High-safety electrical equipment circuit fault detection method

Abstract

The invention relates to the field of circuit fault detection, in particular to a high-safety electrical equipment circuit fault detection method. According to the method, electrical signal data of a plurality of nodes in a circuit are collected through a sensor, and the electrical signal data comprise signal strength and timestamps; filtering and normalizing the electrical signal data to generate a time delay matrix and a physical distance matrix; according to the use time delay matrix and the physical distance matrix, constructing a cost function of the propagation path and generating a cost matrix; finding the shortest path from each node to other nodes by using a path optimization algorithm, and generating a propagation path diagram; performing fault analysis on each node in the propagation path diagram to determine all fault source nodes; and performing probabilistic reasoning by using a Bayesian network, calculating the posterior probability of each fault source node under the given propagation path and signal propagation time delay, and selecting the node with the highest posterior probability as the final fault source node.

Description

Technical Field

[0001] The present invention relates to the field of circuit fault detection, and in particular to a high-safety circuit fault detection method for electrical equipment. Background Art

[0002] Traditional electrical equipment circuit fault detection methods mainly rely on fixed thresholds, human experience and simple signal processing technology. These methods have many safety hazards in practical applications.

[0003] First, traditional fault detection methods usually use fixed thresholds to determine whether a signal is abnormal. This method is difficult to adapt to the dynamic changes in the circuit operating environment. For example, changes in factors such as temperature, humidity, and load can affect the normal range of the signal, resulting in frequent false alarms or missed alarms.

[0004] Secondly, traditional fault detection methods lack an in-depth understanding of circuit topology and cannot accurately evaluate the propagation characteristics of signals along different paths. This makes it difficult to accurately locate the source of the fault in complex circuits and may misjudge the fault location, thereby delaying fault handling time and increasing safety hazards.

[0005] Furthermore, there may be differences in judgment criteria and processing methods, leading to inconsistencies in fault detection results. Finally, traditional fault detection methods fail to fully utilize information such as time delay and physical distance when dealing with multi-node and multi-path propagation. This makes it difficult to accurately distinguish the true faulty node when there are multiple potential fault sources, thereby reducing the reliability and accuracy of fault detection. Summary of the Invention

[0006] The object of the present invention is to provide a highly secure method for detecting circuit faults in electrical equipment to solve the problems raised in the above background technology.

[0007] To achieve the above object, the present invention provides the following technical solution: a highly secure method for detecting circuit faults in electrical equipment, the method comprising the following steps: S1. Collecting electrical signal data from multiple nodes in a circuit through sensors, where the electrical signal data includes signal strength and timestamp; S2. Filter and normalize the electrical signal data to generate a time delay matrix and a physical distance matrix; construct a cost function of the propagation path based on the time delay matrix and the physical distance matrix; and calculate the cost between each node based on the cost function to generate a cost matrix; S3. Use the path optimization algorithm to find the shortest path from each node to other nodes based on the cost matrix and generate a propagation path graph; perform fault analysis on each node in the propagation path graph to determine all fault source nodes; use the Bayesian network for probabilistic reasoning to calculate the posterior probability of each fault source node under a given propagation path and signal propagation time delay, and select the node with the highest posterior probability as the final fault source node.

[0008] As a further improvement of the present technical solution, the electrical signal data includes signals during normal operation and potential fault signals.

[0009] As a further improvement of the present technical solution, the process of generating the time delay matrix and the physical distance matrix specifically includes: Based on the timestamp information in the electrical signal data collected by the sensor, the signal propagation time delay between each node is calculated. The signal propagation time delay between all nodes is calculated to generate a time delay matrix. The time delay matrix is ​​an N×N matrix, where N is the number of nodes, and each element in the matrix represents the signal propagation time delay between two corresponding nodes. A physical layout diagram of the circuit is obtained from a public database. Based on the physical layout diagram of the circuit, the physical distance between each node is obtained. The physical distances between all nodes are recorded to generate a physical distance matrix. The physical distance matrix is ​​also an N×N matrix, in which each element represents the physical distance between two corresponding nodes.

