimated cost h(s) from the starting point to the target point, and the calculation logic is h(s)=|x s -x z |+|y s -y z |, where (x s , y s ) is the coordinate of the starting point, (x z , y z ) is the coordinate of the target point. The calculation logic of the total cost f(s) of the starting point: f(s)=g(s)+h(s); set a parent node pointer for each node, and the parent node pointer of the starting point s is initialized to null, indicating that there is no parent node;

[0043] When the open list is not empty, select the node (n) with the minimum total cost f(n) from the open list. f(n) represents the total cost of the path from the starting point to the target point passing through the node. Remove it from the open list and add it to the closed list. Check whether the current node (n) is the target point z. If it is the target point z, the algorithm ends, and the path from the starting point to the target point is obtained by backtracking the parent node pointer; if it is not the target point z, find all adjacent nodes m of the current node (n). An adjacent node is a node connected to node (n) by an edge. For each adjacent node m, perform the following operations: Calculate the actual cost g temp (m) from the starting point s through the current node (n) to the adjacent node m. g temp (m) = g(n) + w nm , where w nm represents the edge weight from node (n) to node m; if the adjacent node m is in the closed list, it means that the node has been explored, and skip this node; if the adjacent node m is not in the open list, add it to the open list, set g(m) = g temp (m), calculate h(m), and then obtain f(m) = h(m) + g(m) to get the total cost of the adjacent node m. The steps of calculating h(m) are the same as those of calculating h(s). At the same time, set node (n) as the parent node of node m;

[0044] If the adjacent node m is in the open list, compare the sizes of g temp (m) and g(m): If g temp (m) < g(m), it means that the path through the current node (n) to node m is better. Update g(m) = g temp (m), recalculate f(m) = h(m) + g(m), and set node (n) as the parent node of node m; if g temp (m) ≥ g(m), then do not update the information of node m; when the target point z is found, starting from the target point, backtrack through the parent node pointer, and visit the parent nodes of each node in turn until returning to the starting point s. The sequence of nodes passed through is the path from the starting point to the target point. The algorithm ends and outputs the obtained initial path;

[0045] Randomly perturb the initial path through the genetic algorithm. For example, randomly select a node on the path, and then select a legal node from its adjacent nodes to replace it, but ensure that the path after replacement is still a legal path from the starting point to the end point. Repeat the above steps multiple times to generate a specified number of initial populations P = {C1, C2,.....C N}., where C represents the candidate path obtained after random perturbation, and N represents the total number of candidate paths. For each node vi on the path, determine the cable burial depth d i, set the geological condition influence factor as q i , which reflects the influence degree of the geology at this location on the burial depth cost. For example, in areas with hard geology, q i is larger, and the burial depth calculation logic is as follows: Set the construction difficulty coefficient between adjacent nodes as r i , i + 1, then the construction difficulty cost logic is as follows: Then, according to the formula Cost = Cost depth ×0.6 + Cost difficulty ×0.4 to obtain the comprehensive cost function, where 0.6 is the fixed weight constant of the burial depth cost, and 0.4 is the fixed weight constant of the construction difficulty cost; it should be noted that the burial depth cost of the cable accounts for a relatively large proportion in the entire cable laying cost. Different burial depth requirements will affect aspects such as the workload of excavating earthwork, the cable protection measures, and the subsequent maintenance difficulty. Although the construction difficulty cost will also affect the comprehensive cost, compared with the burial depth cost, its influence range and degree are relatively small. The construction difficulty mainly involves factors such as the distribution of underground obstacles and the terrain complexity;

[0046] The fitness function is inversely proportional to the comprehensive cost function, expressed as The higher the fitness, the better the candidate path. Calculate the fitness Fitness(C u ) of each candidate path, as well as the total fitness of all chromosomes in the population. The calculation logic of the total fitness is as follows: u represents the number of the candidate path, and the selection probability of each candidate path is p u , and the calculation logic of the selection probability is as follows: Randomly select a chromosome, generate a random number r between 0 and 1, and accumulate the selection probability p of each chromosome in turn u . When the accumulated value is greater than or equal to r, select the corresponding optimal candidate path;

