An intrusion prevention diagnosis method based on vibration signal feature recognition

Abstract

The application discloses an anti-intrusion diagnosis method based on vibration signal feature recognition and belongs to the technical field of signal processing. The excavator excavation signal is recognized through three signal features, and the excavator excavation signal feature recognition is more sufficient. Meanwhile, the upper and lower limits of the signal parameters are simultaneously updated and limited by using a parameter updating method, the mutual influence between abnormal signal parameters and normal signal parameters can be prevented, the accuracy and efficiency of the excavator excavation signal recognition are obviously improved, the dependence on a large amount of data set training can be greatly reduced, and the practicability and efficiency of the anti-intrusion diagnosis method are improved.

Description

Technical Field

[0001] This invention relates to the field of signal processing technology, and in particular to an intrusion prevention diagnostic method based on vibration signal feature recognition. Background Technology

[0002] Currently, in the identification of excavator excavation signals, the advantages of relying on vibration signal analysis include the low cost of vibration sensors, ease of operation, and reliable identification results. Vibration analysis is an analytical method that obtains identification results by analyzing key components in the collected vibration signals. By calculating the characteristic components in the vibration signals and comparing the characteristic distributions between different signals, the signal to be identified can be determined. Therefore, how to accurately identify excavation signals is a problem that urgently needs to be solved.

[0003] In vibration signal identification, the process mainly involves two steps. The first step is to extract different features of the vibration signal. Current methods include calculating time-domain energy, frequency-domain energy, spectral kurtosis, and waveform distribution. More complex methods exist, such as wavelet decomposition, empirical mode decomposition, and their improvements. These methods primarily analyze the vibration signal by decomposing it into different components. The second step involves identifying and classifying the extracted feature components. This is mainly done through traditional classification methods and machine learning / deep learning methods. The latter two methods require a certain amount of data and are highly dependent on the quality of the data. This invention, starting from the perspective of traditional classification methods, calculates three signal feature components from the collected vibration data and uses a parameter update method to limit the upper and lower limits of the calculated feature components.

[0004] In vibration signal analysis and identification, there are few methods specifically for vibration signal analysis and identification of excavator excavation. Most existing methods rely on large datasets for training and verification. The method of this invention extracts several features that conform to the special characteristics of excavator excavation signals under the condition of traditional data classification, and classifies them. It judges whether the current signal meets the characteristics of excavator excavation by continuously updating the upper and lower limits of the threshold parameters, which has a certain degree of specificity.

[0005] This invention provides an intrusion prevention diagnostic method based on vibration signal feature recognition. It identifies excavator excavation signals through three signal features, resulting in more comprehensive excavator excavation signal feature recognition. Simultaneously, it utilizes a parameter update method to update the upper and lower limits of signal parameters at the same time, which can prevent mutual interference between abnormal and normal signal parameters, significantly improving the accuracy and efficiency of excavator excavation signal recognition. This method can also greatly reduce the dependence on training with large datasets, thereby improving the practicality and efficiency of the intrusion prevention diagnostic method of this invention. Summary of the Invention

[0006] To address the aforementioned problems, this invention provides an intrusion prevention diagnostic method based on vibration signal feature recognition. This invention significantly improves the accuracy of excavator excavation signal recognition through specific vibration signal feature extraction and parameter update methods, reduces reliance on large datasets, and enhances the practicality and efficiency of the method. It also reduces the dependence on large datasets for training, thus improving the practicality and efficiency of the intrusion prevention diagnostic method of this invention.

[0007] To achieve the above objectives, the technical solution adopted by the present invention is as follows.

