Equipment fault early warning method and system under multi-source internet-of-things perception and storage medium

Abstract

The invention provides an equipment fault early warning method and system under multi-source Internet of Things sensing and a storage medium, and relates to the technical field of equipment fault early warning, and the method comprises the steps: building a fault propagation rate model through obtaining the spatial position coordinates of a plurality of sensing nodes in multi-source Internet of Things sensing equipment; calling the model to calculate propagation delay indexes of sensing node pairs, outputting a propagation delay index matrix, outputting fault response time anchor points in combination with fault response time of the sensing nodes, performing time sequence delay calibration on the multiple sensing nodes, outputting a calibration time sequence, and extracting calibration multi-source sensing data to perform fault type identification. And outputting a fault early warning signal. According to the method and the device, the technical problems of insufficient data fusion precision and insufficient fault early warning accuracy and timeliness caused by difference of response time of the multi-source sensing nodes in the prior art are solved, and the technical effects of realizing time sequence delay calibration of the multi-source sensing data and further improving the fault early warning accuracy and timeliness are achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of equipment fault warning, and particularly to a method, a system and a storage medium for equipment fault warning under multi-source Internet of Things perception. Background Art

[0002] With the continuous development of the Internet of Things technology, equipment fault warning based on multi-source Internet of Things perception has gradually become an important technical means to ensure the safe and reliable operation of industrial equipment. In the Internet of Things perception system, multiple perception nodes are deployed at each key part of the equipment to collect the operation data of the equipment in real time. The existing equipment fault warning methods under Internet of Things perception generally focus on abnormal detection of data of a single perception node, or simple preliminary integration and feature extraction of multi-source perception data to identify potential fault risks. However, the existing methods ignore the differences in the response times of multi-source perception nodes and fail to fully consider the propagation laws of each sensing node under the spatial position distribution and physical topology, which leads to the difficulty in accurately corresponding the timing relationship of data during the data fusion process, resulting in insufficient data fusion accuracy. This multi-node timing deviation is likely to cause the mixing of fault features, reduce the accuracy of multi-modal data fusion, and further affect the timeliness and reliability of fault warning, restricting the effective utilization of multi-source perception data in equipment health state perception and fault warning. Summary of the Invention

[0003] The present invention provides a method, a system and a storage medium for equipment fault warning under multi-source Internet of Things perception, which solve the technical problems in the prior art that due to the differences in the response times of multi-source perception nodes, the data fusion accuracy is insufficient, and the accuracy and timeliness of fault warning are insufficient, and achieve the technical effect of realizing the time series delay calibration of multi-source perception data, and further improving the accuracy and timeliness of fault warning.

[0004] In view of the above problems, on the one hand, the present invention provides a method for equipment fault warning under multi-source Internet of Things perception, and the method includes: obtaining multiple spatial position coordinates corresponding to multiple perception nodes in a multi-source Internet of Things perception device; establishing a fault propagation rate model according to the multiple spatial position coordinates; combining the multiple perception nodes to obtain perception node pairs, calling the fault propagation rate model to calculate the propagation delay index of each perception node pair, and outputting a propagation delay index matrix; collecting multiple fault response times of the multiple perception nodes; outputting a fault response time anchor point based on the propagation delay index matrix and the multiple fault response times, performing time series delay calibration on the multiple perception nodes with the fault response time anchor point, outputting a calibrated time series, and extracting calibrated multi-source perception data under the calibrated time series; and identifying the fault type by using the calibrated multi-source perception data, and outputting a fault warning signal.

[0005] Preferably, a fault propagation rate model is established according to the multiple spatial position coordinates. The method further includes: inputting the multiple spatial position coordinates to obtain the signal propagation paths and propagation medium types corresponding to the multiple sensing nodes; collecting historical fault response signals of the multiple sensing nodes under known fault event samples according to the signal propagation paths and propagation medium types; inversely analyzing the historical fault response signal samples to set a corresponding preset propagation rate for each sensing node; and constructing a fault propagation rate model based on the relationship between each sensing node and the corresponding preset propagation rate for identifying the multiple propagation rates corresponding to the multiple sensing nodes.

[0006] Preferably, the multiple sensing nodes in the multi-source IoT sensing device include at least two of an acoustic sensing device, an image sensing device, a temperature sensing device, a current or voltage sensing device, and a pressure sensing device.

[0007] Preferably, the method for calculating the propagation delay index of each sensing node pair by invoking the fault propagation rate model includes: obtaining a first sensing node and a second sensing node of each sensing node pair; obtaining a first spatial position coordinate corresponding to the first sensing node and a second spatial position coordinate corresponding to the second sensing node; invoking the fault propagation rate model to identify a first propagation rate corresponding to the first sensing node and a second propagation rate corresponding to the second sensing node; and outputting the propagation delay index of each sensing node pair according to the first spatial position coordinate, the second spatial position coordinate, the first propagation rate, and the second propagation rate.

[0008] Preferably, the method for outputting the propagation delay index of each sensing node pair according to the first spatial position coordinate, the second spatial position coordinate, the first propagation rate, and the second propagation rate includes: outputting a first propagation path length and a second propagation path length according to the first spatial position coordinate and the second spatial position coordinate; obtaining a first propagation response index, where the first propagation response index is the ratio of the first propagation path length to the first propagation rate; obtaining a second propagation response index, where the second propagation response index is the ratio of the second propagation path length to the second propagation rate; calculating a delay index difference between the first propagation response index and the second propagation response index, and outputting the propagation delay index.

[0009] Preferably, the method for outputting a fault response time anchor based on the propagation delay index matrix and the multiple fault response times includes: constructing a minimum residual optimization model; calculating the propagation delay index matrix and the multiple fault response times by using the minimum residual optimization model to output the fault response time anchor; where the expression of the minimum residual optimization model is: ; where is the calculated fault response time anchor point, is the optimization variable selected from the candidate fault response time anchor points, is the actual corresponding fault response time of the i-th sensing node, is the propagation path length corresponding to the i-th sensing node, is the propagation rate corresponding to the i-th sensing node, is the propagation response index obtained by theoretical calculation, is the number of sensing nodes.

