Spring operating mechanism fault monitoring method and device, computer equipment and medium

Abstract

The invention relates to a fault monitoring method and device for a spring operating mechanism, computer equipment and a medium, and the method comprises the steps: firstly carrying out the detection processing of data collected by a plurality of different types of sensors of the spring operating mechanism, and obtaining a detection result; and then, under the condition that the representation of the detection result is abnormal, performing fault detection processing on the detection result based on an isolated forest algorithm constructed by a plurality of randomly generated isolated trees to obtain a fault detection result of the spring operating mechanism. According to the method and the device, the data acquired by the plurality of different types of sensors are comprehensively detected and processed, so that the problem that errors are easy to occur in single-sensor data detection is avoided, and the fault detection accuracy is effectively improved. According to the method, fault detection processing is carried out on possibly abnormal data by adopting an isolated forest algorithm, effective detection can be realized under a label-free condition, the adaptability of the whole monitoring system to different running states and data distribution is greatly enhanced, and the method can be widely applied to fault monitoring scenes of various spring operating mechanisms.

Description

Technical Field

[0001] The present application relates to the technical field of high-voltage circuit breaker monitoring, and in particular to a method, device, computer equipment, and medium for monitoring faults of a spring operating mechanism. Background Art

[0002] In power systems, the primary function of GIS (Gas Insulated Switchgear) high-voltage circuit breakers is to connect or disconnect load current. However, the spring mechanism of GIS circuit breakers can fail to operate after long-term operation. Furthermore, when the spring mechanism is in an energy-storage state, the metal material is subjected to static pressure for a long time. Metal fatigue can cause insufficient static pressure in the spring mechanism, frequently leading to failed opening and closing operations or reduced mechanism speed. In severe cases, this can cause over-tripping, posing a significant threat to the safe and stable operation of the power grid and becoming a key factor limiting GIS reliability.

[0003] Current GIS high-voltage circuit breaker inspection technologies primarily involve periodic inspections and online monitoring. Periodic inspections have significant drawbacks, including a high degree of blindness and difficulty accurately identifying potential defects within each device. Current online monitoring, on the other hand, mostly relies on a single physical parameter, resulting in limited accuracy and low reliability. Summary of the Invention

[0004] Based on this, it is necessary to provide a spring operating mechanism fault monitoring method, device, computer equipment and medium that can solve the above technical problems.

[0005] In a first aspect, the present application provides a spring operating mechanism fault monitoring method, the method comprising:

[0006] Detecting and processing data collected by multiple different types of sensors of the spring operating mechanism to obtain detection results;

[0007] When the test results indicate anomalies, fault detection processing is performed on the test results based on a pre-built isolation forest algorithm to obtain fault detection results for the spring operating mechanism. The isolation forest algorithm is constructed based on multiple randomly generated isolated trees. The fault detection results are used to evaluate the abnormal conditions of each component in the spring operating mechanism and the severity of the fault.

[0008] In one embodiment, the above-mentioned pre-built isolation forest algorithm is used to perform fault detection processing on the detection result to obtain the fault detection result, including:

[0009] Perform feature extraction on the data collected by various types of sensors to obtain multiple sensor feature vectors;

[0010] The anomaly index of each sensor’s feature vector is calculated using the isolation forest algorithm;

[0011] The fault detection result is determined based on the abnormality index and a pre-set index interval, wherein different index intervals correspond to different severity levels.

[0012] In one embodiment, the calculation of the abnormality index of each sensor feature vector using the isolation forest algorithm includes:

[0013] For each sensor feature vector, the isolation forest algorithm is used to perform isolation tree traversal on the sensor feature vector to obtain the average height of the sensor feature vector in each isolation tree;

[0014] The average height of each sensor's feature vector is normalized to obtain the abnormal index of each sensor's feature vector.

[0015] In one embodiment, the construction process of the isolation forest algorithm includes:

[0016] A plurality of sample feature vectors obtained in advance are randomly selected to obtain a plurality of target sample points and target cutting points; each target sample point includes a plurality of vector dimensions;

[0017] Generate a hyperplane based on the target cutting point, and divide the data space according to the hyperplane to obtain two subspaces; one subspace corresponds to a child node of the isolated tree;

[0018] For each target sample point, the data of different vector dimensions in the target sample point are divided into different subspaces according to the target cutting point to obtain the corresponding child nodes;

[0019] The partitioning process is performed on each child node until the child node meets the termination condition after the partitioning, and an isolated tree is obtained. The isolation forest algorithm is constructed by multiple isolated trees.

[0020] In one embodiment, the random selection process is performed on the pre-acquired multiple sample feature vectors to obtain multiple target sample points and target cutting points, including:

[0021] According to the pre-set sample length, each sample feature vector is randomly selected to obtain a sample subset; the sample subset includes multiple target sample points; the sample subset is used to create the root node of the isolation tree;

[0022] Randomly assign the vector dimension of the sample feature vector to obtain the target dimension;

[0023] The target cutting points are randomly generated from the target sample points of the sample subset according to the target dimension.

[0024] In one embodiment, the above method of dividing the data of different vector dimensions in the target sample point into different subspaces according to the target cutting point to obtain corresponding child nodes includes:

[0025] According to the target dimension of the target cutting point, the data smaller than the target dimension in the target sample point is divided into the first subspace to obtain the left child node;

[0026] Divide the data in the target sample point that is greater than or equal to the target dimension into the second subspace to obtain the right child node.

[0027] In one embodiment, determining the fault detection result based on the abnormality index and the pre-set index range includes:

[0028] Determine the target index range corresponding to the abnormal index;

[0029] Determine the fault detection result based on the alarm level corresponding to the target index range;

[0030] Among them, the alarm level corresponding to the target index interval includes at least one of the first level alarm, the second level alarm and the third level alarm, and the fault detection result includes at least one of the slight deviation, the moderate deviation and the severe deviation of the sensor collected data.

[0031] In a second aspect, the present application also provides a spring operating mechanism fault monitoring device. The device comprises:

[0032] A data detection module is used to detect and process data collected by multiple different types of sensors of the spring operating mechanism to obtain detection results;

[0033] The fault detection module is used to perform fault detection processing on the detection results based on a pre-built isolation forest algorithm when the detection results indicate an abnormality, thereby obtaining a fault detection result for the spring operating mechanism; the isolation forest algorithm is constructed based on multiple randomly generated isolated trees; the fault detection results are used to evaluate the abnormal conditions of each component in the spring operating mechanism and the severity of the fault.

