Methods, devices, equipment, and storage media for anomaly monitoring of nonlinear waveform signals
By acquiring the recurrence graph and correlation recurrence graph of the nonlinear signal, a detection control chart is established, which solves the problem of monitoring subtle local anomalies in nonlinear signals and improves the monitoring accuracy.
Patent Information
- Application Number
- CN202311148259.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2023-09-07
- Publication Date
- 2026-11-17
- Estimated Expiration
- 2043-09-07
AI Technical Summary
In existing technologies, it is difficult to effectively monitor subtle local anomalies in nonlinear signals.
By acquiring the target recursion graph of the target nonlinear signal, determining the associated recursion graph, establishing a detection control graph for abnormal signals, estimating control limits, and determining whether the target statistical characteristic quantity is within the control limits, the abnormality of the nonlinear signal can be monitored.
It improves the accuracy of monitoring subtle local anomalies in nonlinear signals and enhances the distinction between normal and abnormal signal characteristic values.
Smart Images

Figure CN117272197B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of signal processing and analysis technology, and in particular to a method, apparatus, device, and storage medium for anomaly monitoring of nonlinear waveform signals. Background Technology
[0002] With the development of sensor and computer technologies, massive amounts of data are being collected, including nonlinear waveform signals such as vibration, temperature, pressure, audio, electrocardiograms, pulse, and electromagnetic waves. This data covers all aspects of national welfare, including industrial production, healthcare, transportation, agriculture, and military applications, and contains a vast amount of information. For example, radar echo signals not only contain the echo characteristics of the target but also its structure, shape, and subtle movements. Furthermore, the target's speed and trajectory can be deduced from the echo signal. Traditional wavelet transform analysis of nonlinear signals often focuses on changes in a specific portion of wavelet coefficients, easily neglecting changes in other parts. This makes it difficult to detect and identify anomalies in nonlinear waveform signals, leading to difficulties in effectively monitoring subtle local anomalies. Summary of the Invention
[0003] The main objective of this invention is to provide a method, apparatus, device, and storage medium for monitoring anomalies in nonlinear waveform signals, aiming to solve the technical problem that it is difficult to effectively monitor subtle local anomalies in nonlinear signals in the prior art.
[0004] To achieve the above objectives, the present invention provides a method for anomaly monitoring of nonlinear waveform signals, the method comprising the following steps:
[0005] Acquire the target nonlinear signal and determine the target recurrence graph of the target nonlinear signal;
[0006] Based on the target recursion graph and the baseline recursion graph, determine the associated recursion graph;
[0007] Determine the target statistical features of the recurrence graph;
[0008] Establish a detection control chart for abnormal signals and estimate the control limits of the detection control chart;
[0009] Determine whether the target statistical characteristic quantity is within the control limit, obtain the determination result, and monitor the target nonlinear signal based on the determination result.
[0010] Optionally, before determining the associated recursive graph based on the target recursive graph and the baseline recursive graph, the method further includes:
[0011] Acquire multiple first nonlinear signals, wherein the first nonlinear signals are nonlinear signals without anomalies;
[0012] Multiple recursive graphs are obtained by sequentially analyzing multiple first nonlinear signals using the recursive graph method.
[0013] The baseline recursive graph is determined based on the plurality of recursive graphs.
[0014] Optionally, determining the target statistical feature of the association recursion graph includes:
[0015] Determine a first preset threshold for the associated recursive graph;
[0016] The element whose value in the recursive graph is greater than the first preset threshold is taken as the first element;
[0017] The target statistical feature quantity of the association recursion graph is determined based on the first element.
[0018] Optionally, determining the target statistical feature of the association recursion graph includes:
[0019] Determine the second preset threshold for each element in the associated recursive graph;
[0020] The element values of each element in the recursive graph are compared sequentially with the second preset threshold corresponding to each element to obtain the comparison results;
[0021] The element whose comparison result is greater than the second preset threshold is taken as the second element;
[0022] The target statistical feature quantity of the association recursion graph is determined based on the second element.
[0023] Optionally, establishing a detection control chart for the abnormal signal and estimating the control limits of the detection control chart includes:
[0024] Acquire multiple second nonlinear signals, wherein the second nonlinear signals are nonlinear signals without anomalies;
[0025] A statistical feature set is determined, and a detection control chart based on the statistical feature set is established, wherein the statistical feature set is composed of statistical feature quantities of multiple second nonlinear signals;
[0026] Multiple sample sets are determined based on the statistical feature set, wherein each sample set has the same sample size, and the elements of each sample set are determined from the statistical feature set.
[0027] Determine multiple sample means for multiple sample sets, and determine the control limits of the detection control chart based on the multiple sample means.
[0028] Optionally, the control limits include an upper control limit and a lower control limit; wherein,
[0029] Determining the control limits of the detection control chart based on the mean values of the multiple samples includes:
[0030] Select the significance level;
[0031] The upper control limit and the lower control limit are determined based on the significance level and the mean of the plurality of samples.