[0047] Perform a mutation operation on the corresponding optimal candidate path and iteratively output the optimal cable path: Randomly select a node vj in the corresponding optimal candidate path and replace it with another legal node v in the map new , continuously repeat the above steps, and set the fitness threshold When in the population , the algorithm ends and outputs the cable path corresponding to the optimal candidate path; it should be noted that the genetic algorithm is an optimization algorithm that simulates natural selection and genetic mechanisms. It continuously optimizes the burial depth and construction cost of the path by simulating the selection and mutation operations in the biological evolution process. During the optimization process, factors such as the actual laying situation of the cable, geological conditions, and construction difficulty are considered to achieve the best path planning effect.

[0048] The present invention is further configured to use a CNN network architecture model, adopt a three-layer convolution-activation-pooling structure, with the number of convolution kernels in each layer being 32, 64, and 128 in sequence, and the stride being 1; the fully connected layer outputs a hot spot probability map, which is iterated through the binary cross-entropy loss function, and is trained for 50 rounds with a batch size of 32 through the Adam optimizer. The data is divided into a 70% training set, a 20% validation set, and a 10% test set. Training is terminated when the validation set loss does not decrease for 10 consecutive rounds;

[0049] The Internet platform inputs the thermal imaging map into the convolutional layer, sets the hot spot probability threshold θ = 0.9 to locate the high-confidence region. The pooling layer outputs a pixel-level hot spot probability map, superimposes the probability heat map on the original thermal imaging map, and marks the hot spot region with a rectangular box, and outputs a coordinate list.

[0050] Please refer to Figure 2 As shown, the second aspect of the present invention provides a power cable path finding device based on an Internet platform, specifically as follows:

[0051] The signal acquisition module injects a 1kHz + 10kHz signal into the target cable according to the transmitter, and the receiver receives the signal and introduces the Kalman filtering algorithm to process random noise and measurement errors, and extracts effective features;

[0052] The initial path generation module generates an initial path according to the A* algorithm in combination with the map information of the target cable and the historical cable path data;

[0053] The optimization module optimizes the initial path according to the genetic algorithm, simulates the evolution mechanism, and outputs the optimal path plan through selection and mutation iteration;

[0054] The fault determination module locates the hot spot by analyzing the thermal imaging map of the target cable through a convolutional neural network.

[0055] The above is the description of the present invention and should not be regarded as a limitation thereof. Although several exemplary embodiments of the present invention have been described, those skilled in the art will easily understand that many modifications can be made to the exemplary embodiments without departing from the novel teachings and advantages of the present invention. Therefore, all such modifications are intended to be included within the scope of the present invention defined by the claims. It should be understood that the above is the description of the present invention and should not be considered limited to the specific embodiments disclosed, and modifications to the disclosed embodiments and other embodiments are intended to be included within the scope of the appended claims. The present invention is defined by the claims and their equivalents.

Claims

1. A method for finding the path of a power cable based on an Internet platform, characterized in that, It includes the following steps: The transmitter injects a 1kHz + 10kHz signal into the target cable, and the receiver receives the signal and introduces the Kalman filtering algorithm to process random noise and measurement errors; The receiver is built-in with a thermal imaging camera to collect the thermal imaging map of the cable in real time; Use the A* algorithm to generate the initial path of the cable, and use the genetic algorithm to optimize the initial path to find the optimal cable path plan; Use a convolutional neural network to analyze the thermal imaging map of the target cable to locate the overheating point.