[0010] Preferably, time series delay calibration is performed on the multiple sensing nodes with the fault response time anchor point, and a calibrated time series is output. The expression is: ; where is the response time after time series delay calibration of the i-th sensing node, is the preset original time point.

[0011] Preferably, the calibrated multi-source sensing data is used for fault type identification, and a fault warning signal is output. The method includes: performing data fusion on the calibrated multi-source sensing data to obtain fused sensing data; extracting the fault feature vector of the fused sensing data, and performing fault type identification with the fault feature vector to output the fault warning signal corresponding to the fault type.

[0012] In a second aspect, the present invention also provides a device fault warning system under multi-source IoT sensing. The system includes: a node coordinate acquisition module for acquiring multiple spatial position coordinates corresponding to multiple sensing nodes in a multi-source IoT sensing device; a propagation rate modeling module for establishing a fault propagation rate model according to the multiple spatial position coordinates; a propagation delay calculation module for combining the multiple sensing nodes to obtain sensing node pairs, calling the fault propagation rate model to calculate the propagation delay index of each sensing node pair, and outputting a propagation delay index matrix; a response time acquisition module for acquiring multiple fault response times of the multiple sensing nodes; a delay calibration module for outputting a fault response time anchor point based on the propagation delay index matrix and the multiple fault response times, performing time series delay calibration on the multiple sensing nodes with the fault response time anchor point, outputting a calibrated time series, and extracting the calibrated multi-source sensing data under the calibrated time series; a fault identification and warning module for using the calibrated multi-source sensing data for fault type identification and outputting a fault warning signal.

[0013] In a third aspect, the present invention also provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps in the above-mentioned device fault warning method under multi-source IoT sensing are implemented.

[0014] One or more technical solutions provided in the present invention have at least the following beneficial effects: By obtaining multiple spatial position coordinates corresponding to multiple sensing nodes in multi-source Internet of Things sensing devices, key spatial dimension information is provided for subsequent accurate assessment of fault propagation. A fault propagation rate model is established based on the multiple spatial position coordinates to quantitatively describe the propagation speed and manner of faults between sensing nodes at different spatial positions, laying a theoretical foundation for subsequent accurate calculation of the propagation delay index, enabling the entire solution to deeply analyze fault propagation from the perspectives of time and space. By combining the multiple sensing nodes to obtain sensing node pairs, the fault propagation rate model is called to calculate the propagation delay index for each sensing node pair, and a propagation delay index matrix is output, concretizing the abstract fault propagation law into quantifiable indices, which can clearly reflect the time delay differences caused by fault propagation between different sensing node pairs, providing a precise quantitative basis for subsequent time series delay calibration of sensing nodes. By collecting multiple fault response times of the multiple sensing nodes, the original time information of the sensing nodes under actual fault conditions is obtained, and these original time information are the benchmark data sources for subsequent time series delay calibration. Based on the propagation delay index matrix and the multiple fault response times, a fault response time anchor point is output, and the multiple sensing nodes are calibrated for time series delay using the fault response time anchor point, a calibrated time series is output, and the calibrated multi-source sensing data under the calibrated time series is extracted. Through calibration, the inconsistency of multi-source sensing data in time is eliminated, enabling the data of different sensing nodes to be compared and fused on a unified time scale, providing a high-quality data basis for subsequent accurate fault type identification using the calibrated multi-source sensing data. Using the calibrated multi-source sensing data for fault type identification and outputting a fault warning signal, the calibrated multi-source sensing data can more accurately reflect the true operating state of the device. Using these data for fault type identification can accurately determine the type of the fault, and then timely output a fault warning signal, effectively avoiding problems such as device damage and production interruption caused by untimely or false fault warnings, and improving the reliability and stability of device operation.

[0015] In summary, through the acquisition of spatial position coordinates and the establishment of a propagation rate model, the present invention realizes the dynamic modeling of fault propagation among multi-source sensing nodes, and can accurately characterize the delay characteristics between nodes. Combining the propagation delay index matrix and the actual fault response time, multi-node timing delay calibration is performed based on the fault response time anchor points, solving the problem of data timing asynchrony caused by the difference in response time of multi-source sensing nodes. The calibrated multi-source sensing data can more truly reflect the operating state of the device, eliminate the aliasing of fault characteristics, significantly improve the fusion accuracy and consistency of multi-modal sensing data, and thus achieve more accurate and timely fault type identification and early warning, meeting the high-reliability requirements of industrial equipment for health management under complex working conditions.

[0016] The above description is only an overview of the technical solution of the present invention. In order to be able to understand the technical means of the present invention more clearly, it can be implemented according to the content of the specification. And in order to make the above and other objects, features and advantages of the present invention more obvious and understandable, the following specifically illustrates the specific embodiments of the present invention. Brief Description of the Drawings

[0017] Figure 1 It is a schematic flow chart of the device fault early warning method under multi-source IoT sensing provided by an embodiment of the present invention.

[0018] Figure 2 It is a schematic flow chart of calculating the propagation delay index for each pair of sensing nodes in the device fault early warning method under multi-source IoT sensing provided by an embodiment of the present invention.

[0019] Figure 3 It is a schematic structural diagram of the device fault early warning system under multi-source IoT sensing provided by an embodiment of the present invention.