[0032] Optionally, determining whether the target statistical feature is within the control limit and obtaining the determination result includes:
[0033] When it is determined that the target statistical feature quantity is less than the upper control limit and greater than the lower control limit, the determination result is that the target nonlinear signal is a normal signal;
[0034] When the target statistical feature is determined to be greater than or equal to the upper control limit or less than or equal to the lower control limit, the determination result is that the target nonlinear signal is an abnormal signal.
[0035] Furthermore, to achieve the above objectives, the present invention also proposes an anomaly monitoring device for nonlinear waveform signals, the anomaly monitoring device for nonlinear waveform signals comprising:
[0036] An acquisition module is used to acquire a target nonlinear signal and determine the target recurrence graph of the target nonlinear signal;
[0037] The determination module is used to determine the associated recursive graph based on the target recursive graph and the baseline recursive graph;
[0038] The determining module is further configured to determine the target statistical feature quantity of the association recursion graph;
[0039] A module is established to create a detection control chart for abnormal signals and to estimate the control limits of the detection control chart.
[0040] The judgment module is used to determine whether the target statistical feature quantity is within the control limit, obtain the judgment result, and monitor the target nonlinear signal based on the judgment result.
[0041] Furthermore, to achieve the above objectives, the present invention also proposes an anomaly monitoring device for nonlinear waveform signals. The anomaly monitoring device for nonlinear waveform signals includes: a memory, a processor, and an anomaly monitoring program for nonlinear waveform signals stored in the memory and executable on the processor. The anomaly monitoring program for nonlinear waveform signals is configured to implement the steps of the anomaly monitoring method for nonlinear waveform signals as described above.
[0042] Furthermore, to achieve the above objectives, the present invention also proposes a storage medium storing an anomaly monitoring program for nonlinear waveform signals, wherein when the anomaly monitoring program for nonlinear waveform signals is executed by a processor, the steps of the anomaly monitoring method for nonlinear waveform signals as described above are implemented.
[0043] The present invention proposes an anomaly monitoring method, apparatus, device, and storage medium for nonlinear waveform signals. This involves acquiring a target nonlinear signal and determining its target recursion graph; determining a correlation recursion graph based on the target and baseline recursion graphs; determining the target statistical feature quantity of the correlation recursion graph; establishing a detection control graph for the anomaly signal and estimating its control limits; and determining whether the target statistical feature quantity is within the control limits to obtain a judgment result, which is then used to monitor the target nonlinear signal. This method utilizes the spatiotemporal correlation characteristics of nonlinear signals to extract their recursive features. Monitoring these recursive features helps determine whether the target nonlinear signal is an anomaly. The recursive features not only effectively preserve the characteristic information of normal signals but also enhance the distinction between normal and anomaly signal feature values, thereby improving the accuracy of anomaly nonlinear signal monitoring, especially for detecting subtle local anomalies in nonlinear signals. Attached Figure Description
[0044] Figure 1 This is a schematic diagram of the structure of an anomaly monitoring device for nonlinear waveform signals in the hardware operating environment involved in the embodiments of the present invention;
[0045] Figure 2 This is a flowchart illustrating the first embodiment of the nonlinear waveform signal anomaly monitoring method of the present invention;
[0046] Figure 3 This is a schematic diagram of generating a reference recursive graph based on multiple recursive graphs in the first embodiment of the nonlinear waveform signal anomaly monitoring method of the present invention;
[0047] Figure 4 This is a mapping diagram between the nonlinear signal and the target recurrence graph in the first embodiment of the nonlinear waveform signal anomaly monitoring method of the present invention;
[0048] Figure 5 This is the detection result of the control graph of the abnormal nonlinear signal in the first embodiment of the abnormal monitoring method for nonlinear waveform signals of the present invention;
[0049] Figure 6 This is a flowchart illustrating the second embodiment of the nonlinear waveform signal anomaly monitoring method of the present invention;
[0050] Figure 7This is a schematic diagram illustrating the specific process of monitoring nonlinear signals in the second embodiment of the nonlinear waveform signal anomaly monitoring method of the present invention;
[0051] Figure 8 This is a structural block diagram of the first embodiment of the anomaly monitoring device for nonlinear waveform signals of the present invention.
[0052] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation
[0053] It should be understood that the specific embodiments described herein are for illustrative purposes only and are not intended to limit the scope of the invention.
[0054] Reference Figure 1 , Figure 1 This is a schematic diagram of the structure of an anomaly monitoring device for nonlinear waveform signals in the hardware operating environment involved in the embodiments of the present invention.
[0055] like Figure 1 As shown, the anomaly monitoring device for the nonlinear waveform signal may include: a processor 1001, such as a central processing unit (CPU), a communication bus 1002, a user interface 1003, a network interface 1004, and a memory 1005. The communication bus 1002 is used to enable communication between these components. The user interface 1003 may include a display screen or an input unit such as a keyboard; optionally, the user interface 1003 may also include a standard wired interface or a wireless interface. The network interface 1004 may optionally include a standard wired interface or a wireless interface (such as a Wireless-Fidelity (Wi-Fi) interface). The memory 1005 may be a high-speed random access memory (RAM) or a stable non-volatile memory (NVM), such as a disk drive. The memory 1005 may also optionally be a storage device independent of the aforementioned processor 1001.