[0034] In a third aspect, the present application further provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of the method in the first aspect when executing the computer program.

[0035] In a fourth aspect, the present application further provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the method in the first aspect.

[0036] In a fifth aspect, the present application further provides a computer program product, which includes a computer program that implements the steps of the method in the first aspect when executed by a processor.

[0037] The above-mentioned spring operating mechanism fault monitoring method, device, computer equipment and medium first detect and process the data collected by multiple different types of sensors of the spring operating mechanism to obtain a detection result; then, when the detection result represents an abnormality, the detection result is processed by an isolation forest algorithm constructed based on multiple randomly generated isolated trees to obtain a fault detection result of the spring operating mechanism. This application avoids the problem of error proneness in single sensor data detection by performing comprehensive detection and processing on data collected by multiple different types of sensors, and effectively improves the accuracy of fault detection. The use of the isolation forest algorithm to perform fault detection and processing on potentially abnormal data can achieve effective detection under label-free conditions, greatly enhancing the adaptability of the entire monitoring system to different operating states and data distributions, and can be widely used in fault monitoring scenarios of various spring operating mechanisms. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 2. A diagram showing an application environment of a spring operating mechanism fault monitoring method according to an embodiment;

[0039] Figure 2 Schematic diagram of a flow chart of a spring operating mechanism fault monitoring method according to one embodiment;

[0040] Figure 3 A schematic diagram of a process for performing fault detection in one embodiment;

[0041] Figure 4 Schematic diagram of a process for calculating anomaly index of each sensor feature vector in one embodiment;

[0042] Figure 5 1 is a flow chart of a construction process of an isolation forest algorithm in one embodiment;

[0043] Figure 6 A schematic diagram of a process for obtaining multiple target sample points and target cutting points in one embodiment;

[0044] Figure 7 A schematic diagram of a process for obtaining a child node in one embodiment;

[0045] Figure 8 A schematic diagram of a process for determining a fault detection result in one embodiment;

[0046] Figure 9 is a structural block diagram of a spring operating mechanism fault monitoring device in one embodiment;

[0047] Figure 10 FIG. 1 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0048] In order to make the purpose, technical solutions and advantages of this application more clear, the following further describes this application in detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain this application and are not intended to limit this application.

[0049] The spring operating mechanism fault monitoring method provided in the embodiment of the present application can be applied to Figure 1 In the application environment shown. Among them, the terminal 102 communicates with the server 104 through the network. The data storage system can store the data that the server 104 needs to process. The data storage system can be integrated on the server 104, or it can be placed on the cloud or other network servers. The server 104 detects and processes the data collected by multiple different types of sensors of the spring operating mechanism to obtain a detection result; when the detection result indicates an abnormality, the detection result is subjected to fault detection processing based on a pre-built isolation forest algorithm to obtain a fault detection result of the spring operating mechanism. Among them, the terminal 102 can be, but is not limited to, an intelligent control cabinet and an on-site monitoring cabinet. The server 104 can be implemented as an independent server or a server cluster consisting of multiple servers.

[0050] In an exemplary embodiment, Figure 2 As shown, a spring operating mechanism fault monitoring method is provided, which is applied to Figure 1 Taking the server 104 in the example as an example, the following steps are included:

[0051] S201 , performing detection processing on data collected by multiple different types of sensors of the spring operating mechanism to obtain detection results.

[0052] The spring operating mechanism is a crucial device used in power systems to operate circuit breakers and other equipment. Different types of sensors include closing spring pressure sensors, operating mechanism arm linear velocity sensors, closing coil current sensors, opening coil current sensors, and energy storage motor current sensors. These sensors are used to collect data on various aspects of the spring operating mechanism during operation, providing a basis for subsequent monitoring and fault diagnosis.

[0053] In this embodiment, the server first establishes communication with various sensors installed on the spring operating mechanism and receives data collected by them. For example, it obtains real-time spring pressure data from the closing spring pressure sensor, receives crank arm speed data from the operating mechanism crank arm linear speed sensor, and receives corresponding current data from the closing coil current sensor, opening coil current sensor, and energy storage motor current sensor.

[0054] The server pre-stores standard reference value ranges for each sensor data point. For each received data point, the server compares it against the corresponding standard range. For example, if the received pressure value falls within the preset normal pressure range, the data is marked as normal; if it falls outside this range, it is marked as a possible abnormality. This processing is repeated for all sensor data to obtain the overall detection result.

[0055] The spring operating mechanism is equipped with a closing spring pressure sensor, an operating mechanism crank arm linear speed sensor, a closing coil current sensor, an opening coil current sensor and an energy storage motor current sensor.

[0056] For time t Data from sensors , define the corresponding physical parameter standard reference value and the allowable fluctuation range [S i min , S i max ] .

[0057] If any sensor data All satisfied , then mark the group of data as normal; if a sensor data Out of range [S i min , S i max ] , then the data set is judged as possibly abnormal and the next step of detection is carried out.

[0058] S202 , when the detection result indicates an abnormality, performing fault detection processing on the detection result based on a pre-built isolation forest algorithm to obtain a fault detection result of the spring operating mechanism.

[0059] Among them, the isolation forest algorithm is constructed based on multiple randomly generated isolation trees; the fault detection results are used to evaluate the abnormal conditions of each component in the spring operating mechanism and the severity of the fault.

[0060] In this embodiment, when data marked as potentially abnormal is detected in the test results, the isolation forest algorithm is activated for further testing. The server first extracts a portion of data related to the current abnormal data type from its stored historical data, combining this data with the currently abnormal data to form a training dataset. For example, if the closing spring pressure data is abnormal, the server extracts previous closing spring pressure data and other related auxiliary data (such as other sensor data from the same time period).

[0061] This data is used to train an isolation forest, generating multiple isolated trees. During the construction of each isolation tree, the server randomly selects a characteristic dimension of the data (such as pressure value or current value trend) and randomly determines a cut point on that dimension, dividing the data into two subsets. This process is repeated until each leaf node contains only one data point or the tree reaches a preset height.