[0020] Description of the reference numerals: Node coordinate acquisition module 10, propagation rate modeling module 20, propagation delay calculation module 30, response time acquisition module 40, delay calibration module 50, fault identification and early warning module 60. Detailed Embodiments

[0021] The embodiments of the present invention provide a device fault early warning method, system and storage medium under multi-source IoT sensing, solving the technical problems in the prior art that due to the difference in the response time of multi-source sensing nodes, the data fusion accuracy is insufficient, and the accuracy and timeliness of fault early warning are insufficient, and achieving the technical effect of realizing the calibration of the timing delay of multi-source sensing data, thereby improving the accuracy and timeliness of fault early warning.

[0022] Embodiment 1, as Figure 1 shown, the embodiments of the present invention provide a device fault early warning method under multi-source IoT sensing, and the method includes: Step S100: Obtain multiple spatial position coordinates corresponding to multiple sensing nodes in the multi-source Internet of Things sensing device.

[0023] Furthermore, at least two of the multiple sensing nodes in the multi-source Internet of Things sensing device include an acoustic sensing device, an image sensing device, a temperature sensing device, a current or voltage sensing device, and a pressure sensing device.

[0024] Specifically, the multi-source Internet of Things sensing device refers to an aggregate of various types of sensors deployed in a device or industrial environment, capable of sensing data such as the operating state of the device and physical signals. A sensing node refers to a physical sensor unit in the multi-source Internet of Things sensing device that actually undertakes the data acquisition function. In the embodiments of the present invention, at least two of the multiple sensing nodes in the multi-source Internet of Things sensing device include an acoustic sensing device, an image sensing device, a temperature sensing device, a current or voltage sensing device, and a pressure sensing device. Among them, the acoustic sensing device refers to a sensor that collects acoustic signals generated during the operation of the device through the principle of sound wave propagation, such as an acoustic emission sensor or an ultrasonic sensor; the image sensing device is a sensor based on the principle of optical imaging, such as an industrial camera or an infrared thermal imager; the temperature sensing device, such as a thermocouple or a thermistor, is used to sense temperature signals; the current or voltage sensing device detects electrical signals through a current sensor or a voltage sensor; the pressure sensing device is a pressure sensor or a piezoelectric sensor, used to detect pressure changes inside or around the device. By introducing at least two different types of sensors, multiple physical characteristics of the device operating state can be covered. For example, by simultaneously utilizing the complementarity of acoustic signals and temperature signals, joint monitoring of mechanical friction anomalies and heat accumulation can be achieved. Through the integration of multi-modal and multi-type sensing nodes, richer physical dimension data can be provided, laying a more comprehensive and diversified sensing foundation for subsequent fault propagation rate models and data fusion analysis.

[0025] First, through a known three-dimensional model of the device or a sensor installation data sheet, obtain the accurate spatial position coordinate information of each multi-source sensing node in the device space. These spatial position coordinates are generally presented in the form of three-dimensional rectangular coordinates. For example, a certain temperature sensor is located at (1.2m, 0.5m, 0.3m), and a certain image sensor is located at (0.8m, -0.3m, 0.7m), etc. Through the digital twin model or CAD model of the device, map the spatial position coordinates of the sensing nodes to the actual device structure to ensure that the positions of all sensing nodes are accurately marked under the same spatial reference coordinate system. Further, store these coordinate data in a position database for subsequent propagation rate modeling and timing delay calibration.

[0026] The spatial position coordinates of multiple sensing nodes obtained through this step can provide an accurate basis for the spatial geometric relationship for subsequent modeling of the fault propagation rate, ensuring that the propagation paths and propagation distances between multi-source sensing nodes can be quantified and modeled, and improving the spatial consistency and computational accuracy of multi-source sensing data fusion.

[0027] Step S200: Establish a fault propagation rate model based on the multiple spatial position coordinates.

[0028] Specifically, the fault propagation rate model refers to a mathematical model that combines the spatial positions of sensing nodes and historical fault signal samples to deduce the propagation speed law of the sensing response of each sensing node under a fault event. First, input the spatial position coordinates of each sensing node, and combine the device structure, material information, and sensor type to deduce the propagation paths of the sensing signals between the nodes. Further obtain the known historical fault response signal data of the device (such as mechanical shock sound waves, heat diffusion characteristics, etc.), and by playing back these historical fault events, record the response times of each sensing node to the known faults. Then, based on these historical samples, invert the propagation rates of the sensing nodes in different propagation media (such as the speed of sound in metal components, the speed of sound in air, etc.), thereby constructing the preset propagation rate of each sensing node. Combine these propagation rates with the spatial position relationship to generate a complete propagation rate model.

[0029] By establishing the propagation rate model through this step, it can truly reflect the signal propagation laws of different sensing nodes at different spatial positions and physical paths, making the subsequent calculation of the propagation delay index more scientific and accurate, and ensuring the comparability of the time series of data.

[0030] Step S300: Combine the multiple sensing nodes to obtain sensing node pairs, call the fault propagation rate model to calculate the propagation delay index of each sensing node pair, and output a propagation delay index matrix.

[0031] Specifically, a sensing node pair refers to the combination of any two sensing nodes and is the basic unit for analyzing the signal propagation delay relationship between nodes; the propagation delay index is a quantitative index representing the delay magnitude of the signal propagating from one sensing node to another; the propagation delay index matrix is a two-dimensional matrix, and each element in the matrix represents the propagation delay index of a node pair.

[0032] Enumerate all pairs of sensing nodes pairwise to generate all possible pairs of sensing nodes. For each pair of sensing nodes, call the propagation rate model respectively to obtain the propagation path length and propagation rate of this node pair. Divide the propagation path length by the propagation rate to obtain the propagation response metric, and then calculate the propagation delay metric by comparing the differences in the propagation response metrics between node pairs. Finally, summarize the propagation delay metrics of all node pairs in matrix form to form a propagation delay metric matrix for subsequent time calibration optimization. For example, there are four sensing nodes A, B, C, and D, and the combined node pairs are such as A - B, A - C, A - D, B - C, etc. Using the fault propagation rate model, their propagation delays are obtained respectively (for example, A - B is 0.001 seconds, A - C is 0.002 seconds, A - D is 0.001 seconds, B - C is 0.003 seconds, etc.), and they are summarized into a 4×4 propagation delay metric matrix.