[0056] Those skilled in the art will understand that Figure 1 The structure shown does not constitute a limitation on the anomaly monitoring device for nonlinear waveform signals, and may include more or fewer components than shown, or combine certain components, or have different component arrangements.
[0057] like Figure 1 As shown, the memory 1005, which serves as a storage medium, may include an operating system, a network communication module, a user interface module, and an anomaly monitoring program for nonlinear waveform signals.
[0058] exist Figure 1 In the nonlinear waveform signal anomaly monitoring device shown, the network interface 1004 is mainly used for data communication with the network server; the user interface 1003 is mainly used for data interaction with the user; the processor 1001 and memory 1005 in the nonlinear waveform signal anomaly monitoring device of the present invention can be set in the nonlinear waveform signal anomaly monitoring device, and the nonlinear waveform signal anomaly monitoring device calls the nonlinear waveform signal anomaly monitoring program stored in the memory 1005 through the processor 1001 and executes the nonlinear waveform signal anomaly monitoring method provided in the embodiment of the present invention.
[0059] Based on the above hardware structure, an embodiment of the nonlinear waveform signal anomaly monitoring method of the present invention is proposed.
[0060] Reference Figure 2 , Figure 2 This is a flowchart illustrating the first embodiment of an anomaly monitoring method for nonlinear waveform signals according to the present invention.
[0061] In this embodiment, the anomaly monitoring method for nonlinear waveform signals includes the following steps:
[0062] Step S10: Obtain the target nonlinear signal and determine the target recursion graph of the target nonlinear signal.
[0063] It should be noted that the executing entity in this embodiment can be a computing service device with data processing, network communication, and program execution functions, such as a mobile phone, tablet computer, or personal computer, or an electronic device or an anomaly monitoring device for nonlinear waveform signals capable of performing the above functions. The following description uses the anomaly monitoring device for nonlinear waveform signals as an example to illustrate this embodiment and the subsequent embodiments.
[0064] It should be noted that the target nonlinear signal can be one of the nonlinear waveform signals such as vibration, temperature, pressure, audio, electrocardiogram, pulse, and electromagnetic waves; the target recursion graph of the target nonlinear signal can be determined by the recursion graph method.
[0065] Step S20: Determine the associated recursion graph based on the target recursion graph and the baseline recursion graph.
[0066] In practical implementation, the baseline recursive graph can be determined using the recursive graph method. The recursive graph method is a good method for nonlinear signal analysis, which can convert one-dimensional nonlinear signals into two-dimensional matrix data, and can better extract the features of the signal.
[0067] Understandably, the correlation recurrence graph is used to represent the correlation between the target nonlinear signal and the reference recurrence graph.
[0068] In one embodiment, before determining the associated recursive graph based on the target recursive graph and the baseline recursive graph, the method further includes:
[0069] Acquire multiple first nonlinear signals, wherein the first nonlinear signals are nonlinear signals without anomalies;
[0070] Multiple recursive graphs are obtained by sequentially analyzing multiple first nonlinear signals using the recursive graph method.
[0071] The baseline recursive graph is determined based on the plurality of recursive graphs.
[0072] It should be noted that the recursive graph method transforms the nonlinear signal into a high-dimensional phase space by defining recursive graph parameters, and then obtains a two-dimensional recursive graph matrix. This two-dimensional recursive graph matrix can be regarded as a two-dimensional graph, thus obtaining an image containing the characteristic information of the nonlinear signal. Finally, by analyzing the two-dimensional image, the features of the recursive graph can be extracted, which can better characterize the changes of the nonlinear signal.
[0073] In specific implementations, such as Figure 3 As shown, the eight matrix diagrams on the left are the recursion graphs obtained from analyzing the first nonlinear signal, and the matrix diagram on the right is the baseline recursion graph determined based on multiple recursion graphs.
[0074] In the specific implementation, the baseline recursion graph is defined as follows:
[0075] Suppose {X1,X2,…,X...} n Given a set of nonlinear signals without anomalies, the recursion graph method is used to analyze this data, resulting in a recursion graph for each nonlinear signal as {R1, R2, ..., R...}. n}, at this point, define the two-dimensional matrix of the baseline recursion graph. for:
[0076]
[0077] Because the recursion matrix R corresponding to the recursion graph i It contains only two elements: 0 and 1. In the resulting two-dimensional matrix, all elements are greater than 0 and less than n. Therefore, a two-dimensional matrix can be obtained. Each element Values range from 0 to 1:
[0078]
[0079] It is understandable that in a two-dimensional matrix It can be used to represent the common characteristics of normal nonlinear signals, a two-dimensional matrix. The size of the element values has the following properties:
[0080] (1) If a two-dimensional matrix A certain element A value close to 1 means that in all recursive graphs R1, R2, ..., R n In most recursive graphs, there exist elements R1(i,j) = R2(i,j) = ... = R n (i,j)=1; when most of the recursion graphs contain R i When (i,j)=1, it indicates that in the original nonlinear signal, most signals have some similar signal characteristics at the corresponding positions. For example... Figure 3 As shown, if the element at the same position in all left-hand matrices has a value of 1, then the corresponding element at the same position in the right-hand matrix also has a value of 1. Therefore, when the element... The closer the value of is to 1, the stronger the two-dimensional matrix it forms. In recursive graphs, a greater proportion of the recursive graphs share certain common characteristics.