[0062] For data points marked as potentially anomaly, the server inserts them into the constructed isolation forest and traverses each isolated tree. Based on information such as the depth of the data point in the isolated tree, the average height of the data point in the forest is calculated. This average height is then converted into an anomaly score using a specific algorithm. For example, if a data point is shallow in most isolated trees, it indicates that it is significantly different from other normal data points, resulting in a higher anomaly score.

[0063] Based on the calculated anomaly score, the server determines the spring operating mechanism's fault condition. If the anomaly score exceeds a certain threshold, a fault is determined. The server then assesses the severity of the fault based on the anomaly score range, classifying it as a minor, moderate, or severe fault, ultimately generating the final fault detection result.

[0064] Another implementation method involves distributing preprocessed data across the system's storage nodes. When fault detection is needed, the master node distributes the data to different compute nodes for parallel processing based on the data type and characteristics. Each compute node constructs an isolation tree on its local data subset using the isolation forest algorithm and calculates anomaly scores for each local data point. Each compute node aggregates the results to the master node, which then performs comprehensive analysis and processing.

[0065] The above-mentioned spring operating mechanism fault monitoring method first detects and processes the data collected by multiple different types of sensors of the spring operating mechanism to obtain a detection result; then, when the detection result represents an abnormality, the detection result is processed by an isolation forest algorithm constructed based on multiple randomly generated isolated trees to obtain a fault detection result of the spring operating mechanism. This application avoids the problem of error proneness in single sensor data detection by performing comprehensive detection and processing on data collected by multiple different types of sensors, and effectively improves the accuracy of fault detection. The use of the isolation forest algorithm to perform fault detection and processing on potentially abnormal data can achieve effective detection under label-free conditions, greatly enhancing the adaptability of the entire monitoring system to different operating states and data distributions, and can be widely used in fault monitoring scenarios of various spring operating mechanisms.

[0066] In an exemplary embodiment, based on the above embodiment, see Figure 3The embodiment of the present application relates to a process of performing fault detection processing on a detection result based on a pre-built isolation forest algorithm to obtain a fault detection result, including the following steps:

[0067] S301 , performing feature extraction processing on data collected by various types of sensors to obtain multiple sensor feature vectors.

[0068] The sensor feature vector is a vector formed by combining the features of multiple sensors. Each dimension represents a specific feature of a sensor, integrating key information from different sensors.

[0069] In this embodiment of the present application, the server receives data from various sensors of the spring operating mechanism, including data such as closing spring pressure, opening and closing speed, closing coil current, opening coil current, and energy storage motor current. For the closing spring pressure data, the server calculates its average, maximum, minimum, and slope of pressure change within a time window (e.g., the past 10 minutes) as features. For the opening and closing speed data, features such as average speed and speed fluctuation range are extracted. For the current data, features such as current peak value, steady-state current value, and current rise and fall time are analyzed.

[0070] These features extracted from different sensors are combined in a specific order to form a sensor feature vector. For example, a feature vector might be a sample feature vector [average closing spring pressure, average opening and closing speed, peak closing coil current, steady-state opening coil current, and energy storage motor current rise time]. Each feature vector represents the comprehensive operating status information of the spring operating mechanism at a specific moment.

[0071] S302: Calculate the abnormality index of each sensor feature vector using the isolation forest algorithm.

[0072] The anomaly index is a value calculated by the isolation forest algorithm based on information such as the data's position and depth within the isolation tree. It measures the degree to which the data deviates from a normal distribution. A higher anomaly index indicates a greater likelihood of anomaly, and a greater likelihood of a faulty sensor feature or device component.

[0073] In this embodiment, the server randomly extracts a portion of the historical sensor feature vector data it stores and uses it, along with the currently extracted feature vector, as a training set to construct an isolation forest. When constructing the isolation forest, the server randomly selects a dimension of the feature vector and randomly determines a cut point on that dimension, splitting the data into two subsets. This process is repeated until each leaf node contains only one data point or the preset tree height is reached, thereby generating multiple isolation trees.

[0074] For each current sensor feature vector, the server traverses each isolated tree, calculates its average depth in the forest and other related information, and converts this information into an anomaly index based on internal rules. For example, if a feature vector has a shallow depth in most isolated trees, it indicates that it is significantly different from other normal data points, and its anomaly index will be relatively high.

[0075] The specific calculation steps of the algorithm are as follows:

[0076] For all sensor feature vectors collected in real time =[ ], , d represents the dimension of sensor feature vector; randomly extracted from X A subset of sample points long enough X Insert the root node.

[0077] Randomly specify a dimension q from the d dimensions and randomly generate a cutting point p in the current data. .

[0078] A hyperplane is generated at this cutting point p, dividing the current data space into two subspaces: sample points with a specified dimension less than p are placed in the left child node, and those greater than or equal to p are placed in the right child node.

[0079] Repeat the above steps until all leaf nodes have only one sample point or the isolated tree (iTree) has reached the specified height. Repeat the above steps until t isolated trees are generated.

[0080] For each data point , let it traverse each isolated tree (iTree) and calculate the point The average height in the forest h( ), normalize the average height of all points. The calculation formula of the outlier score is shown in formula (1)(2)(3):

[0081] (1)

[0082] in,

[0083] (2)

[0084] (3)

[0085] S303: Determine a fault detection result according to the abnormality index and a preset index range.

[0086] Among them, different index intervals correspond to different degrees of severity.

[0087] In this embodiment, the server pre-sets multiple index intervals, such as [0 sample feature vector - sample feature vector 0.3] for a normal interval, [0.3 sample feature vector - sample feature vector 0.6] for a slightly abnormal interval, [0.6 sample feature vector - sample feature vector 0.8] for a moderately abnormal interval, and [0.8 sample feature vector - sample feature vector 1] for a severely abnormal interval. After calculating the abnormality index of a sensor feature vector, the server determines the interval to which the index belongs. If the abnormality index is within the normal range, the server determines that the spring operating mechanism is currently operating normally; if it is in the slight abnormality range, the server records the abnormality and may send a prompt message to the operation and maintenance personnel, suggesting attention but not taking emergency measures for the time being; if it is in the moderate abnormality range, the server issues an alarm and notifies the operation and maintenance personnel to conduct further inspection and possible maintenance; when the abnormality index is in the serious abnormality range, the server immediately takes emergency measures, such as cutting off the power supply of related equipment or starting backup equipment, and notifies the operation and maintenance personnel to carry out emergency repairs. At the same time, the severity of the fault is determined according to the range of the abnormality index, providing a reference for subsequent maintenance work.