[0033] By obtaining the propagation delay metric matrix through this step, it provides accurate propagation delay constraints for subsequent time series calibration based on global propagation residual optimization, improving the accuracy of delay calibration.

[0034] Step S400: Collect the multiple fault response times of the multiple sensing nodes.

[0035] Specifically, the fault response time refers to the timestamp when the sensing node detects the response signal after the occurrence of the device fault event, which is the key timing information in the sensing node data. During the operation of the device, the response data of each sensing node is collected in real time, and the time points when each node senses the fault feature signal are recorded to form a set of fault response time series. These fault response time series record the sensing delays of each node for the same fault event and are important original data for subsequent time calibration.

[0036] Step S500: Output a fault response time anchor point based on the propagation delay metric matrix and the multiple fault response times, perform time series delay calibration on the multiple sensing nodes with the fault response time anchor point, output the calibrated time series, and extract the calibrated multi-source sensing data under the calibrated time series.

[0037] Specifically, the fault response time anchor point refers to the reference time point that can best represent the actual fault occurrence time globally, which is obtained by optimizing the residual function; time series delay calibration refers to correcting the timing data of each sensing node according to the anchor point and the propagation delay to ensure the alignment of multi-node data, and the corrected result is recorded as the calibrated time series. The calibrated multi-source sensing data is the multi-source sensing data extracted based on the calibrated time series, and these data are aligned in time and can be more accurately fused and analyzed.

[0038] First, using the minimum residual optimization model, the propagation delay metric matrix and the original response time series of the sensing nodes are jointly input to solve for a globally optimal fault response time anchor. Then, based on this fault response time anchor, the response time of each node is sequentially delay-calibrated and corrected according to the propagation response metric difference to obtain a set of calibrated time series with consistent timing and minimized errors. Finally, based on the calibrated timestamps, the calibrated multi-source sensing data of each node is extracted to ensure the timing consistency of subsequent data fusion analysis.

[0039] Through this step, the time series delay calibration of multiple nodes is completed, significantly reducing the timing error in multi-modal data fusion and improving the accuracy of the fusion results.

[0040] Step S600: Use the calibrated multi-source sensing data to identify the fault type and output a fault warning signal.

[0041] Specifically, fault type identification refers to using the fused multi-modal features to classify and judge the fault type by using pattern recognition or machine learning methods; the fault warning signal is a warning message or control instruction issued for the detected potential fault type. Based on the calibrated multi-source sensing data, first, feature engineering processing is performed to extract the key fault feature vectors of the calibrated multi-source sensing data. Then, machine learning algorithms (such as support vector machines, random forests, deep neural networks, etc.) or rule-based fault feature matching are used to classify or perform pattern recognition on the multi-source feature vectors to judge the type and severity of the fault. Finally, according to the recognition result, a corresponding fault warning signal is generated for use by the monitoring center or the upper-layer control system to achieve real-time warning and remote monitoring of the equipment.

[0042] This step uses the multi-modal calibrated sensing data to identify the fault type and outputs a reliable warning signal, achieving more accurate and intelligent equipment fault warning and improving the safety and operation and maintenance efficiency of industrial equipment.

[0043] Further, step S200 includes: Step S210: Input the multiple spatial position coordinates to obtain the signal propagation paths and propagation medium types corresponding to the multiple sensing nodes.

[0044] Step S220: According to the signal propagation paths and propagation medium types, collect the historical fault response signals of the multiple sensing nodes under known fault event samples.

[0045] Step S230: Invert the historical fault response signal samples and set a corresponding preset propagation rate for each sensing node.

[0046] Step S240: Based on the relationship between each sensing node and the corresponding preset propagation rate, construct a fault propagation rate model for identifying the multiple propagation rates corresponding to the multiple sensing nodes.

[0047] Specifically, the propagation path refers to the possible physical path between the fault source (the device to be detected) and each sensing node; the propagation medium type refers to the medium during signal transmission, such as air, metal, hydraulic fluid, etc. Take the multiple spatial position coordinates corresponding to the multiple sensing nodes obtained in step S100 as input information, and determine the propagation path between each sensing node and the potential fault source according to the structure, materials of the device, and sensor distribution, in combination with CAD models or BIM models, etc. Identify the propagation medium type according to the medium involved in the path (for example, acoustic wave propagation in air, stress wave propagation in metal, etc.).

[0048] A fault event sample refers to the typical fault event data with known occurrences and known response times of each sensing node. Extract multiple typical fault events (such as mechanical shock, thermal runaway, etc.) from the historical operation records of the device, and record the original signal response sequences and timestamps of each sensing node under these faults. During acquisition, ensure that the response signals of each node can be repeatedly obtained and cover various combinations of propagation paths and media to enhance the generalization ability of the propagation rate model.

[0049] Inversion refers to calculating the propagation rate based on the response time difference, spatial position, and path information; the preset propagation rate refers to the propagation rate set for each sensing node during the model initialization stage. Through the least squares method or genetic algorithm, etc., use the response time difference of each node to known faults, the propagation path length, and the propagation medium type to inversely calculate the propagation rate of each node pair. Then, statistically analyze the propagation rates of each node under multiple historical samples and perform mean or weighted mean processing to obtain the preset propagation rate of each sensing node.

[0050] Based on the above preset propagation rate, combined with the spatial distribution of the nodes, construct a rate model containing multi-dimensional information of propagation rate, path length, and propagation medium. This model can adopt matrix form, graph neural network, or propagation delay expression based on physical formulas, and can quickly query the propagation rate of each node pair for generating the subsequent propagation delay index matrix.