[0081] (2) If a two-dimensional matrix A certain element A value close to 0 means that in all recursive graphs R1, R2, ..., R n In most recursive graphs, there exist elements R1(i,j) = R2(i,j) = ... = R n (i,j)=0. When most recursive graphs contain R... i When (i,j) = 0, it indicates that in the original nonlinear signal, most signals do not exhibit a certain "recursive" characteristic at the corresponding positions. For example... Figure 3 As shown, if the element at the same position in all left-hand matrices has a value of 0, then the corresponding element at the same position in the right-hand matrix will also have a value of 0. Therefore, when the element... The closer the value of is to 0, the stronger the two-dimensional matrix is. In a recursive graph, a larger proportion of the recursive graphs do not have a certain feature.
[0082] In practical implementation, the association recursion graph is also equivalent to a matrix of the association recursion graph, and can be defined as a matrix R. diff The target recursion graph of the target nonlinear signal is R. new The matrix R of the recursive graph diff for:
[0083]
[0084] Understandably, based on the baseline recursion graph Based on the properties, the recurrence graph R can be deduced. diff It has the following properties:
[0085] (1) If element Rdiff If the value of (i,j) is close to 1, it indicates that R new (i,j) and The element values differ greatly, and the recursion graph R new The significant difference between this element's position and the baseline recursive graph indicates that the target nonlinear signal X is located at the corresponding signal segment. new Recursive diagram of the constitutive baseline The signals in that group exhibit different characteristics. And because they form the baseline recursive graph... All signals are normal. If R diff The large number of points with element values close to 1 allows us to infer the target nonlinear signal X. new It exhibits many characteristics that differ from normal signals. With R... diff The more points in the middle whose element value is close to 1, the stronger the target nonlinear signal X. new The more abnormal features a target nonlinear signal has, the more likely it is to be considered an abnormal signal.
[0086] (2) If element R diff If the value of (i, j) is close to 0, it indicates that R new (i, j) and The element values differ very little. This means that in the corresponding signal segment of the signal, the target nonlinear signal X... new Recursive diagram of the constitutive baseline The signals in that group share almost identical characteristics. This is because they constitute the baseline recursive graph. That set of signals is a normal signal; if R diff The large number of points with element values close to 0 allows us to infer the target nonlinear signal X. new It shares many characteristics with normal signals. With the correlation recursion graph R... diff The more points whose element values are close to 0, the more likely the nonlinear signal of the target can be considered a normal signal.
[0087] In specific implementations, such as Figure 4 As shown, the first nonlinear signal can be mapped to the recursive graph.
[0088] Step S30: Determine the target statistical feature quantity of the association recursion graph.
[0089] In practical implementation, a threshold T can be defined using the "Top-r" method based on the properties of the recurrence graph. threshold Compare the recurrence graph R diff The element values of each element in the table are related to the threshold value, if R... diff (i, j) ≥ T threshold If R diff (i, j) < Tthreshold If the selected element is not selected, then ignore that element; calculate the sum of squares of the selected element points to obtain a recursive graph R based on association. diff The statistical characteristic quantity (i.e., the target statistical characteristic quantity) is defined here as A. New , can be represented as:
[0090]
[0091] In the formula, I(x) is an indicator function; when x ≥ 0, I(x) = 1; when x < 0, I(x) = 0. r (i, j) is a predefined threshold parameter, which functions similarly to T. threshold .
[0092] It is understandable that the target statistical feature A can be inferred from the properties of the recurrence graph. New It is an association recursion graph R diff The cumulative effect of all points where all element values are greater than the threshold parameter reflects the accumulation of all abnormal features in the original signal that differ from the normal signal; the target statistical feature quantity A New The larger the value, the greater the difference between the original signal and the normal signal; by analyzing the target statistical characteristic quantity A... New Monitoring can detect whether there are any abnormalities in the nonlinear signal.
[0093] In one embodiment, determining the target statistical feature quantity of the association recursion graph includes:
[0094] Determine a first preset threshold for the associated recursive graph;
[0095] The element whose value in the recursive graph is greater than the first preset threshold is taken as the first element;
[0096] The target statistical feature quantity of the association recursion graph is determined based on the first element.
[0097] It is understandable that all elements in the recursive graph need to be compared with the first preset threshold to filter out the first preset threshold in the recursive graph; the method of using the first preset threshold to filter the first element refers to the hard threshold method; the first preset threshold can be set in advance.
[0098] In practical implementation, the hard thresholding method is used to determine the target statistical features as follows:
[0099]
[0100] In the formula, T constant This indicates the first preset threshold.
[0101] In this embodiment, the target statistical feature can be determined when the cause of the abnormal signal is unknown, using the hard threshold method.
[0102] In one embodiment, determining the target statistical feature quantity of the association recursion graph includes:
[0103] Determine the second preset threshold for each element in the associated recursive graph;
[0104] The element values of each element in the recursive graph are compared sequentially with the second preset threshold corresponding to each element to obtain the comparison results;
[0105] The element whose comparison result is greater than the second preset threshold is taken as the second element;
[0106] The target statistical feature quantity of the association recursion graph is determined based on the second element.