[0088] For example, It is recorded in x The training data of samples constitutes the anomaly index of iTree, which ranges from [0,1] and is used to measure the degree to which the data deviates from the normal distribution. Here, the judgment threshold is set to 0.6. When , the real-time input data is marked as normal; when When the data is abnormal, the system triggers a fault alarm.

[0089] The present embodiment accurately detects spring operating mechanism faults by extracting features from sensor data and calculating an anomaly index using the isolation forest algorithm. The feature extraction process converts raw data into more valuable information, reducing data processing complexity and improving analysis efficiency. The application of the isolation forest algorithm effectively identifies abnormal data even in complex operating environments, enhancing the accuracy and reliability of fault detection.

[0090] In an exemplary embodiment, based on the above embodiment, see Figure 4 The embodiment of the present application relates to a process of calculating the anomaly index of each sensor feature vector using the isolation forest algorithm, including the following steps:

[0091] S401 , for each sensor feature vector, performing isolation tree traversal processing on the sensor feature vector using the isolation forest algorithm to obtain an average height of the sensor feature vector in each isolation tree.

[0092] In the embodiment of the present application, the generated tree includes multiple isolated trees. At the same time, the sensor feature vector data currently required to be detected is received.

[0093] For each sensor feature vector, the server sequentially places it into each isolated tree in the isolation forest for traversal. For example, in an isolation tree, starting from the root node, the feature vector's value in a certain feature dimension of the current node (such as the closing spring pressure feature dimension) is compared with the node's pre-set split value. If the value is less than the split value, the node moves to the left child node; otherwise, it moves to the right child node. This process repeats until it reaches a leaf node, and the number of node levels traversed is recorded as the height in the isolation tree. After performing this operation on all isolated trees, the average height of the sensor feature vector in each isolation tree is calculated.

[0094] S402 : Normalize the average height of each sensor feature vector to obtain an abnormality index of each sensor feature vector.

[0095] In this embodiment of the present application, the server obtains the average height data of all sensor feature vectors after traversing the isolation tree and determines the maximum and minimum values in these data. Assume that after statistics, the maximum average height of all feature vectors is sample feature vector 8, and the minimum average height is sample feature vector 3.

[0096] The server normalizes the average height of each sensor feature vector as follows (this is just a simple linear normalization example, not involving complex formulas): If the average height of a feature vector is 5.5, then its normalized anomaly index is (5.5 - 3) / (8 - 3) = 0.5). This process converts the average height of all sensor feature vectors into an anomaly index between 0 and 1, facilitating subsequent fault diagnosis based on a set threshold. For example, if an anomaly index greater than 0.6 indicates a possible fault, then a feature vector with a normalized anomaly index of 0.5 is temporarily considered normal.

[0097] This embodiment of the present application uses an isolation forest algorithm to perform isolation tree traversal and normalization on sensor feature vectors to calculate an anomaly index, effectively identifying anomalies in spring operating mechanism operating data. This method, which does not rely on a large amount of prior fault data labels, has strong adaptability and generalization capabilities, and can accurately detect potential faults in complex and changing power equipment operating environments.

[0098] In an exemplary embodiment, based on the above embodiment, see Figure 5 The construction process of the isolation forest algorithm in the embodiment of the present application includes the following steps:

[0099] S501 , performing random selection processing on a plurality of sample feature vectors obtained in advance to obtain a plurality of target sample points and target cutting points.

[0100] Each target sample point consists of multiple vector dimensions. The sample feature vector is extracted from the historical sensor data of the spring operating mechanism and contains vector information about multiple sensor characteristics, such as closing spring pressure, opening and closing speed, and closing coil current. Target sample points are randomly selected sample data points used to construct an isolation tree. Each target sample point contains information from multiple vector dimensions, corresponding to the characteristics of different sensors. These dimensions form the basis for the algorithm's data partitioning and tree structure construction.

[0101] The target cut point is a value selected on a vector dimension that is used to divide the data space into two parts. For example, for the closing spring pressure dimension, if the target cut point is the sample feature vector 10MPa, then sample points with pressure values less than the sample feature vector 10MPa will be divided into one subspace, and sample points with pressure values greater than or equal to the sample feature vector 10MPa will be divided into the other subspace.

[0102] In this embodiment, the server extracts a set of sample feature vectors from a large amount of stored historical data about spring operating mechanism sensors. Assume this set contains 1000 sample feature vectors, each containing information in 5 dimensions, corresponding to the closing spring pressure, closing and opening speed, closing coil current, opening coil current, and energy storage motor current.

[0103] The server selects 100 sample feature vectors as target sample points according to a random rule. For each target sample point, the server randomly selects a dimension from the 5 sample feature vector dimensions, such as the closing spring pressure dimension, and then randomly determines a target cut point within the value range of this dimension. Assuming that the closing spring pressure dimension ranges from 5 MPa to 20 MPa, the server randomly selects 12 MPa as the target cut point.

[0104] S502: Generate a hyperplane according to the target cutting point, and divide the data space according to the hyperplane to obtain two subspaces.

[0105] Among them, a subspace corresponds to a child node of the isolated tree;

[0106] In this embodiment, the server generates a hyperplane in the data space based on the selected target cutpoint sample feature vector (12 MPa). This hyperplane divides the data space into two subspaces based on the closing spring pressure dimension: one subspace contains all target sample points whose closing spring pressure is less than the sample feature vector (12 MPa), and the other subspace contains all target sample points whose closing spring pressure is greater than or equal to the sample feature vector (12 MPa). These two subspaces correspond to two child nodes of the isolation tree, respectively.

[0107] S503 : For each target sample point, divide the data of different vector dimensions in the target sample point into different subspaces according to the target cutting point to obtain corresponding child nodes.

[0108] In this embodiment of the present application, the server traverses all target sample points and, for each target sample point, divides it into a corresponding subspace based on the comparison result of its closing spring pressure value with the sample feature vector of the target cut point, 12MPa. For example, if the closing spring pressure value of a target sample point is a sample feature vector of 10MPa, then it will be divided into the subspace of the sample feature vector with a pressure less than the sample feature vector of 12MPa; if the pressure value of another target sample point is a sample feature vector of 15MPa, then it will be divided into the subspace of the sample feature vector with a pressure greater than or equal to the sample feature vector of 12MPa.