[0010] As a further improvement of this technical solution, the process of constructing the cost function of the propagation path and generating the cost matrix specifically includes: The cost function is used to evaluate the total cost of electrical signal data propagating from one node to another. This function combines information from the time delay matrix and the physical distance matrix. The cost function multiplies the signal propagation time delay and physical distance by their respective weight parameters, and then adds them together to obtain the total cost. The weight parameters are determined by cross-validation. Using the determined weight parameters, the total cost between each node is calculated according to the cost function of the propagation path; the total cost between all nodes is calculated to generate an N×N cost matrix, where each element in the cost matrix represents the total cost between the corresponding two nodes.

[0011] As a further improvement of this technical solution, the process of generating the propagation path map specifically includes: Initialize the distance matrix, set the distance of all nodes to infinity, and the distance of the starting node to 0; select the node with the smallest distance as the current node; Update the distance from the current node to all adjacent nodes. If the distance to the adjacent node through the current node is shorter, update the distance to the adjacent node. Repeat the above steps until the distances of all nodes are calculated; Based on the calculation results of the shortest path, a propagation path graph is generated, which shows the shortest path from each node to other nodes.

[0012] As a further improvement of this technical solution, the process of determining all fault source nodes specifically includes: Perform preliminary abnormal screening on the signal propagation time delay corresponding to each node in the propagation path diagram. By calculating the average signal propagation time delay of all nodes, if the signal propagation time delay corresponding to each node exceeds 10% of the average signal propagation time delay, it is determined to be a preliminary abnormal node; For each preliminary abnormal node, analyze its impact range by checking the shortest path from the preliminary abnormal node to all other nodes; record the number of other nodes and the number of paths affected by each preliminary abnormal node; exclude the node that only affects one path, and treat the remaining preliminary abnormal nodes as fault source nodes.

[0013] As a further improvement of this technical solution, the process of generating the final fault source node specifically includes: Construct a Bayesian network, where nodes represent physical nodes in the circuit, edges represent physical connections, and the conditional probability table represents the probability of abnormal signals appearing in other nodes when a node fails. A prior probability is set for each fault source node based on historical fault data, and the likelihood function is calculated on each path. For each fault source node, the shortest path from it to all other nodes is checked, the probability of signal anomalies on the path is evaluated, and the probabilities are multiplied to obtain the likelihood function value of the fault source node. The signal strength is normalized to convert its value between 0 and 1. The maximum and minimum values ​​of the signal strength of all nodes are calculated, and the normalization formula is used to convert the signal strength value of each node to a value between 0 and 1. This is done by subtracting the minimum value from each signal strength value and dividing it by the difference between the maximum and minimum values. Using the Bayesian formula, the posterior probability of each fault source node is calculated according to the prior probability of each fault source node. The posterior probability values ​​of all fault source nodes are compared, and the node with the highest posterior probability is selected as the final fault source node.

[0014] Compared with the prior art, the present invention has the following beneficial effects: This highly secure method for detecting circuit faults in electrical equipment utilizes probabilistic reasoning techniques based on sensor acquisition and Bayesian networks, combined with a time delay matrix, a physical distance matrix, and a path optimization algorithm to achieve highly secure and accurate fault detection. The data collected by the sensors includes not only signals during normal operation but also potential fault signals, ensuring comprehensive and real-time fault detection. By defining a cost function for the propagation path and combining information from the time delay matrix and physical distance matrix, the total cost between each node is calculated to generate a cost matrix. This step makes the evaluation of signal propagation paths more comprehensive and accurate, better reflecting the actual propagation characteristics of the signal in the circuit.

[0015] In addition, a path optimization algorithm is used to find the shortest path from each node to other nodes based on the cost matrix, and a propagation path graph is generated. This enables comprehensive consideration of multiple signal propagation paths during fault analysis, thereby improving the accuracy of fault location. By performing fault analysis on each node in the propagation path graph and determining the initial abnormal nodes and their impact range, this method can effectively identify potential fault sources. Bayesian networks are used for probabilistic reasoning to calculate the posterior probability of each fault source node under a given propagation path and signal propagation time delay, and the node with the highest posterior probability is selected as the final fault source node. This process, by combining path analysis and probability models, not only improves the accuracy of fault location, but also effectively distinguishes the true fault node in complex multi-node and multi-path situations. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 Schematic diagram of the method steps of the present invention. DETAILED DESCRIPTION

[0017] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making any creative efforts shall fall within the scope of protection of the present invention.