[0107] It should be noted that the method of setting different second preset thresholds for different elements based on their position and size can be called the soft threshold method; the second element is the element whose value in the recursive graph is greater than the second preset threshold.
[0108] In practical implementation, the soft thresholding method is used to determine the target statistical features as follows:
[0109]
[0110] In the formula, This represents the second preset threshold for the (i,j)th element in the recursive graph.
[0111] It is understandable that the second preset threshold is different for different elements in the recursive graph, and the second preset threshold of different elements is related to the different positions and sizes of the elements in the recursive graph.
[0112] It should be noted that when sufficient anomalous signal data is available, the anomalous characteristics of the anomalous signals can be determined, thereby enabling the determination of the correlation recursion graph R. diff The positions and sizes of all elements corresponding to the abnormal feature information in the graph R; based on the association recursion graph R diff The location distribution and abnormal feature information of the element points are used to set different second preset thresholds for each element point.
[0113] In this embodiment, using the soft thresholding method to determine the target statistical features of the target recursion graph ensures that the obtained target statistical features better reflect the abnormal information of the target nonlinear signal.
[0114] Step S40: Establish a detection control chart for abnormal signals and estimate the control limits of the detection control chart.
[0115] In practical implementation, it can be utilized Control chart method, establishing a statistical characteristic A based on the target New The detection control chart is used to estimate the control chart parameters (i.e., control limits), and then the target nonlinear signal is detected as an abnormal signal based on the control limits.
[0116] Step S50: Determine whether the target statistical characteristic quantity is within the control limit, obtain the determination result, and monitor the target nonlinear signal based on the determination result.
[0117] It is understandable that the target statistical characteristic can be interpreted as the difference (i.e., deviation value) between the target nonlinear signal and the baseline recurrence plot.
[0118] It should be noted that the judgment result can be either that the target statistical characteristic is not within the control limit or that the target statistical characteristic is within the control limit. When the judgment result is that the target statistical characteristic is not within the control limit, it indicates that the detected target nonlinear signal is an abnormal signal. When the judgment result is that the target statistical characteristic is within the control limit, it indicates that the detected target nonlinear signal is a normal signal.
[0119] In specific implementations, such as Figure 5 As shown, nonlinear signals can be monitored to determine whether they are abnormal signals by controlling the limits.
[0120] This embodiment acquires a target nonlinear signal and determines its target recursion graph; based on the target recursion graph and a baseline recursion graph, it determines an associated recursion graph; it determines the target statistical feature quantity of the associated recursion graph; it establishes a detection control graph for abnormal signals and estimates the control limits of the detection control graph; it determines whether the target statistical feature quantity is within the control limits, obtaining a judgment result, and monitors the target nonlinear signal based on the judgment result. Through this method, the spatiotemporal correlation characteristics of the nonlinear signal can be utilized to extract its recursive features. By monitoring these recursive features, it is possible to determine whether the target nonlinear signal is an abnormal signal. The recursive features not only effectively preserve the feature information of normal signals but also enhance the distinction between normal and abnormal signal feature values, thereby improving the detection accuracy of abnormal nonlinear signals, especially improving the detection effect when local subtle anomalies occur in the nonlinear signal.
[0121] refer to Figure 6 , Figure 6 This is a flowchart illustrating a second embodiment of the anomaly monitoring method for nonlinear waveform signals according to the present invention.
[0122] Based on the first embodiment described above, the anomaly monitoring method for nonlinear waveform signals in this embodiment, which involves establishing a detection control chart for the abnormal signal and estimating the control limits of the detection control chart, includes:
[0123] Step S401: Acquire multiple second nonlinear signals, wherein the second nonlinear signals are nonlinear signals without anomalies.
[0124] Step S402: Determine the statistical feature set and establish a detection control chart based on the statistical feature set, wherein the statistical feature set is composed of statistical feature quantities of multiple second nonlinear signals.
[0125] It should be noted that the statistical characteristics of the second nonlinear signal are determined in the same way as the target statistical characteristics of the target nonlinear signal. It is also necessary to first compare the second nonlinear signal with the reference recursion graph to determine the correlation recursion graph, and then use the "Top-r" method to define the threshold according to the properties of the correlation recursion graph, and then determine the statistical characteristics of the second nonlinear signal from the correlation recursion graph.
[0126] Step S403: Determine multiple sample sets based on the statistical feature set, wherein each sample set has the same sample size, and the elements of each sample set are determined from the statistical feature set.
[0127] It should be noted that a preset number can be set in advance, and then the preset number of sample sets can be determined based on the set of statistical features. It is also important to ensure that the sample size of each sample set is the same, and that the elements in each sample set are randomly determined from the set of statistical features.
[0128] Step S404: Determine multiple sample means for multiple sample sets, and determine the control limits of the detection control chart based on the multiple sample means.
[0129] In practice, the control limits of the detection control chart can be determined based on the mean of multiple samples using the control chart method.