[0109] Although feature information of other dimensions is not directly used as the basis for division in this division, they will be divided into corresponding sub-nodes along with the sample points so that they can continue to play a role in the subsequent sub-node division process.

[0110] S504 , performing a partitioning process on each child node until the child node meets the termination condition after partitioning, obtaining an isolated tree, and constructing an isolation forest algorithm from the multiple isolated trees.

[0111] In an embodiment of the present application, the server repeats steps one to three for each of the two child nodes after division, i.e., randomly selecting target sample points, determining target cutting points, generating hyperplanes, and dividing sample points are performed again, and the subtree structure is continuously recursively constructed until the child node meets the termination condition. The termination condition can be set as the number of sample points in the child node is less than a certain threshold (such as 5 sample feature vectors) or the depth of the isolated tree reaches a preset value (such as 10 sample feature vector layers). After multiple isolated trees are constructed according to the above rules, these isolated trees form an isolation forest algorithm. The server can use this isolation forest algorithm to perform anomaly detection on the new sensor feature vector to determine whether the spring operating mechanism is faulty.

[0112] This embodiment of the application constructs an isolation forest algorithm through specific steps, effectively utilizing historical data from spring operating mechanisms to identify abnormal operating conditions. Operations such as random selection and hyperplane partitioning enable the algorithm to adapt to complex data distributions, improving its ability to detect different fault types and reducing false positives and missed negatives.

[0113] In an exemplary embodiment, based on the above embodiment, see Figure 6 The embodiment of the present application relates to a process of randomly selecting a plurality of sample feature vectors obtained in advance to obtain a plurality of target sample points and target cutting points, including the following steps:

[0114] S601 , randomly selecting each sample feature vector according to a preset sample length to obtain a sample subset.

[0115] The sample length is the number of samples selected from all sample feature vectors when constructing the isolation tree. By setting an appropriate sample length, we can control computing resource consumption and improve algorithm efficiency while ensuring algorithm accuracy. The sample subset includes multiple target sample points; the sample subset is used to create the root node of the isolation tree.

[0116] In the embodiment of the present application, the server stores a large number of sample feature vectors corresponding to the historical operation data of the spring operating mechanism. Assuming there are 5,000 sample feature vectors, each vector contains feature information of 5 dimensions. The server pre-sets the sample length to 500.

[0117] The server uses a random algorithm to randomly extract 500 sample feature vectors from these 5,000 to form a sample subset. This sample subset will serve as the data foundation for subsequently creating the root node of the isolated tree. For example, these 500 sample feature vectors may include data feature combinations of spring operating mechanisms under different normal and abnormal operating conditions.

[0118] S602: Perform random assignment processing on the vector dimension of the sample feature vector to obtain a target dimension.

[0119] In this embodiment of the present application, the server randomly assigns the five dimensions of the sample feature vectors contained in the sample subset obtained in the previous step. Assuming that the server randomly selects the number 3 between 1 and 5 using a random number generator, the third dimension is determined as the target dimension. If the third dimension corresponds to the closing coil current feature dimension, then subsequent operations will be carried out around this dimension.

[0120] S603: randomly generate target cutting points from the target sample points of the sample subset according to the target dimension.

[0121] In this embodiment of the present application, based on a selected target dimension (the closing coil current feature dimension), the server retrieves all data values for that dimension from the 500 target sample points in the sample subset. Assuming these data values fluctuate within a certain range, the server randomly generates a value within this range as the target cutoff point. For example, if the closing coil current values are between 0.5A and 2A, the server randomly generates 1.2A as the target cutoff point. This target cutoff point is used to partition the data space of the sample subset into two parts, corresponding to the two child nodes of the isolation tree, laying the foundation for the subsequent construction of the isolation tree structure.

[0122] The embodiment of the present application constructs an isolation forest algorithm by performing operations such as random selection of sample feature vectors, target dimension specification, and target cutting point generation. This can effectively utilize the diversity and randomness of data, avoid the algorithm's excessive reliance on specific data patterns, and improve the accuracy and reliability of spring operating mechanism fault detection.

[0123] In an exemplary embodiment, based on the above embodiment, see Figure 7 The embodiment of the present application relates to a process of dividing data of different vector dimensions in a target sample point into different subspaces according to a target cutting point to obtain corresponding child nodes, including the following steps:

[0124] S701 , according to the target dimension of the target cutting point, divide the data of the target sample point that is smaller than the target dimension into a first subspace to obtain a left child node.

[0125] The target dimension is a randomly selected dimension from the multiple dimensions of the sample feature vector and serves as the baseline dimension for data partitioning. For example, it could be one of the dimensions such as closing spring pressure, opening and closing speed, or closing coil current. Subsequent data partitioning operations will revolve around this dimension. The first subspace (left child node) is a portion of the data space partitioned based on the target cut point and target dimension. This subspace contains data from the target sample point that is smaller than the target cut point in the target dimension. This data will form the left child node and form part of the isolated tree branch structure. Further data partitioning and tree construction may be performed based on this subspace.

[0126] In this embodiment of the present application, when constructing the isolation tree, the server assumes that the currently determined target dimension is the opening and closing speed and the target cut point is 2 m / s. The server checks the value of each sample point in the opening and closing speed dimension one by one from the existing target sample point set.

[0127] For example, if the closing and opening speed of a target sample point is 1.5 m / s, because 1.5 m / s is less than 2 m / s, this sample point will be divided into the first subspace, that is, the left child node. The server continues to perform this judgment and division operation on all target sample points until all sample points in this dimension are processed, thereby determining the specific data composition of the left child node.

[0128] S702: Divide the data of the target sample point that is greater than or equal to the target dimension into the second subspace to obtain a right child node.

[0129] Among them, the second subspace (right child node) corresponds to the first subspace, and contains the data in the target sample points that are greater than or equal to the target cut point in the target dimension, forming the right child node. It is also an important part of the isolation tree construction process, which is used to further refine data classification and tree growth.