[0051] Further, as Figure 2 shown, step S300 includes: Step S310: Obtain the first sensing node and the second sensing node of each sensing node pair.

[0052] Step S320: Obtain the first spatial position coordinate corresponding to the first sensing node and the second spatial position coordinate corresponding to the second sensing node.

[0053] Step S330: Invoke the fault propagation rate model to identify the first propagation rate corresponding to the first sensing node and the second propagation rate corresponding to the second sensing node.

[0054] Step S340: Output the propagation delay index for each pair of sensing nodes according to the first spatial position coordinate, the second spatial position coordinate, the first propagation rate, and the second propagation rate.

[0055] Specifically, pair up all the sensing nodes to form a number of pairs of sensing nodes. Each pair of nodes contains a first sensing node and a second sensing node, and their numbers or IDs are recorded respectively. Retrieve the three-dimensional spatial position coordinates of the first sensing node and the second sensing node from the database storing multiple spatial position coordinates, denoted as the first spatial position coordinate and the second spatial position coordinate.

[0056] Using the previously established fault propagation rate model, take the first sensing node and the second sensing node as inputs, and through look-up tables, calculations, and model computations, obtain the first propagation rate corresponding to the first sensing node and the second propagation rate corresponding to the second sensing node. Based on the obtained spatial position coordinates and the identified propagation rates, first calculate the spatial distances from the first sensing node and the second sensing node to the fault source. Then, divide the spatial distance from the first spatial position coordinate to the fault point (the first propagation path length) by the first propagation rate, and divide the spatial distance from the second spatial position coordinate to the fault point (the second propagation path length) by the second propagation rate to obtain their respective propagation response times. The difference between the two is the propagation delay index. Finally, organize the propagation delay indices of all pairs of nodes into a matrix form to obtain the propagation delay index matrix, which is convenient for subsequent global optimization of time anchor points.

[0057] Further, step S340 includes: Step S341: Output the first propagation path length and the second propagation path length according to the first spatial position coordinate and the second spatial position coordinate.

[0058] Step S342: Obtain the first propagation response index, where the first propagation response index is the ratio of the first propagation path length to the first propagation rate.

[0059] Step S343: Obtain the second propagation response index, where the second propagation response index is the ratio of the second propagation path length to the second propagation rate.

[0060] Step S344: Calculate the delay index difference between the first propagation response index and the second propagation response index, and output the propagation delay index.

[0061] Specifically, the propagation response index is the result of dividing the signal propagation path length by the propagation rate, representing the propagation time from the node to the fault point. Using the three-dimensional Euclidean distance formula, the straight-line distances between the first sensing node and the second sensing node and the fault source are calculated respectively to obtain the first propagation path length and the second propagation path length. According to the known first propagation path length and the first propagation rate, the ratio of the first propagation path length to the first propagation rate is calculated and denoted as the first propagation response index. Similarly, the ratio of the second propagation path length to the second propagation rate is calculated and denoted as the second propagation response index. The absolute value of the difference between the two calculated propagation response indexes is obtained to get the propagation delay index between the first sensing node and the second sensing node. The larger the value of this propagation delay index, the more obvious the response difference between the first sensing node and the second sensing node when sensing the same fault event. According to the above calculation method, the propagation delay indexes between all pairs of sensing nodes are determined and combined into a propagation delay index matrix to intuitively reflect the response differences among multiple sensing nodes.

[0062] Further, step S500 includes: Step S510: Construct a minimum residual optimization model.

[0063] Step S520: Use the minimum residual optimization model to calculate the propagation delay index matrix and the multiple fault response times, and output the fault response time anchor points. The expression of the minimum residual optimization model is: ; where is the calculated fault response time anchor point, is the optimization variable selected from the candidate fault response time anchor points, is the actual fault response time corresponding to the i-th sensing node, is the propagation path length corresponding to the i-th sensing node, is the propagation rate corresponding to the i-th sensing node, is the propagation response index obtained by theoretical calculation, is the number of sensing nodes.

[0064] Specifically, the minimum residual optimization model is a mathematical optimization model that obtains the optimal fault response time anchor point by minimizing the sum of the squares of the differences between the theoretical propagation response index and the actual fault response time. The fault response time anchor point refers to the globally unified theoretical starting time point of the fault response, serving as a reference benchmark for the time series calibration of multiple sensing nodes. The candidate fault response time anchor point refers to the value range of the optimization variable T in the minimum residual optimization model, that is, multiple possible starting points of the fault response time tried during the search process. Its range can be set according to the distribution of the actual response times of multiple sensing nodes, the equipment working conditions, or historical empirical values.

[0065] First, based on the propagation delay index matrix and the actual fault response time of each sensing node, a minimum residual optimization model is built. The model takes minimizing the sum of square errors between the theoretical propagation delay index and the actual response time of each sensing node as the objective function, which is formally expressed as follows: ,in, is the calculated fault response time anchor point, is the optimization variable selected within the candidate fault response time anchor point, is the actual fault response time corresponding to the i-th sensing node, is the propagation path length corresponding to the i-th sensing node, is the propagation rate corresponding to the i-th sensing node, is the propagation response index obtained by theoretical calculation, is the number of sensing nodes.

[0066] Call the minimum residual optimization model constructed above and input all perception nodes ( The actual response time of sensor nodes (n is a positive integer) ), propagation path length ( ) and the propagation rate ( ), and the propagation delay indicator matrix. The mean range of all theoretical propagation delay indicators is used as the center of the candidate fault response time anchor point. Then, combined with the actual working conditions of the multi-source IoT perception system (such as the response delay segment after the device is started or the sensor works stably), a certain time range (such as ±10ms or ±5%) is extended forward and backward to form a set of candidate fault response time anchor points. The minimum residual optimization model iteratively searches within the range of candidate fault response time anchor points (T) until the sum of squared errors is minimized and the optimal output is obtained. The value is used as the global fault response time anchor point of the entire perception system. The iterative search process can be implemented with the help of optimization methods such as gradient descent algorithm and least squares fitting, or it can be automatically solved through the existing numerical optimization library to finally obtain an accurate fault response time anchor point.