[0008] An intrusion prevention diagnostic method based on vibration signal feature recognition includes the following steps: Step 1: Assume the vibration signal to be identified is x ( t For parameters that need to be updated STA / LTA The range of values ​​( R min , R max Frequency domain energy within the frequency domain range (1Hz, 100Hz) The range of values Frequency domain energy within the range of (100Hz, 300Hz) The range of values Perform parameter initialization; Step 2: Input the vibration signal to be identified x ( t ), and the vibration signal x ( t Frame segmentation is performed to obtain the frame signal for each frame. x l ( t ); Step 3: Calculate the frame signals from Step 2 respectively. x l ( t ) signal characteristics STA / LTA value R f ( l Frequency domain energy within the frequency domain range (1Hz, 100Hz) Frequency domain energy within the range of (100Hz, 300Hz) ; Step 4: Determine whether the signal characteristics meet the excavator intrusion signal characteristics, which are as follows: in, K , G 1.G2 They are respectively K , G 1. G 2 are signal characteristics STA / LTA value R f ( l The floating threshold coefficients for frequency domain energy within the range of (1Hz, 100Hz) and frequency domain energy within the range of (100Hz, 300Hz). & This means that both conditions must be met simultaneously; Step 5: If the excavator intrusion signal characteristics in Step 4 are met, an alarm will be triggered and it will be determined whether the abnormal parameter update conditions are met. Step 6: If the abnormal parameter update condition in Step 5 is not met, then do not update the parameter; if the abnormal parameter update condition is met, then proceed to Step 9. Step 7: If the signal characteristics do not meet the excavator intrusion signal characteristics in Step 4, do not alarm, and determine whether the normal parameter update conditions are met; Step 8: If the normal parameter update conditions in Step 7 are not met, then do not update the parameters; if the normal parameter update conditions are met, then update the corresponding lower limit of the parameters, and then proceed to Step 9. Step 9: Determine the vibration signal x ( t Check if all frame signals have finished updating; if not, calculate the vibration signal. x ( t The signal characteristics of the next frame signal are determined, and the process returns to step 3; if the process ends, then the frame signal... x 1 ( t If there is a parameter upper limit update, then the vibration signal... x ( t All lower limits of parameters for all framed signals are canceled from being updated, and the process returns to step 2 to re-enter the next segment of vibration signal to be identified, and then uses the previous segment of vibration signal as a reference. x ( t The final parameters are used as initialization parameters for identification and updating until all vibration signals to be identified have been identified and updated.

[0009] Preferably, the frame segmentation process is as follows: in, x ( t () represents the vibration signal to be identified. x l ( t ) represents the framing signal for each frame. t Indicates time,l Indicates the current frame. L Total number of frames w l ( t ) is the rectangular window function for the current frame. inc Indicates frame shift, Len Indicates the frame length.

[0010] Preferably, the signal features STA / LTA value R f ( l The calculation steps are as follows: Step S1: Assume the current frame is l Within the current frame, set the position of the short time window ( t s1 , t s2 ), length is T sta ; Set the position of the long window ( t l1 , t l2 ), length is T lta The positions of the long and short windows are distributed as follows: If the short-time window precedes the long-time window, then the range of the short-time window is ( l , l + T sta The long window range is ( l + T sta , l + T sta + T lta ); If the short-time window is in the middle of the long-time window, that is, the range of the short-time window is ( l + T lta / 2, l + T lta / 2+ T sta The long window range is ( l , l + T lta / 2)∪( l + T lta / 2+ T sta , l + Tsta + T lta ); If the short-time window follows the long-time window, then the range of the short-time window is ( l + T lta , l + T lta + T sta The long window range is ( l , l + T lta ); Step S2: Calculate the following formula: l Short window energy mean of frames STA ( l ), No. l Long-term window energy mean of frames LTA ( l ), No. l Short window energy mean of frames STA ( l ) and long-term window energy mean LTA ( l The ratio of ) R ( l ): Step S3: Calculate the ratio of the short time window to the long time window respectively. R 1( l The ratio of the short window to the middle of the long window. R 2( l The ratio of the short window to the long window. R 3( l The maximum value among the three is taken as the current frame. STA / LTA final value R f ( l ),Right now .

[0011] Preferably, the frequency domain energy in step 3 is within the range of (1Hz, 100Hz). The calculation steps are as follows: Step S1: For each frame's segmented signal x l ( t Perform a Fourier transform to convert the time-domain signal into a frequency-domain signal. X l ( ω ): in, T Represents the entire time domain. j It is an imaginary number. ω For frequency, Indicates Fourier transform; Step S2: For the frequency domain signal X l ( ω Find the frequency domain range. f low , f high The energy within, of which f low Indicates the lower limit of frequency. f high Indicates the upper limit of frequency; Step S3: Frequency domain energy within the frequency range (1Hz, 100Hz) for: in, f s Sampling frequency, Len Indicates the frame length.

[0012] Preferably, the frequency domain energy is within the range of (100Hz, 300Hz). for: in, f s The sampling frequency.