[0130] In the embodiment of the present application, for those target sample points whose opening and closing speed dimension is greater than or equal to 2m / s, the server divides them into the second subspace, i.e., the right child node. For example, the opening and closing speed of another target sample point is 2.5m / s, which meets the condition of being greater than or equal to 2m / s, so it will be placed in the right child node. The server also traverses and judges all target sample points, completes the data filling of the right child node, and thus completes the division of the target sample points according to the target cutting point and the target dimension, forming two child nodes of the isolation tree, laying the foundation for the subsequent further construction of the hierarchy of the isolation tree.

[0131] This embodiment of the application uses clear rules to partition target sample point data into different subspaces to form subnodes. This simple yet effective data partitioning method is a key step in constructing the isolation tree and, in turn, the isolation forest algorithm. It can quickly classify and organize complex data, enabling the algorithm to gradually discover potential abnormal patterns in the data.

[0132] In an exemplary embodiment, based on the above embodiment, see Figure 8 The embodiment of the present application relates to a process of determining a fault detection result based on an abnormality index and a pre-set index interval, including the following steps:

[0133] S801: Determine a target index range corresponding to an abnormality index.

[0134] The target index interval is the specific interval among all pre-defined index intervals that matches the calculated anomaly index. By determining the target index interval, we can clearly identify the category to which the data anomaly belongs and take appropriate measures.

[0135] In this embodiment of the present application, a server receives an anomaly index of 0.5 for a sensor feature vector calculated by the isolation forest algorithm. The server pre-sets index intervals: [0-0.3] for the normal range, [0.3-0.6] for the slightly deviated range, [0.6-0.8] for the moderately deviated range, and [0.8-1] for the severely deviated range. Because 0.5 falls within the [0.3-0.6] range, the server determines that the target index interval corresponding to this anomaly index is the slightly deviated range.

[0136] S802: Determine a fault detection result according to the alarm level corresponding to the target index range.

[0137] Among them, the alarm level corresponding to the target index interval includes at least one of the first level alarm, the second level alarm and the third level alarm, and the fault detection result includes at least one of the slight deviation, the moderate deviation and the severe deviation of the sensor collected data.

[0138] In this embodiment of the present application, the server, based on pre-defined rules, knows that the alarm level corresponding to the slight deviation interval is a Level 1 alarm. Therefore, the server determines that the fault detection result is a slight deviation in the sensor data, and triggers the Level 1 alarm mechanism. The server may record this abnormality and send a prompt message to the operation and maintenance personnel, informing them that a slight deviation in the sensor data is recommended, but no emergency measures are required at this time. Further inspection and analysis can be carried out at an appropriate time.

[0139] The present embodiment determines the fault detection result by matching the abnormality index with a pre-set index range and alarm level, providing maintenance personnel with a clear and intuitive fault judgment standard. This method can quickly and accurately assess the operating status of the spring operating mechanism, allowing maintenance personnel to take appropriate measures according to the severity of the fault.

[0140] In an exemplary embodiment, based on the above embodiment, the method of the embodiment of the present application further includes the following steps:

[0141] Step 1: Based on the pre-set sample length, randomly select each sample feature vector to obtain a sample subset; the sample subset includes multiple target sample points; randomly specify the vector dimension of the sample feature vector to obtain the target dimension; randomly generate the target cutting point from the target sample points in the sample subset according to the target dimension;

[0142] Step 2: Generate a hyperplane based on the target cut point and divide the data space according to the hyperplane to obtain two subspaces; for each target sample point, according to the target dimension of the target cut point, divide the data smaller than the target dimension in the target sample point into the first subspace to obtain the left child node; divide the data greater than or equal to the target dimension in the target sample point into the second subspace to obtain the right child node;

[0143] Step 3: Perform partitioning on each child node until the child node meets the termination condition after partitioning, and obtain an isolated tree. Then, an isolation forest algorithm is constructed from multiple isolated trees.

[0144] Step 4: Detect and process the data collected by multiple different types of sensors of the spring operating mechanism to obtain detection results; when the detection results indicate an abnormality, perform feature extraction on the data collected by each type of sensor to obtain multiple sensor feature vectors; for each sensor feature vector, use the isolation forest algorithm to perform isolation tree traversal on the sensor feature vector to obtain the average height of the sensor feature vector in each isolation tree; normalize the average height of each sensor feature vector to obtain the abnormality index of each sensor feature vector; determine the target index range corresponding to the abnormality index; and determine the fault detection result based on the alarm level corresponding to the target index range.

[0145] It should be understood that, although the various steps in the flowcharts involved in the various embodiments described above are displayed in sequence according to the instructions of the arrows, these steps are not necessarily executed in sequence in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order restriction on the execution of these steps, and these steps can be executed in other orders. Moreover, at least a portion of the steps in the flowcharts involved in the various embodiments described above can include multiple steps or multiple stages, and these steps or stages are not necessarily executed and completed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a portion of steps or stages in other steps.

[0146] Based on the same inventive concept, embodiments of the present application further provide a spring operating mechanism fault monitoring device for implementing the aforementioned spring operating mechanism fault monitoring method. The solution provided by this device is similar to the solution described in the aforementioned method. Therefore, the specific limitations of one or more of the following embodiments of the spring operating mechanism fault monitoring device can be found in the aforementioned definitions of the spring operating mechanism fault monitoring method and will not be further elaborated here.

[0147] In one embodiment, Figure 9 As shown, a spring operating mechanism fault monitoring device 900 is provided, comprising:

[0148] The data detection module 901 is used to detect and process the data collected by multiple different types of sensors of the spring operating mechanism to obtain detection results;

[0149] Fault detection module 902 is used to perform fault detection processing on the detection results based on a pre-built isolation forest algorithm when the detection results indicate an abnormality, thereby obtaining a fault detection result for the spring operating mechanism. The isolation forest algorithm is constructed based on multiple randomly generated isolated trees. The fault detection result is used to evaluate the abnormality of each component in the spring operating mechanism and the severity of the fault.

[0150] In one embodiment, the fault detection module includes:

[0151] A feature extraction unit is used to perform feature extraction processing on the data collected by various types of sensors to obtain multiple sensor feature vectors;

[0152] An index calculation unit, used to calculate the abnormal index of each sensor feature vector using an isolation forest algorithm;

[0153] The fault detection unit is used to determine the fault detection result according to the abnormality index and the pre-set index interval, wherein different index intervals correspond to different severity levels.