[0067] Furthermore, the multiple sensing nodes are subjected to time series delay calibration based on the fault response time anchor point, and a calibration time series is output, which is expressed as follows: ;in, is the response time of the i-th sensor node after time series delay calibration, The preset original time point.

[0068] Specifically, the fault response time anchor point obtained by minimum residual optimization is , for the original response time series collected by each sensing node, based on the propagation path length ( ) and the propagation rate ( ), combined with the preset original time point ( ), perform time series delay calibration to eliminate the time delay differences in multi-source perception data, align the response times of each perception node under the same global time benchmark, and obtain the calibrated response time of each perception node under the global time benchmark, that is, the calibrated time series. The specific calibration expression is as follows: The expression is: ; where is the response time after time series delay calibration of the i-th perception node, is the preset original time point, usually set to 0.

[0069] The core of this calibration step lies in quantifying and compensating the time delay of each perception node, synchronizing the response time series of each node within the framework of the theoretical propagation model, avoiding time drift problems caused by different node distribution positions or propagation medium differences, and thus providing a timing basis for subsequent multi-modal data fusion and fault type identification to achieve more efficient and reliable device fault warning.

[0070] Furthermore, step S600 includes: Step S610: Use the calibrated multi-source perception data for data fusion to obtain fused perception data.

[0071] Step S620: Extract the fault feature vector of the fused perception data, perform fault type identification based on the fault feature vector, and output the fault warning signal corresponding to the fault type.

[0072] Specifically, data fusion refers to integrating multi-source data from different perception nodes according to a certain fusion strategy to form complete, consistent, and minimized-redundancy global perception information. The fault feature vector refers to the set of features that can characterize the fault mode extracted from the fused multi-modal perception data, which can be represented by statistical features, time-frequency domain features, or deep features (such as convolutional feature vectors, etc.).

[0073] First, obtain the calibrated multi-source perception data of each perception node, and perform data fusion operations on the premise of time alignment. Data fusion can be carried out based on methods such as weighted average, multi-modal feature splicing, Bayesian inference, or deep neural network multi-modal fusion. For example, use tools such as TensorFlow or PyTorch for tensor splicing or self-attention mechanism integration to obtain fused perception data. Subsequently, extract the feature vectors used to distinguish fault types from the fused perception data. These features can include statistical metrics of time series (such as mean, standard deviation, peak value, etc.), or spectral energy distribution (such as the amplitude spectrum after Fourier transform), or deep features extracted through a convolutional neural network. Finally, based on known training samples or expert experience, use classifiers such as support vector machines, random forests, or convolutional neural networks to perform pattern recognition on the fault feature vectors, output the recognition results of the fault types, and correspondingly generate corresponding fault warning signals. To ensure real-time performance and scalability, this process is usually executed on an edge computing gateway or a cloud server to ensure that the fused perception data can be quickly converted into accurate warning results.

[0074] Taking a wind turbine device as an example, using multi-modal data collected by multiple perception nodes installed in the unit (including acoustic sensors, temperature sensors, and current sensors), after the time series delay calibration is completed, calibrated acoustic signals, temperature change curves, and current fluctuation sequences are obtained. Then, perform data fusion on the calibrated multi-modal perception data: convert the acoustic signal into a time-frequency spectrogram through short-time Fourier transform, and splice the time series signals of temperature and current to the channel dimension of the acoustic time-frequency spectrogram after normalization to form a multi-channel fused perception data tensor. Then, input this fused perception data tensor into a fault diagnosis model based on a convolutional neural network. The convolutional layer extracts local features of the fused perception data in the spatial and time-frequency domains through multiple convolutional kernels, and the pooling layer performs feature compression to remove noise and redundant information. Subsequently, high-level semantic feature mapping and fusion are performed in the fully connected layer, and finally the probability distribution of each fault type is output in the softmax layer. For example, the fault diagnosis model is trained to distinguish three fault types: bearing wear, gear defect, and normal state. When the model infers a certain set of calibrated fusion data, the softmax layer outputs the highest probability for the bearing wear category, indicating that the fault type matched by the current multi-modal perception data is bearing wear. Finally, according to the output result of the fault diagnosis model, generate a corresponding fault warning signal and push it to the operation and maintenance personnel or the remote monitoring center to achieve intelligent warning.

[0075] By fusing the calibrated multi-source perception data, the complementary advantages of multi-modal perception can be fully utilized, the limitations of a single perception method can be overcome, and the accuracy and robustness of fault recognition can be effectively improved. Finally, based on the feature vectors extracted from the fused perception data, the fault type is recognized and a warning signal is output, significantly improving the comprehensive perception ability of the device operating state and the intelligent warning level, and realizing the closed-loop control from multi-source data acquisition to accurate warning.

[0076] In summary, the device fault warning method under multi-source Internet of Things perception provided by the embodiments of the present invention has the following beneficial effects: A device fault warning method under multi-source Internet of Things perception provided by an embodiment of the present invention first establishes a fault propagation rate model that can reflect the response differences of multiple nodes based on the spatial position coordinates corresponding to multiple sensing nodes in a multi-source Internet of Things sensing device; by combining sensing node pairs and invoking the fault propagation rate model, a propagation delay index matrix for each pair of sensing nodes is derived, effectively quantifying the delay deviation caused by spatial topology and propagation rate differences; further, the actual fault response time of each sensing node is collected, and a fault response time anchor point is derived through a minimum residual optimization model. Combining the propagation delay index and a preset time point, the time series delay calibration formula is used to correct the time series of all sensing node data; finally, based on the calibrated multi-source perception data, data fusion and feature extraction are used to derive fault feature vectors and identify fault types, realizing high-precision and low-latency fault warning that integrates multi-dimensional perception data.