2. The method for finding the path of a power cable based on an Internet platform according to claim 1, characterized in that, Extract the map information and historical cable path data of the target cable in the Internet platform, and generate the initial path of the cable through the A* algorithm. The specific steps are as follows: Extract the historical cable path data in the Internet platform and the map information of the target cable. The map information includes the cable laying area and obstacle distribution data. Divide the map into discrete nodes vi, where vi represents each position node on the map. Connect any two nodes and set them as the edge set, so as to obtain each edge set of the map, obtain the movement cost corresponding to each edge set, obtain the starting point s and the target point z of the cable laying, and then initialize the open list and the closed list. The open list is used to store the nodes to be explored. Initially, add the starting point s to the open list. The closed list is used to store the nodes that have been explored and is initially empty; Initialize the node cost. For the starting point s, set the actual cost from the starting point to itself as g(s) = 0, calculate the estimated cost h(s) from the starting point to the target point, and the calculation logic is h(s) = |x s -x z | + |y s -y z |, where (x s , y s ) is the coordinate of the starting point, (x z , y z ) is the coordinate of the target point. The calculation logic for the total cost f(s) of the starting point is: f(s) = g(s) + h(s); Set a parent node pointer for each node, and initialize the parent node pointer of the starting point s to null, indicating that there is no parent node; When the open list is not empty, select the node (n) with the minimum total cost f(n) from the open list. f(n) represents the total cost of the path from the starting point to the target point passing through the node. Remove it from the open list and add it to the closed list. Check whether the current node (n) is the target point z. If it is the target point z, the algorithm ends, and the path from the starting point to the target point is obtained by backtracking the parent node pointer; if it is not the target point z, find all adjacent nodes m of the current node (n). Adjacent nodes are the nodes connected to the node (n) by edges; for each adjacent node m, perform the following operations: Calculate the actual cost g temp (m) from the starting point s through the current node (n) to the adjacent node m. g temp (m) = g(n) + w nm , where w nm represents the edge weight from the node (n) to the node m; if the adjacent node m is in the closed list, it means that the node has been explored, and skip this node; if the adjacent node m is not in the open list, add it to the open list, and set g(m) = g temp (m), calculate h(m) and then obtain f(m) = h(m) + g(m) to get the total cost of the adjacent node m. The steps of calculating h(m) are the same as those of calculating h(s). At the same time, set the node (n) as the parent node of the node m; If the adjacent node m is in the open list, compare the magnitudes of g temp (m) and g(m): If g temp (m) < g(m), it indicates that the path through the current node (n) to node m is better. Update g(m) = g temp (m), recalculate f(m) = h(m) + g(m), and set node (n) as the parent node of node m; if g temp (m) ≥ g(m), then do not update the information of node m; when the target point z is found, starting from the target point, backtrack through the parent node pointers and visit the parent nodes of each node in turn until returning to the starting point s. The sequence of nodes passed through is the path from the starting point to the target point. The algorithm ends and outputs the obtained initial path.

3. The method for finding the path of a power cable based on an Internet platform according to claim 1, wherein Use the genetic algorithm to calculate the fitness function and the comprehensive cost function of the initial path, and select the optimal candidate path. The specific steps are as follows: Randomly perturb the initial path through a genetic algorithm and repeat the above steps multiple times to generate a specified number of initial populations P = {C1, C2,.....C N}, where C represents the candidate path obtained after random perturbation, N represents the total number of candidate paths. For each node vi on the path, determine the cable burial depth d i at this node, and set the geological condition influence factor as q i . Then the burial depth calculation logic is: Set the construction difficulty coefficient between adjacent nodes as r i , i + 1. Then the construction difficulty cost logic is: Then, according to the formula Cost = Cost depth ×0.6 + Cost difficulty ×0.4 to obtain the comprehensive cost function Cost, where 0.6 is the fixed weight constant of the burial depth cost, and 0.4 is the fixed weight constant of the construction difficulty cost; The fitness function is inversely proportional to the comprehensive cost function, expressed as Calculate the fitness Fitness(C u ) of each candidate path, as well as the total fitness of all chromosomes in the population. The calculation logic for the total fitness is as follows: u represents the number of the candidate path, and the selection probability of each candidate path is p u , and the calculation logic for the selection probability is as follows: Randomly select a chromosome, generate a random number r between 0 and 1, and accumulate the selection probability p of each chromosome in turn u . When the accumulated value is greater than or equal to r, select the corresponding optimal candidate path.