[0130] In one embodiment, the control limits include an upper control limit and a lower control limit; wherein,
[0131] Determining the control limits of the detection control chart based on the mean values of the multiple samples includes:
[0132] Select the significance level;
[0133] The upper control limit and the lower control limit are determined based on the significance level and the mean of the plurality of samples.
[0134] In the specific implementation, the control limits of the detection control chart are estimated as follows:
[0135] (1) Assume there are k normal nonlinear signals (i.e., second nonlinear signals), and the statistical characteristic quantities of this group of signals can be determined as {A1, A2, ..., A by the recurrence plot method k} (i.e., the statistical characteristic set);
[0136] (2) Use the exhaustive method to generate a new sample set with a sample size of n c where n c is smaller than k;
[0137] (3) Calculate the mean value of the new sample set
[0138] (4) Repeat steps (2) and (3) B times to obtain B sample mean values
[0139] (5) Sort these B samples in ascending order to obtain
[0140]
[0141] (6) Select a significance level α, and according to P(S A <UCL)=1-α / 2 and P(S A >LCL)=α / 2, calculate the upper control limit UCL and the lower control limit LCL.
[0142] In specific implementation, as Figure 7 shown, a reference recurrence plot is determined through normal nonlinear signals, then the acquired new nonlinear signal is compared with the reference recurrence plot to determine a joint recurrence plot, then the "Top-r" method is used to define a threshold, and finally a target statistical characteristic quantity is determined according to the threshold and the joint recurrence plot, so that the monitoring result of the target statistical characteristic quantity can be determined through the exhaustive method.
[0143] In an embodiment, the step of determining whether the target statistical characteristic quantity is within the control limits to obtain a determination result includes:
[0144] When it is determined that the target statistical characteristic quantity is less than the upper control limit and greater than the lower control limit, the determination result is obtained that the target nonlinear signal is a normal signal;
[0145] When it is determined that the target statistical characteristic quantity is greater than or equal to the upper control limit or less than or equal to the lower control limit, the determination result is obtained that the target nonlinear signal is an abnormal signal.
[0146] It should be noted that the upper control limit is UCL, and the lower control limit is LCL.
[0147] In specific implementation, the target statistical characteristic A is processed through the control limits New performs monitoring, when A New ≥UCL or A New ≤LCL, the signal is an abnormal signal; if LCL<A New <UCL, the signal is a normal signal.
[0148] In this embodiment, a plurality of second nonlinear signals are acquired, where the second nonlinear signals are anomaly-free nonlinear signals; a statistical feature set is determined, and a detection control chart based on the statistical feature set is established, where the statistical feature set is composed of statistical feature quantities of the plurality of second nonlinear signals; a plurality of sample sets are determined according to the statistical feature set, where each sample set has the same sample size, and elements of each sample set are determined from the statistical feature set; a plurality of sample means of the plurality of sample sets are determined, and the control limits of the detection control chart are determined based on the plurality of sample means. Through the above manner, the control upper limit and the control lower limit can be estimated by establishing the control chart method based on the statistical feature quantities, thereby effectively improving the detection effect when local subtle anomalies occur in the nonlinear signal.
[0149] In addition, the embodiment of the present invention further proposes a storage medium, where the storage medium stores a program for monitoring anomalies of nonlinear waveform signals, and when the program for monitoring anomalies of nonlinear waveform signals is executed by a processor, the steps of the method for monitoring anomalies of nonlinear waveform signals as described above are implemented.
[0150] refer to Figure 8 , Figure 8 which is a structural block diagram of the first embodiment of the device for monitoring anomalies of nonlinear waveform signals according to the present invention.
[0151] as shown in Figure 8 , the device for monitoring anomalies of nonlinear waveform signals proposed by the embodiment of the present invention comprises:
[0152] an acquisition module 10, configured to acquire a target nonlinear signal and determine a target recurrence plot of the target nonlinear signal.
[0153] a determination module 20, configured to determine an associated recurrence plot according to the target recurrence plot and a reference recurrence plot.
[0154] the determination module 20 is further configured to determine a target statistical feature quantity of the associated recurrence plot.
[0155] an establishment module 30, configured to establish a detection control chart for abnormal signals, and estimate the control limits of the detection control chart.
[0156] a judgment module 40, configured to judge whether the target statistical feature quantity is within the control limits, obtain a judgment result, so as to monitor the target nonlinear signal according to the judgment result.
[0157] It should be understood that the above are merely illustrative examples and do not constitute any limitation on the technical solutions of the present invention. In specific applications, those skilled in the art can make settings as needed, and the present invention does not impose any restrictions on this.
[0158] This embodiment acquires a target nonlinear signal and determines its target recursion graph; based on the target recursion graph and a baseline recursion graph, it determines an associated recursion graph; it determines the target statistical feature quantity of the associated recursion graph; it establishes a detection control graph for abnormal signals and estimates the control limits of the detection control graph; it determines whether the target statistical feature quantity is within the control limits, obtaining a judgment result, and monitors the target nonlinear signal based on the judgment result. Through this method, the spatiotemporal correlation characteristics of the nonlinear signal can be utilized to extract its recursive features. By monitoring these recursive features, it is possible to determine whether the target nonlinear signal is an abnormal signal. The recursive features not only effectively preserve the feature information of normal signals but also enhance the distinction between normal and abnormal signal feature values, thereby improving the detection accuracy of abnormal nonlinear signals, especially improving the detection effect when local subtle anomalies occur in the nonlinear signal.