[0077] Overall, through the acquisition of spatial position coordinates and the establishment of a propagation rate model in the embodiments of the present invention, dynamic modeling of fault propagation among multi-source sensing nodes is realized, and the delay characteristics between nodes can be accurately characterized. Combining the propagation delay index matrix and the actual fault response time, multi-node time series delay calibration is performed based on the fault response time anchor point, solving the problem of data time series out-of-sync caused by the response time differences of multi-source sensing nodes. The calibrated multi-source perception data can more truly reflect the device operating state, eliminate the aliasing of fault features, significantly improve the fusion accuracy and consistency of multi-modal perception data, and further realize higher-accuracy and timeliness fault type recognition and warning, meeting the high-reliability requirements of industrial equipment for health management under complex working conditions.

[0078] Embodiment 2, as Figure 3 shown, based on the same inventive concept as in the foregoing Embodiment 1, an embodiment of the present invention provides a device fault warning system under multi-source Internet of Things perception, and the system includes: A node coordinate acquisition module 10, configured to acquire a plurality of spatial position coordinates corresponding to a plurality of sensing nodes in a multi-source Internet of Things sensing device.

[0079] The propagation rate modeling module 20 is used to establish a fault propagation rate model based on the multiple spatial position coordinates.

[0080] The propagation delay calculation module 30 is used to combine the multiple sensing nodes to obtain sensing node pairs, call the fault propagation rate model to calculate the propagation delay index of each sensing node pair, and output a propagation delay index matrix.

[0081] The response time acquisition module 40 is used to acquire the multiple fault response times of the multiple sensing nodes.

[0082] The delay calibration module 50 is used to output a fault response time anchor point based on the propagation delay index matrix and the multiple fault response times, perform time series delay calibration on the multiple sensing nodes with the fault response time anchor point, output a calibrated time series, and extract the calibrated multi-source sensing data under the calibrated time series.

[0083] The fault identification and warning module 60 is used to identify the fault type using the calibrated multi-source sensing data and output a fault warning signal.

[0084] Furthermore, the multiple sensing nodes in the multi-source IoT sensing device include at least two of an acoustic sensing device, an image sensing device, a temperature sensing device, a current or voltage sensing device, and a pressure sensing device.

[0085] Furthermore, the propagation rate modeling module 20 is further used to perform the following steps: Input the multiple spatial position coordinates, obtain the signal propagation paths and propagation medium types corresponding to the multiple sensing nodes; according to the signal propagation paths and propagation medium types, collect the historical fault response signals of the multiple sensing nodes under known fault event samples; invert the historical fault response signal samples, set a corresponding preset propagation rate for each sensing node; based on the relationship between each sensing node and the corresponding preset propagation rate, construct a fault propagation rate model for identifying the multiple propagation rates corresponding to the multiple sensing nodes.

[0086] Furthermore, the propagation delay calculation module 30 is further used to perform the following steps: Obtain the first sensing node and the second sensing node of each sensing node pair; obtain the first spatial position coordinate corresponding to the first sensing node and the second spatial position coordinate corresponding to the second sensing node; call the fault propagation rate model to identify the first propagation rate corresponding to the first sensing node and the second propagation rate corresponding to the second sensing node; output the propagation delay index of each sensing node pair according to the first spatial position coordinate, the second spatial position coordinate, the first propagation rate, and the second propagation rate.

[0087] Further, the propagation delay calculation module 30 is further configured to perform the following steps: Output a first propagation path length and a second propagation path length according to the first spatial position coordinate and the second spatial position coordinate; obtain a first propagation response index, where the first propagation response index is the ratio of the first propagation path length to the first propagation rate; obtain a second propagation response index, where the second propagation response index is the ratio of the second propagation path length to the second propagation rate; calculate a delay index difference between the first propagation response index and the second propagation response index, and output a propagation delay index.

[0088] Further, the delay calibration module 50 is further configured to perform the following steps: Construct a minimum residual optimization model; use the minimum residual optimization model to calculate the propagation delay index matrix and the multiple fault response times, and output a fault response time anchor point; where the expression of the minimum residual optimization model is: ; where is the calculated fault response time anchor point, is an optimization variable selected from candidate fault response time anchor points, is the actual fault response time corresponding to the i-th sensing node, is the propagation path length corresponding to the i-th sensing node, is the propagation rate corresponding to the i-th sensing node, is the theoretically calculated propagation response index, is the number of sensing nodes.

[0089] Further, the fault identification and warning module 60 performs time series delay calibration on the multiple sensing nodes with the fault response time anchor point, and outputs a calibrated time series, and the expression is: ; where is the response time after time series delay calibration of the i-th sensing node, is a preset original time point.

[0090] Further, the fault identification and warning module 60 is further configured to perform the following steps: Use the calibrated multi-source sensing data for data fusion to obtain fused sensing data; extract a fault feature vector of the fused sensing data, and perform fault type identification with the fault feature vector, and output a fault warning signal corresponding to the fault type.

[0091] Through the foregoing detailed description of the device fault warning method under multi-source Internet of Things perception, those skilled in the art can clearly know the device fault warning system under multi-source Internet of Things perception in this embodiment. For the system disclosed in Embodiment 2, since it corresponds to the method disclosed in Embodiment 1, it has corresponding functional modules and beneficial effects. For the relevant parts, refer to the description in the method section.

[0092] Embodiment 3, based on the same inventive concept as the device fault warning method under multi-source Internet of Things perception in Embodiment 1 described above, the present invention also provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, it implements each step of the above-mentioned device fault warning method embodiment under multi-source Internet of Things perception and can achieve the same technical effects. To avoid repetition, it will not be elaborated here.