[0154] In one embodiment, the index calculation unit includes:

[0155] The feature traversal subunit is used to perform isolation tree traversal processing on each sensor feature vector using the isolation forest algorithm to obtain the average height of the sensor feature vector in each isolation tree;

[0156] The data normalization subunit is used to normalize the average height of the feature vectors of each sensor to obtain the abnormal index of the feature vector of each sensor.

[0157] In one embodiment, the apparatus comprises:

[0158] A vector selection unit is used to randomly select multiple sample feature vectors obtained in advance to obtain multiple target sample points and target cutting points; each target sample point includes multiple vector dimensions;

[0159] The space partitioning unit is used to generate a hyperplane according to the target cutting point and partition the data space according to the hyperplane to obtain two subspaces; one subspace corresponds to a child node of the isolated tree;

[0160] A data partitioning unit is used to partition the data of different vector dimensions in each target sample point into different subspaces according to the target cutting point to obtain corresponding child nodes;

[0161] The node partitioning unit is used to perform partitioning processing on each child node until the child node meets the end condition after partitioning, obtains an isolated tree, and constructs an isolation forest algorithm from multiple isolated trees.

[0162] In one embodiment, the vector selection unit includes:

[0163] The vector selection subunit is used to randomly select each sample feature vector according to a preset sample length to obtain a sample subset; the sample subset includes multiple target sample points; the sample subset is used to create the root node of the isolation tree;

[0164] The dimension assignment subunit is used to randomly assign the vector dimension of the sample feature vector to obtain the target dimension;

[0165] The cutting point generation subunit is used to randomly generate target cutting points from target sample points in the sample subset according to the target dimension.

[0166] In one embodiment, the data partitioning unit includes:

[0167] The first partitioning subunit is configured to partition the data of the target sample point that is smaller than the target dimension into a first subspace according to the target dimension of the target cutting point, thereby obtaining a left child node;

[0168] The second partitioning subunit is used to partition the data in the target sample point that is greater than or equal to the target dimension into the second subspace to obtain a right child node.

[0169] In one embodiment, the fault detection unit includes:

[0170] An interval determination subunit, used to determine a target index interval corresponding to the abnormality index;

[0171] The result determination subunit is used to determine the fault detection result according to the alarm level corresponding to the target index interval; wherein the alarm level corresponding to the target index interval includes at least one of a first-level alarm, a second-level alarm, and a third-level alarm, and the fault detection result includes at least one of a slight deviation, a moderate deviation, and a severe deviation in the sensor collected data.

[0172] Each module in the spring operating mechanism fault monitoring device can be implemented in whole or in part through software, hardware, or a combination thereof. Each module can be embedded in or independent of a processor in a computer device in the form of hardware, or can be stored in a memory in the computer device in the form of software, so that the processor can call and execute the corresponding operations of each module.

[0173] In one embodiment, a computer device is provided. The computer device may be a terminal, and its internal structure diagram may be as follows: Figure 10As shown. The computer device includes a processor, memory, an input / output interface, a communication interface, a display unit, and an input device. The processor, memory, and input / output interface are connected via a system bus, and the communication interface, display unit, and input device are connected to the system bus via the input / output interface. The processor of the computer device provides computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores an operating system and computer programs. The internal memory provides an environment for the operating system and computer programs stored in the non-volatile storage medium. The input / output interface of the computer device is used to exchange information between the processor and external devices. The communication interface of the computer device is used to communicate with external terminals via wired or wireless communication, which can be achieved via Wi-Fi, mobile cellular networks, NFC (near-field communication), or other technologies. When executed by the processor, the computer program implements a method for monitoring spring operating mechanism faults. The display unit of the computer device is used to produce a visually visible image and can be a display screen, a projection device, or a virtual reality imaging device. The display screen can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device can be a touch layer covering the display screen, or a button, trackball or touchpad set on the computer device casing, or an external keyboard, touchpad or mouse.

[0174] Those skilled in the art will understand that Figure 10 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than shown in the figure, or combine certain components, or have a different component arrangement.

[0175] In one embodiment, a computer device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and when the processor executes the computer program, the following steps are implemented:

[0176] Detecting and processing data collected by multiple different types of sensors of the spring operating mechanism to obtain detection results;

[0177] When the test results indicate anomalies, fault detection processing is performed on the test results based on a pre-built isolation forest algorithm to obtain fault detection results for the spring operating mechanism. The isolation forest algorithm is constructed based on multiple randomly generated isolated trees. The fault detection results are used to evaluate the abnormal conditions of each component in the spring operating mechanism and the severity of the fault.

[0178] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:

[0179] Perform feature extraction on the data collected by various types of sensors to obtain multiple sensor feature vectors;

[0180] The anomaly index of each sensor’s feature vector is calculated using the isolation forest algorithm;

[0181] The fault detection result is determined based on the abnormality index and a pre-set index interval, wherein different index intervals correspond to different severity levels.

[0182] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:

[0183] For each sensor feature vector, the isolation forest algorithm is used to perform isolation tree traversal on the sensor feature vector to obtain the average height of the sensor feature vector in each isolation tree;

[0184] The average height of each sensor's feature vector is normalized to obtain the abnormal index of each sensor's feature vector.

[0185] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:

[0186] A plurality of sample feature vectors obtained in advance are randomly selected to obtain a plurality of target sample points and target cutting points; each target sample point includes a plurality of vector dimensions;

[0187] Generate a hyperplane based on the target cutting point, and divide the data space according to the hyperplane to obtain two subspaces; one subspace corresponds to a child node of the isolated tree;

[0188] For each target sample point, the data of different vector dimensions in the target sample point are divided into different subspaces according to the target cutting point to obtain the corresponding child nodes;

[0189] The partitioning process is performed on each child node until the child node meets the termination condition after the partitioning, and an isolated tree is obtained. The isolation forest algorithm is constructed by multiple isolated trees.

[0190] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:

[0191] According to the pre-set sample length, each sample feature vector is randomly selected to obtain a sample subset; the sample subset includes multiple target sample points; the sample subset is used to create the root node of the isolation tree;

[0192] Randomly assign the vector dimension of the sample feature vector to obtain the target dimension;

[0193] The target cutting points are randomly generated from the target sample points of the sample subset according to the target dimension.