[0018] See also Figure 1 The present invention provides a technical solution: a highly secure method for detecting circuit faults in electrical equipment, comprising the following steps: S1. Collecting electrical signal data from multiple nodes in a circuit through sensors, where the electrical signal data includes signal strength and timestamp; S2. Filter and normalize the electrical signal data to generate a time delay matrix and a physical distance matrix; construct a cost function of the propagation path based on the time delay matrix and the physical distance matrix; and calculate the cost between each node based on the cost function to generate a cost matrix; S3. Use the path optimization algorithm to find the shortest path from each node to other nodes based on the cost matrix and generate a propagation path graph; perform fault analysis on each node in the propagation path graph to determine all fault source nodes; use the Bayesian network for probabilistic reasoning to calculate the posterior probability of each fault source node under a given propagation path and signal propagation time delay, and select the node with the highest posterior probability as the final fault source node.

[0019] Specifically include: In S1, sensors are used to collect electrical signal data from multiple nodes in the circuit. Sensors are installed at each physical node in the circuit, such as Node A, Node B, Node C, Node D, and Node E. The electrical signal data collected by each sensor includes signal strength and timestamp. The data collected by the sensor includes both normal operating signals and potential fault signals. This process provides comprehensive and dynamic data support for fault detection. The data collected by the sensors includes not only normal operating signals but also potential fault signals, ensuring comprehensive and real-time fault detection.

[0020] S2. Filter and normalize the electrical signal data to generate a time delay matrix and a physical distance matrix; construct a cost function for the propagation path based on the time delay matrix and the physical distance matrix; and calculate the cost between each node based on the cost function to generate a cost matrix, specifically including: The electrical signal data collected by the sensor usually contains noise, which affects subsequent analysis and modeling. Therefore, we need to filter the signal to reduce the impact of noise. We choose to use a low-pass filter to remove high-frequency noise. The low-pass filter retains the low-frequency signal while filtering out the high-frequency noise. The low-pass filter algorithm smoothes the high-frequency components in the signal strength data of each node. In order to make the signal strength of different nodes uniformly comparable, we need to normalize the signal strength and convert its value to between 0 and 1. Calculate the maximum and minimum signal strengths of all nodes and use the normalization formula to convert the signal strength value of each node to a value between 0 and 1. This is done by subtracting the minimum value from each signal strength value and then dividing it by the difference between the maximum and minimum values. Based on the timestamp information in the electrical signal data collected by the sensor, the signal propagation time delay between each node is calculated. For example, if the timestamp of the electrical signal data at node A is 100 milliseconds and the timestamp at node B is 105 milliseconds, then the time delay for the electrical signal data to propagate from node A to node B is 5 milliseconds. The signal propagation time delay between all nodes is calculated to generate a time delay matrix. The time delay matrix is ​​an N×N matrix, where N is the number of nodes. Each element in the matrix represents the signal propagation time delay between two corresponding nodes. Obtain a physical layout diagram of the circuit from a public database. Based on the physical layout diagram, obtain the physical distance between each node. For example, the physical distance between node A and node B is 1 meter, the physical distance between node A and node C is 2 meters, and so on. Record the physical distances between all nodes and generate a physical distance matrix. The physical distance matrix is ​​also an N×N matrix, where each element represents the physical distance between two corresponding nodes. To evaluate the total cost of electrical signal data propagating from one node to another, a cost function for the propagation path is defined. This function combines information from the time delay matrix and the physical distance matrix. The cost function is a weighted sum of time delay and physical distance. Specifically, the time delay and physical distance are multiplied by their respective weight parameters, and then the sum is added to obtain the total cost. For example, if the weight parameter of time delay is 0.7 and the weight parameter of physical distance is 0.3, then the total cost is calculated as 0.7 times the time delay plus 0.3 times the physical distance. The weight parameters are determined through cross-validation. Using the determined weight parameters, the total cost between each node is calculated based on the propagation path cost function. For example, if the time delay from node A to node B is 5 milliseconds, the physical distance is 1 meter, and the weight parameters are 0.7 and 0.3, respectively, then the total cost between node A and node B is 0.7 × 5 + 0.3 × 1 = 3.5 + 0.3 = 3.8. 0.7 × 5 + 0.3 × 1 = 3.5 + 0.3 = 3.8. The total cost between all nodes is calculated to generate an N × N cost matrix. Each element in the cost matrix represents the total cost between the corresponding two nodes.