4. The method for finding the path of a power cable based on an Internet platform according to claim 1, wherein Perform a mutation operation on the corresponding optimal candidate path and iterate to output the optimal cable path. Specifically: Randomly select a node vj in the corresponding optimal candidate path and replace it with another legal node v in the map new , and continuously repeat the above steps, setting the fitness threshold When the fitness value of the optimal candidate path in the population ≥ , the algorithm ends and outputs the cable path corresponding to the optimal candidate path 5. The method for finding the path of a power cable based on an Internet platform according to claim 1, wherein The transmitter injects a 1kHz + 10kHz signal into the target cable. Specifically: Connect the transmitter to the target cable, and then inject the 1kHz + 10kHz signal as the signal frequency into the target cable. The receiver uses a phase-locked loop to obtain the 1kHz + 10kHz signal, and amplifies the obtained signal through an amplifier. The gain Gamp of the amplifier is adjusted according to the intensity of the received signal to ensure that the amplitude of the amplified signal is within an appropriate range.

6. The method for finding the path of a power cable based on an Internet platform according to claim 5, characterized in that, The receiver receives the signal and introduces the Kalman filtering algorithm to process random noise and measurement errors. Specifically: Establish a state space model, introduce the amplified signal into the Kalman filtering algorithm, calculate the Kalman gain by Kalman filtering, update the state estimate and error covariance. The construction logic of the state space model is as follows: Let the true state of the cable signal be Q k , representing the amplitude of the signal at time k. The true state transition equation is expressed as Q k = Q k-1 + W k-1 , where W k-1 is the process noise; the observation equation is E k = H k E k-1 + R k-1 , where R k-1 is the measurement noise, and H k is the observation matrix; combine the true state transition equation and the observation equation to construct the state space model; Kalman filtering performs state estimation based on the state space model. According to the state estimate at the previous moment, then use the state transition equation to predict the state estimate at the current moment. The formula is expressed as , where F k is the state transition matrix, G k is the control input matrix, and U k is the control input vector; predict the error covariance P k | k-1 at the current moment. The formula is expressed as P k | k-1 = F k P k-1 | k-1 F k + Q k ; according to the error covariance P k | k-1 at the current moment and the observation noise covariance R k , calculate the Kalman gain K k . The calculation logic of the Kalman gain is: K k = P k | k-1 H k (H k P k | k-1 + R k ) -1 ; update the predicted state estimate k according to the observation value E at the current moment to obtain the state estimate at the current moment. The update state calculation logic is: Update the error covariance at the current moment. The update error calculation logic is: P k | k =(I - K k H k )P k | k-1 , where I is the identity matrix, and by continuously performing prediction and update steps until the correction of the random noise and measurement error of the signal by the Kalman filter is completed; the receiver is built with a thermal imaging camera to generate a cable thermal imaging map in real time, and the port of the receiver is connected to the Internet platform to send the cable thermal imaging map to the Internet platform.

7. The method for finding the path of a power cable based on an Internet platform according to claim 1, wherein Use the CNN network architecture model, adopt a 3-layer convolution-activation-pooling structure, the number of convolution kernels in each layer is 32, 64, and 128 in turn, and the stride is 1; the fully connected layer outputs the overheating point probability map, iterates through the binary cross-entropy loss function, and uses the Adam optimizer with a batch size of 32 and trains for 50 rounds. Divide the data into 70% training set, 20% validation set, and 10% test set. When the validation set loss does not decrease for 10 consecutive rounds, terminate the training; The Internet platform inputs the thermal imaging map into the convolutional layer, sets the overheating point probability threshold θ = 0.9 to locate the high-confidence region. The pooling layer outputs the pixel-level overheating point probability map, superimposes the probability heat map on the original thermal imaging map, and uses a rectangular box to mark the overheating point region and outputs the coordinate list.

8. An electric cable path finding device based on an Internet platform, which is used to implement the electric cable path finding method based on an Internet platform according to any one of claims 1-7, characterized in that, It includes: Signal acquisition module: According to the transmitter injecting a 1kHz + 10kHz signal into the target cable, the receiver receives the signal and introduces the Kalman filtering algorithm to process random noise and measurement errors, and extracts effective features; Initial path generation module: Generate an initial path based on the A* algorithm in combination with the map information of the target cable and the historical cable path data; Optimization module: Optimize the initial path according to the genetic algorithm, simulate the evolutionary mechanism, and output the optimal path plan through selection and mutation iteration; Fault determination module: Locate the overheating point based on the thermal imaging map of the target cable analyzed by the convolutional neural network.

Images