[0159] In one embodiment, the determining module 20 is further configured to:
[0160] Acquire multiple first nonlinear signals, wherein the first nonlinear signals are nonlinear signals without anomalies;
[0161] Multiple recursive graphs are obtained by sequentially analyzing multiple first nonlinear signals using the recursive graph method.
[0162] The baseline recursive graph is determined based on the plurality of recursive graphs.
[0163] In one embodiment, the determining module 20 is further configured to:
[0164] Determine a first preset threshold for the associated recursive graph;
[0165] The element whose value in the recursive graph is greater than the first preset threshold is taken as the first element;
[0166] The target statistical feature quantity of the association recursion graph is determined based on the first element.
[0167] In one embodiment, the determining module 20 is further configured to:
[0168] Determine the second preset threshold for each element in the associated recursive graph;
[0169] The element values of each element in the recursive graph are compared sequentially with the second preset threshold corresponding to each element to obtain the comparison results;
[0170] The element whose comparison result is greater than the second preset threshold is taken as the second element;
[0171] The target statistical feature quantity of the association recursion graph is determined based on the second element.
[0172] In one embodiment, the establishment module 30 is further configured to:
[0173] Acquire multiple second nonlinear signals, wherein the second nonlinear signals are nonlinear signals without anomalies;
[0174] A statistical feature set is determined, and a detection control chart based on the statistical feature set is established, wherein the statistical feature set is composed of statistical feature quantities of multiple second nonlinear signals;
[0175] Multiple sample sets are determined based on the statistical feature set, wherein each sample set has the same sample size, and the elements of each sample set are determined from the statistical feature set.
[0176] Determine multiple sample means for multiple sample sets, and determine the control limits of the detection control chart based on the multiple sample means.
[0177] In one embodiment, the control limits include an upper control limit and a lower control limit; wherein,
[0178] The establishment module 30 is also used for:
[0179] Select the significance level;
[0180] The upper control limit and the lower control limit are determined based on the significance level and the mean of the plurality of samples.
[0181] In one embodiment, the determining module 40 is further configured to:
[0182] When it is determined that the target statistical feature quantity is less than the upper control limit and greater than the lower control limit, the determination result is that the target nonlinear signal is a normal signal;
[0183] When the target statistical feature is determined to be greater than or equal to the upper control limit or less than or equal to the lower control limit, the determination result is that the target nonlinear signal is an abnormal signal.
[0184] It should be noted that the workflow described above is merely illustrative and does not limit the scope of protection of this invention. In practical applications, those skilled in the art can select some or all of the workflow to achieve the purpose of this embodiment according to actual needs, and no restrictions are imposed here.
[0185] In addition, for technical details not described in detail in this embodiment, please refer to the abnormal monitoring method of nonlinear waveform signals provided in any embodiment of the present invention, which will not be repeated here.
[0186] Furthermore, it should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or system that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or system. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or system that includes that element.
[0187] The sequence numbers of the above embodiments of the present invention are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.
[0188] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as read-only memory (ROM) / RAM, magnetic disk, optical disk) and includes several instructions to cause a terminal device (which may be a mobile phone, computer, server, or network device, etc.) to execute the methods described in the various embodiments of the present invention.
[0189] The above are merely preferred embodiments of the present invention and do not limit the scope of the patent. Any equivalent structural or procedural transformations made based on the description and drawings of the present invention, or direct or indirect applications in other related technical fields, are similarly included within the scope of patent protection of the present invention.
Claims
1. A method of anomaly monitoring of a non-linear waveform signal, characterized by, The anomaly monitoring method for the nonlinear waveform signal includes: Acquire the target nonlinear signal and determine the target recurrence graph of the target nonlinear signal; Based on the target recursion graph and the baseline recursion graph, determine the associated recursion graph; Determine the target statistical features of the recurrence graph; The step of determining the target statistical feature quantity of the association recursion graph includes: A threshold parameter is defined using the Top-r method. Element points in the association recursion graph whose element values are greater than the threshold parameter are filtered out. A second preset threshold is set based on the position and size of each element in the association recursion graph. The second preset threshold is different for different elements in the association recursion graph. The second preset threshold for different elements is related to the different positions and sizes of the elements in the association recursion graph. Based on the position distribution and abnormal feature information of the element points in the association recursion graph, a different second preset threshold is set for each element point. The sum of squares of the element values of the selected element points is calculated to obtain the target statistical feature of the correlation recursion graph. Establish a detection control chart for abnormal signals and estimate the control limits of the detection control chart, wherein the control limits include an upper control limit and a lower control limit; The steps for determining control limits include: Select the significance level; The upper control limit and the lower control limit are determined based on the significance level and the mean of multiple samples; The determination of the target statistical feature quantity of the association recursion graph further includes: The element values of each element in the recursive graph are compared sequentially with the second preset threshold corresponding to each element to obtain the comparison results; The element whose comparison result is greater than the second preset threshold is taken as the second element; The target statistical feature quantity of the association recursion graph is determined based on the second element; Determine whether the target statistical characteristic quantity is within the control limit, obtain the determination result, and monitor the target nonlinear signal based on the determination result; The step of determining whether the target statistical feature is within the control limit and obtaining the determination result includes: When it is determined that the target statistical feature quantity is less than the upper control limit and greater than the lower control limit, the determination result is that the target nonlinear signal is a normal signal; When it is determined that the target statistical feature quantity is greater than or equal to the upper control limit or less than or equal to the lower control limit, the determination result is that the target nonlinear signal is an abnormal signal; Before determining the associated recursive graph based on the target recursive graph and the baseline recursive graph, the method further includes: Acquire multiple first nonlinear signals, wherein the first nonlinear signals are nonlinear signals without anomalies; Multiple recursive graphs are obtained by sequentially analyzing multiple first nonlinear signals using the recursive graph method. The baseline recursive graph is determined based on the plurality of recursive graphs.