[0021] S3 uses a path optimization algorithm to find the shortest path from each node to other nodes based on the cost matrix and generate a propagation path graph, which includes: Initialize the distance matrix, set the distance of all nodes to infinity, and the distance of the starting node to 0; select the node with the smallest distance as the current node; Update the distance from the current node to all adjacent nodes. If the distance to the adjacent node through the current node is shorter, update the distance to the adjacent node. Repeat the above steps until the distances of all nodes are calculated; Based on the calculation results of the shortest path, a propagation path graph is generated. The propagation path graph shows the shortest path from each node to other nodes. For example, the shortest path from node A to node B is A->B, the shortest path from node A to node C is A->E->C, and so on.

[0022] Perform fault analysis on each node in the propagation path diagram to determine all fault source nodes, including: Perform preliminary abnormal screening on the signal propagation time delay corresponding to each node in the propagation path diagram. By calculating the average signal propagation time delay of all nodes, if the signal propagation time delay corresponding to each node exceeds 10% of the average signal propagation time delay, it is determined to be a preliminary abnormal node; For each preliminary abnormal node, analyze its impact range and check the shortest paths from that node to all other nodes to determine whether the abnormal signal is reflected on these paths. Starting from the abnormal signal node, check the shortest paths from it to all other nodes. For example, if the signal of node A is abnormal, check whether the time delay on the path from node A to all other nodes (B, C, D, E) has increased. Record all nodes and paths affected by the abnormal signal of node A. Record the number of other nodes and paths affected by each preliminary abnormal node. For example, node A affects the paths of nodes B, C, D, and E, and node D affects the paths from nodes A to D and A to E. Exclude nodes that only affect one path and treat the remaining preliminary abnormal nodes as fault source nodes.

[0023] Using Bayesian networks for probabilistic reasoning, we calculate the posterior probability of each fault source node under a given propagation path and signal propagation time delay, and select the node with the highest posterior probability as the final fault source node. This includes: Construct a Bayesian network, where nodes represent physical nodes in the circuit (e.g., nodes A, B, C, D, and E), edges represent physical connections, and a conditional probability table represents the probability of abnormal signals occurring in other nodes when a node fails. A priori probability is set for each fault source node. The prior probability is set based on historical fault data or the importance and reliability of the node, and the likelihood function on each path is calculated, that is, the probability of the observed signal propagation path and time data given the fault source node; specifically, for each fault source node, its shortest path to all other nodes is checked, the probability of signal anomalies on the path is evaluated, and these probabilities are multiplied to obtain the likelihood function value of the fault source node; in order to enable a uniform comparison of the signal strength of different nodes, the signal strength is normalized and its value is converted to between 0 and 1. The maximum and minimum values ​​of the signal strength of all nodes are calculated, and the signal strength value of each node is converted to a value between 0 and 1 using the normalization formula. This is done by subtracting the minimum value from each signal strength value and then dividing it by the difference between the maximum and minimum values; Using the Bayesian formula, based on the prior probability of each fault source node, the posterior probability values ​​of all fault source nodes are compared, and the node with the highest posterior probability is selected as the final fault source node. Through this process, combined with path analysis and probability model, the most likely fault source node can be effectively identified, thereby improving the accuracy and reliability of fault location.

[0024] The above shows and describes the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The above embodiments and descriptions are merely preferred examples of the present invention and are not intended to limit the present invention. Various changes and improvements may be made to the present invention without departing from the spirit and scope of the present invention. Such changes and improvements fall within the scope of the present invention. The scope of protection claimed in the present invention is defined by the appended claims and their equivalents.

Claims

1. A highly safe method for detecting circuit faults in electrical equipment, characterized in that: The method steps are as follows: S1. Collecting electrical signal data of multiple nodes in a circuit through sensors, wherein the electrical signal data includes signal strength and timestamp; S2, filtering and normalizing the electrical signal data to generate a time delay matrix and a physical distance matrix; and constructing a cost function of the propagation path based on the time delay matrix and the physical distance matrix; and calculating the cost between each node based on the cost function to generate a cost matrix; S3. Use the path optimization algorithm to find the shortest path from each node to other nodes according to the cost matrix and generate a propagation path graph; Perform fault analysis on each node in the propagation path diagram to determine all fault source nodes; use Bayesian network for probabilistic reasoning to calculate the posterior probability of each fault source node under a given propagation path and signal propagation time delay, and select the node with the highest posterior probability as the final fault source node.