[0194] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:

[0195] According to the target dimension of the target cutting point, the data smaller than the target dimension in the target sample point is divided into the first subspace to obtain the left child node;

[0196] Divide the data in the target sample point that is greater than or equal to the target dimension into the second subspace to obtain the right child node.

[0197] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:

[0198] Determine the target index range corresponding to the abnormal index;

[0199] Determine the fault detection result based on the alarm level corresponding to the target index range;

[0200] Among them, the alarm level corresponding to the target index interval includes at least one of the first level alarm, the second level alarm and the third level alarm, and the fault detection result includes at least one of the slight deviation, the moderate deviation and the severe deviation of the sensor collected data.

[0201] According to some embodiments of the present application, a computer program product is also provided. When executed by a processor, the computer program can implement the above-mentioned method. The computer program product includes one or more computer instructions. When these computer instructions are loaded and executed on a computer, part or all of the above-mentioned method can be implemented in whole or in part according to the processes or functions described in the embodiments of the present application.

[0202] According to some embodiments of the present application, a non-transitory computer-readable storage medium including instructions is also provided, such as a memory including instructions. The instructions can be executed by a processor of an electronic device to perform the above method. For example, the non-transitory computer-readable storage medium can be a ROM, a random access memory (RAM), a CD-ROM, a magnetic tape, a floppy disk, an optical data storage device, etc.

[0203] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with the relevant laws, regulations and standards of relevant countries and regions.

[0204] Those skilled in the art will appreciate that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the above-mentioned embodiments. In particular, any reference to memory, database, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The databases involved in the various embodiments provided herein may include at least one of a relational database and a non-relational database. Non-relational databases may include, but are not limited to, distributed databases based on blockchains. The processors involved in the various embodiments provided herein may be, but are not limited to, general-purpose processors, central processing units (CPUs), graphics processing units (GPUs), digital signal processors (DSPs), programmable logic devices (PLDs), data processing logic devices based on quantum computing, and the like.

[0205] The technical features of the above embodiments can be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0206] The above-described embodiments merely represent several implementation methods of the present application. While the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the present application. It should be noted that a person of ordinary skill in the art may make various modifications and improvements without departing from the spirit of the present application, and these modifications and improvements fall within the scope of protection of the present application. Therefore, the scope of protection of the present application shall be determined by the appended claims.

Claims

1. A spring operating mechanism fault monitoring method, characterized in that: The method comprises: Detecting and processing data collected by multiple different types of sensors of the spring operating mechanism to obtain detection results; When the detection result indicates an abnormality, fault detection processing is performed on the detection result based on a pre-constructed isolation forest algorithm to obtain a fault detection result of the spring operating mechanism; the isolation forest algorithm is constructed based on multiple randomly generated isolated trees; the fault detection result is used to evaluate the abnormality of each component in the spring operating mechanism and the severity of the fault.

2. The method according to claim 1, characterized in that The fault detection processing is performed on the detection result based on the pre-built isolation forest algorithm to obtain a fault detection result, including: Performing feature extraction processing on the data collected by each type of sensor to obtain multiple sensor feature vectors; Calculating an anomaly index of each sensor feature vector using the isolation forest algorithm; The fault detection result is determined according to the abnormality index and a pre-set index interval, wherein different index intervals correspond to different severity levels.

3. The method according to claim 2, characterized in that The calculating of the abnormality index of each sensor feature vector by using the isolation forest algorithm includes: For each sensor feature vector, performing an isolation tree traversal process on the sensor feature vector using the isolation forest algorithm to obtain an average height of the sensor feature vector in each isolation tree; The average height of each sensor feature vector is normalized to obtain an abnormality index of each sensor feature vector.

4. The method according to claim 3, characterized in that The construction process of the isolation forest algorithm includes: Randomly selecting multiple sample feature vectors obtained in advance to obtain multiple target sample points and target cutting points; each of the target sample points includes multiple vector dimensions; Generate a hyperplane according to the target cutting point, and divide the data space according to the hyperplane to obtain two subspaces; one of the subspaces corresponds to a child node of the isolated tree; For each target sample point, dividing the data of different vector dimensions in the target sample point into different subspaces according to the target cutting point to obtain corresponding child nodes; A partitioning process is performed on each of the child nodes until the child nodes meet the termination condition after the partitioning, thereby obtaining the isolated tree, and constructing the isolation forest algorithm using a plurality of the isolated trees.

5. The method according to claim 4, characterized in that The randomly selected processing of the pre-acquired multiple sample feature vectors to obtain multiple target sample points and target cutting points includes: According to a preset sample length, each of the sample feature vectors is randomly selected to obtain a sample subset; the sample subset includes a plurality of the target sample points; the sample subset is used to create a root node of the isolation tree; Randomly assigning the vector dimension of the sample feature vector to obtain a target dimension; The target cutting point is randomly generated from the target sample points of the sample subset according to the target dimension.

6. The method according to claim 4, characterized in that The step of dividing the data of different vector dimensions in the target sample point into different subspaces according to the target cutting point to obtain corresponding child nodes includes: According to the target dimension of the target cutting point, data smaller than the target dimension in the target sample point is divided into a first subspace to obtain a left child node; The data in the target sample point that is greater than or equal to the target dimension is divided into a second subspace to obtain a right child node.

7. The method according to claim 2, characterized in that Determining the fault detection result according to the abnormality index and a preset index interval includes: Determining a target index range corresponding to the abnormal index; determining the fault detection result according to the alarm level corresponding to the target index interval; Among them, the alarm level corresponding to the target index interval includes at least one of a first-level alarm, a second-level alarm, and a third-level alarm, and the fault detection result includes at least one of a slight deviation, a moderate deviation, and a severe deviation in the sensor collected data.

8. A spring operating mechanism fault monitoring device, characterized in that: The device comprises: A data detection module is used to detect and process data collected by multiple different types of sensors of the spring operating mechanism to obtain detection results; A fault detection module is configured to, when the detection result indicates an abnormality, perform fault detection processing on the detection result based on a pre-constructed isolation forest algorithm to obtain a fault detection result for the spring operating mechanism; the isolation forest algorithm is constructed based on multiple randomly generated isolated trees; and the fault detection result is used to assess the abnormality of each component in the spring operating mechanism and the severity of the fault.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images