2. The method of claim 1, wherein, Determining the target statistical feature of the recurrence graph includes: Determine the second preset threshold for each element in the associated recursive graph; The element values of each element in the recursive graph are compared sequentially with the second preset threshold corresponding to each element to obtain the comparison results; The element whose comparison result is greater than the second preset threshold is taken as the second element; The target statistical feature quantity of the association recursion graph is determined based on the second element.
3. The method of claim 1, wherein, The establishment of a detection control chart for abnormal signals and the estimation of the control limits of the detection control chart include: Acquire multiple second nonlinear signals, wherein the second nonlinear signals are nonlinear signals without anomalies; A set of statistical features is determined, and a detection control chart based on the set of statistical features is established, wherein the set of statistical features consists of statistical features of multiple second nonlinear signals; Multiple sample sets are determined based on the statistical feature set, wherein each sample set has the same sample size, and the elements of each sample set are determined from the statistical feature set. Determine multiple sample means for multiple sample sets, and determine the control limits of the detection control chart based on the multiple sample means.
4. An anomaly monitoring device for nonlinear waveform signals, characterized in that, The anomaly monitoring device for the nonlinear waveform signal includes: An acquisition module is used to acquire a target nonlinear signal and determine the target recurrence graph of the target nonlinear signal; The determination module is used to determine the associated recursive graph based on the target recursive graph and the baseline recursive graph; The determining module is further configured to determine the target statistical feature quantity of the association recursion graph; A threshold parameter is defined using the Top-r method. Element points in the association recursion graph whose element values are greater than the threshold parameter are filtered out. A second preset threshold is set based on the position and size of each element in the association recursion graph. The second preset threshold is different for different elements in the association recursion graph. The second preset threshold for different elements is related to the different positions and sizes of the elements in the association recursion graph. Based on the position distribution and abnormal feature information of the element points in the association recursion graph, a different second preset threshold is set for each element point. The sum of squares of the element values of the selected element points is calculated to obtain the target statistical feature of the correlation recursion graph. A module is established to create a detection control chart for abnormal signals and to estimate the control limits of the detection control chart, wherein the control limits include an upper control limit and a lower control limit. The steps for determining control limits include: Select the significance level; The upper control limit and the lower control limit are determined based on the significance level and the mean of multiple samples; The determination of the target statistical feature quantity of the association recursion graph further includes: The element values of each element in the recursive graph are compared sequentially with the second preset threshold corresponding to each element to obtain the comparison results; The element whose comparison result is greater than the second preset threshold is taken as the second element; The target statistical feature quantity of the association recursion graph is determined based on the second element; The judgment module is used to determine whether the target statistical feature quantity is within the control limit, obtain a judgment result, and monitor the target nonlinear signal based on the judgment result; The judgment module is further configured to determine that the target nonlinear signal is a normal signal when the target statistical feature quantity is less than the control upper limit and greater than the control lower limit. When it is determined that the target statistical feature quantity is greater than or equal to the upper control limit or less than or equal to the lower control limit, the determination result is that the target nonlinear signal is an abnormal signal; The determining module is further configured to: Acquire multiple first nonlinear signals, wherein the first nonlinear signals are nonlinear signals without anomalies; Multiple recursive graphs are obtained by sequentially analyzing multiple first nonlinear signals using the recursive graph method. The baseline recursive graph is determined based on the plurality of recursive graphs.
5. An anomaly monitoring device for nonlinear waveform signals, characterized in that, The device includes: a memory, a processor, and an anomaly monitoring program for nonlinear waveform signals stored in the memory and executable on the processor, the anomaly monitoring program for nonlinear waveform signals being configured to implement the steps of the anomaly monitoring method for nonlinear waveform signals as described in any one of claims 1 to 3.
6. A storage medium, characterized in that, The storage medium stores an anomaly monitoring program for nonlinear waveform signals. When the processor executes the anomaly monitoring program for nonlinear waveform signals, it implements the steps of the anomaly monitoring method for nonlinear waveform signals as described in any one of claims 1 to 3.
Citation Information
Patent Citations
Industrial production process abnormal state detection method based on heterogeneous recursive pattern
CN107578166A