[0013] Preferably, the abnormal parameter update conditions are as follows: in, δ The range coefficient for parameter updates.

[0014] Preferably, the upper limit of the parameters corresponding to the update is as follows: in, α m This represents the upper limit smoothing factor for all parameters.

[0015] Preferably, the normal parameter update conditions are as follows: in, δ For the range coefficients of parameter updates, K , G 1. G 2 are signal characteristics STA / LTA value R f ( l Frequency domain energy within the frequency domain range (1Hz, 100Hz) Frequency domain energy within the range of (100Hz, 300Hz) The floating threshold coefficient.

[0016] Preferably, the lower limit of the parameter corresponding to the update is as follows: in, α R for R f ( l ) Parameter lower limit smoothing factor, α e1 for Lower bound smoothing factor of parameters α e2 for Lower limit smoothing factor for parameters.

[0017] By adopting the above technical solution, the present invention has the following beneficial effects.

[0018] (1) The present invention identifies excavator excavation signals through three signal features, which makes the excavator excavation signal feature identification more complete; at the same time, the parameter update method is used to update the upper and lower limits of the signal parameters simultaneously, which can prevent the mutual influence between abnormal signal parameters and normal signal parameters and improve the accuracy of excavator excavation signal identification.

[0019] (2) This invention divides the continuous vibration signal into multiple segments (frames). Through frame processing, the characteristics of the signal in different time periods can be observed more closely, improving processing efficiency and recognition accuracy; three key signal features are extracted from the vibration signal. STA / LTA Energy ratio, frequency domain energy in the range of 1-100Hz and frequency domain energy in the range of (100Hz, 300Hz) can quickly and accurately identify and distinguish different types of vibration signals, improving the identification accuracy of intrusive vibration signals from excavators and other equipment.

[0020] (3) This invention calculates the average value of the signal energy within the short-time window and the long-time window, and then determines its... STA / LTA Energy ratio accurately reflects signal changes over short time scales, helping to identify and detect abrupt signal changes, thus enabling precise identification of the signal characteristics of excavator excavation. Fourier transform converts the time-domain signal to the frequency domain and calculates the signal energy within a specific frequency band, aiding in the identification of signal characteristics within a specific frequency range, facilitating the identification of specific frequency characteristics of excavator excavation. The calculation and identification of these three signal characteristics contribute to improving the efficiency and accuracy of excavator vibration signal identification.

[0021] (4) By continuously updating the upper and lower limits of signal parameters, the present invention dynamically adjusts the recognition threshold to adapt to signal recognition under different conditions. It can adapt to signal changes in different environments, avoid misjudgment and missed judgment caused by fixed thresholds, and improve the adaptability and accuracy of the recognition system.

[0022] (5) This invention significantly improves the recognition accuracy of excavator excavation signals by using specific vibration signal feature extraction and parameter update methods, reduces the dependence on large datasets, and improves the practicality and efficiency of the method. It can reduce the dependence on large datasets for training, improve the practicality and efficiency of the intrusion prevention diagnosis method of this invention, and can be widely used in the fields of pipeline, cable, road and other anti-excavation monitoring; industrial manufacturing, construction, transportation and environmental monitoring. Attached Figure Description

[0023] The following provides a detailed discussion of the manufacture and application of preferred embodiments of the present invention. However, it should be understood that the present invention provides many applicable inventive concepts that can be embodied in various specific environments. The specific embodiments discussed are merely illustrative of specific ways of manufacturing and using the present invention and do not limit the scope of the invention. For those skilled in the art, other drawings can be obtained from these drawings without any creative effort.

[0024] Figure 1 This is a flowchart of the present invention.

[0025] Figure 2 This is a comparison diagram of the different signal characteristics of vibration signals from non-excavator and excavator machines according to the present invention.

[0026] Figure 3 This is a table comparing the false recognition probabilities of the present invention under different feature count conditions. Detailed Implementation

[0027] The following provides a detailed discussion of the manufacture and application of preferred embodiments of the present invention. However, it should be understood that the present invention provides many applicable inventive concepts that can be embodied in various specific environments. The specific embodiments discussed are merely illustrative of specific ways of manufacturing and using the invention and do not limit the scope of the invention.