2. The high-safety electrical equipment circuit fault detection method according to claim 1, characterized in that: The electrical signal data includes signals during normal operation and potential fault signals.

3. The high-safety electrical equipment circuit fault detection method according to claim 2, characterized in that: The process of generating the time delay matrix and the physical distance matrix specifically includes: According to the timestamp information in the electrical signal data collected by the sensor, the signal propagation time delay between each node is calculated; the signal propagation time delay between all nodes is calculated to generate a time delay matrix, which is an N×N matrix, where N is the number of nodes, and each element in the matrix represents the signal propagation time delay between two corresponding nodes; A physical layout diagram of the circuit is obtained from a public database, and the physical distance between each node is obtained based on the physical layout diagram of the circuit; the physical distances between all nodes are recorded to generate a physical distance matrix, which is also an N×N matrix, in which each element represents the physical distance between two corresponding nodes.

4. The high-safety electrical equipment circuit fault detection method according to claim 3, characterized in that: The process of constructing the cost function of the propagation path and generating the cost matrix specifically includes: The cost function is used to evaluate the total cost of electrical signal data propagating from one node to another. The function combines the information of the time delay matrix and the physical distance matrix. The form of the cost function is to multiply the signal propagation time delay and the physical distance by their respective weight parameters, and then add them together to obtain the total cost, where the weight parameters are determined by the cross-validation method. Using the determined weight parameters, the total cost between each node is calculated according to the cost function of the propagation path; the total cost between all nodes is calculated to generate an N×N cost matrix, in which each element of the cost matrix represents the total cost between the corresponding two nodes.

5. The high-safety electrical equipment circuit fault detection method according to claim 1, characterized in that: The process of generating the propagation path graph specifically includes: Initialize the distance matrix, set the distances of all nodes to infinity, and the distance of the starting node to 0; select the node with the smallest distance as the current node; Update the distance from the current node to all adjacent nodes. If the distance to the adjacent node through the current node is shorter, update the distance of the adjacent node. Repeat the above steps until the distances of all nodes have been calculated; Based on the calculation results of the shortest path, a propagation path graph is generated, which shows the shortest path from each node to other nodes.

6. The high-safety electrical equipment circuit fault detection method according to claim 5, characterized in that: The process of determining all fault source nodes specifically includes: Perform preliminary abnormal screening on the signal propagation time delay corresponding to each node in the propagation path diagram, and calculate the average signal propagation time delay of all nodes. If the signal propagation time delay corresponding to each node exceeds 10% of the average signal propagation time delay, it is determined as a preliminary abnormal node; For each preliminary abnormal node, analyze its impact range by checking the shortest path from the preliminary abnormal node to all other nodes; record the number of other nodes and the number of paths affected by each preliminary abnormal node; exclude a node that only affects one path, and use the remaining preliminary abnormal nodes as fault source nodes.

7. The high-safety electrical equipment circuit fault detection method according to claim 6, characterized in that: The process of generating the final fault source node specifically includes: Construct a Bayesian network, where the nodes in the network represent the physical nodes in the circuit, the edges represent the physical connection relationships, and the conditional probability table represents the probability of abnormal signals appearing in other nodes when a node fails; A priori probability is set for each fault source node. The prior probability is set based on historical fault data, and the likelihood function on each path is calculated. For each fault source node, the shortest path from it to all other nodes is checked, the signal abnormality probability on the path is evaluated, and the probability is multiplied to obtain the likelihood function value of the fault source node; the signal strength is normalized and its value is converted to between 0 and 1. The maximum and minimum values ​​of the signal strength of all nodes are calculated, and the signal strength value of each node is converted to a value between 0 and 1 using the normalization formula, by subtracting the minimum value from each signal strength value and dividing it by the difference between the maximum and minimum values; Using the Bayesian formula, the posterior probability of each fault source node is calculated according to the prior probability of each fault source node, the posterior probability values ​​of all fault source nodes are compared, and the node with the highest posterior probability is selected as the final fault source node.

Images