[0093] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present invention. Various modifications to these embodiments will be obvious to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to these embodiments shown herein, but will be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A device fault warning method under multi-source Internet of Things perception, characterized in that The method includes: Obtaining multiple spatial position coordinates corresponding to multiple sensing nodes in a multi-source Internet of Things sensing device; Establishing a fault propagation rate model based on the multiple spatial position coordinates; Combining the multiple sensing nodes to obtain sensing node pairs, invoking the fault propagation rate model to calculate the propagation delay index of each sensing node pair, and outputting a propagation delay index matrix; Collecting multiple fault response times of the multiple sensing nodes; Outputting a fault response time anchor point based on the propagation delay index matrix and the multiple fault response times, performing time series delay calibration on the multiple sensing nodes with the fault response time anchor point, outputting a calibrated time series, and extracting calibrated multi-source sensing data under the calibrated time series; Using the calibrated multi-source sensing data to identify a fault type and outputting a fault warning signal.

2. The device fault warning method under multi-source Internet of Things perception according to claim 1, characterized in that, When establishing a fault propagation rate model based on the multiple spatial position coordinates, the method further includes: Inputting the multiple spatial position coordinates to obtain the signal propagation paths and propagation medium types corresponding to the multiple sensing nodes; Collecting historical fault response signals of the multiple sensing nodes under known fault event samples according to the signal propagation paths and propagation medium types; Inverting the historical fault response signal samples and setting a corresponding preset propagation rate for each sensing node; Constructing a fault propagation rate model based on the relationship between each sensing node and the corresponding preset propagation rate for identifying the multiple propagation rates corresponding to the multiple sensing nodes.

3. The device fault warning method under multi-source IoT perception according to claim 1, characterized in that, The multiple sensing nodes in the multi-source Internet of Things sensing device include at least two of an acoustic sensing device, an image sensing device, a temperature sensing device, a current or voltage sensing device, and a pressure sensing device.

4. The device fault warning method under multi-source Internet of Things perception according to claim 1, characterized in that, When invoking the fault propagation rate model to calculate the propagation delay index of each sensing node pair, the method includes: Obtaining a first sensing node and a second sensing node of each sensing node pair; Obtaining a first spatial position coordinate corresponding to the first sensing node and a second spatial position coordinate corresponding to the second sensing node; Invoking the fault propagation rate model to identify a first propagation rate corresponding to the first sensing node and a second propagation rate corresponding to the second sensing node; Outputting the propagation delay index of each sensing node pair according to the first spatial position coordinate, the second spatial position coordinate, the first propagation rate, and the second propagation rate.

5. The device fault warning method under multi-source IoT perception according to claim 4, characterized in that, When outputting the propagation delay index of each sensing node pair according to the first spatial position coordinate, the second spatial position coordinate, the first propagation rate, and the second propagation rate, the method includes: Outputting a first propagation path length and a second propagation path length according to the first spatial position coordinate and the second spatial position coordinate; Obtaining a first propagation response index, where the first propagation response index is the ratio of the first propagation path length to the first propagation rate; Obtaining a second propagation response index, where the second propagation response index is the ratio of the second propagation path length to the second propagation rate; Calculating a delay index difference between the first propagation response index and the second propagation response index and outputting a propagation delay index.

6. The device fault warning method under multi-source IoT perception according to claim 1, characterized in that, Output a fault response time anchor point based on the propagation delay metric matrix and the multiple fault response times. The method includes: Construct a minimum residual optimization model; Use the minimum residual optimization model to calculate the propagation delay metric matrix and the multiple fault response times, and output a fault response time anchor point; Wherein, the expression of the minimum residual optimization model is: ； Among them, is the calculated fault response time anchor point, is the optimization variable selected from the candidate fault response time anchor points, is the actual fault response time corresponding to the i-th sensing node, is the propagation path length corresponding to the i-th sensing node, is the propagation rate corresponding to the i-th sensing node, is the propagation response index obtained by theoretical calculation, is the number of sensing nodes.

7. The device fault warning method under multi-source IoT perception according to claim 6, wherein, Use the fault response time anchor point to perform time series delay calibration on the multiple sensing nodes, and output a calibrated time series. The expression is: ； Among them, is the response time after time series delay calibration for the i-th sensing node, is the preset original time point.

8. The device fault warning method under multi-source Internet of Things perception according to claim 1, characterized in that, Use the calibrated multi-source sensing data to identify the fault type and output a fault warning signal. The method includes: Use the calibrated multi-source sensing data for data fusion to obtain fused sensing data; Extract the fault feature vector of the fused sensing data, and use the fault feature vector to identify the fault type and output the fault warning signal corresponding to the fault type.

9. Device fault warning system under multi-source Internet of Things perception, characterized in that, The system is used to execute the device fault warning method under multi-source IoT sensing according to any one of claims 1-8, and includes: A node coordinate acquisition module, configured to acquire multiple spatial position coordinates corresponding to multiple sensing nodes in a multi-source IoT sensing device; A propagation rate modeling module, configured to establish a fault propagation rate model according to the multiple spatial position coordinates; A propagation delay calculation module, configured to combine the multiple sensing nodes to obtain a sensing node pair, call the fault propagation rate model to calculate the propagation delay metric of each sensing node pair, and output a propagation delay metric matrix; A response time acquisition module, configured to acquire multiple fault response times of the multiple sensing nodes; A delay calibration module, configured to output a fault response time anchor point based on the propagation delay metric matrix and the multiple fault response times, perform time series delay calibration on the multiple sensing nodes with the fault response time anchor point, output a calibrated time series, and extract the calibrated multi-source sensing data under the calibrated time series; A fault identification and warning module, configured to use the calibrated multi-source sensing data to identify the fault type and output a fault warning signal.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the device fault warning method under multi-source IoT sensing according to any one of claims 1-8.

Images