[0028] This invention identifies excavator excavation signals using three features, avoiding the problem of insufficient feature recognition; at the same time, it uses a parameter update method to simultaneously update the upper and lower limits of signal parameters, which can prevent mutual interference between abnormal and normal signal parameters and improve the accuracy of excavator excavation signal recognition.

[0029] like Figure 1 The method for intrusion prevention diagnosis based on vibration signal feature recognition, as shown, includes the following steps.

[0030] Step 1: Parameter initialization, assuming the vibration signal to be identified is x ( t ), the range of values ​​for the parameter STA / LTA that needs to be updated ( R min , R max Frequency domain energy within the frequency domain range (1Hz, 100Hz) The range of values Frequency domain energy within the range of (100Hz, 300Hz) The range of values Perform parameter initialization.

[0031] Step 2: Input vibration signal x ( t For multiple equal-length vibration signals to be identified collected by the sensor x ( t ), and transmit the first vibration signal in sequence. x ( t The signal is input into the intrusion prevention diagnostic method of the present invention for subsequent signal processing, identification and alarm operations.

[0032] Step 3: Framing the vibration signal to obtain the framed signal. x l ( t Input the vibration signal to be identified. x ( t ), and the vibration signal x ( t Frame segmentation is performed to obtain the frame signal for each frame. x l ( t ).

[0033] The formula for the frame segmentation process is as follows: in, x ( t () represents the vibration signal to be identified. x l ( t ) represents the framing signal for each frame. t Indicates time, l Indicates the current frame. L Total number of frames w l ( t ) is the rectangular window function for the current frame. inc Indicates frame shift, Len Indicates the frame length.

[0034] Step 4: Calculate the signal characteristics of the framed signals, and calculate the framed signals from Step 2 respectively. x l ( t ) signal characteristics STA / LTA value R f ( l Frequency domain energy within the frequency domain range (1Hz, 100Hz) Frequency domain energy within the range of (100Hz, 300Hz) .

[0035] Among them, the framing signal in step 4 x l ( t ) signal characteristics STA / LTA value R f ( l The calculation steps are as follows.

[0036] Step S1: Assume the current frame is l Within the current frame, set the position of the short time window ( t s1 , t s2 ), length is T sta ; Set the position of the long window ( t l1 , t l2 ), length is T lta The positions of the long and short windows are distributed as follows: If the short-time window precedes the long-time window, then the range of the short-time window is ( l , l + T sta The long window range is ( l + T sta , l + T sta + T lta ).

[0037] If the short-time window is in the middle of the long-time window, that is, the range of the short-time window is ( l + T lta / 2, l + T lta / 2+ Tsta The long window range is ( l , l + T lta / 2)∪( l + T lta / 2+ T sta , l + T sta + T lta ).

[0038] If the short-time window follows the long-time window, then the range of the short-time window is ( l + T lta , l + T lta + T sta The long window range is ( l , l + T lta ).

[0039] Step S2: Calculate the following formula: l Short window energy mean of frames STA ( l ), No. l Long-term window energy mean of frames LTA ( l ), No. l Short window energy mean of frames STA ( l ) and long-term window energy mean LTA ( l The ratio of ) R ( l ):

[0040] Step S3: Calculate the ratio of the short time window to the long time window respectively. R 1( l The ratio of the short window to the middle of the long window. R 2( l The ratio of the short window to the long window. R 3( l The maximum value among the three is taken as the current frame. STA / LTA final value R f ( l ),Right now .

[0041] In step 4, the frequency domain energy within the range of (1Hz, 100Hz) Frequency domain energy within the range of (100Hz, 300Hz) The calculation steps are as follows.

[0042] Step S1: For each frame's segmented signal x l ( t Perform a Fourier transform to convert the time-domain signal into a frequency-domain signal. X l ( ω ): in, T Represents the entire time domain. j It is an imaginary number. ω For frequency, This represents the Fourier transform.

[0043] Step S2: For the frequency domain signal X l ( ω Find the frequency domain range. f low , f high The energy within, of which f low Indicates the lower limit of frequency. f high This indicates the upper limit of the frequency.

[0044] Step S3: Frequency domain energy within the frequency range (1Hz, 100Hz) for: in, f s Sampling frequency, Len Indicates the frame length.

[0045] Step S4: Frequency domain energy within the frequency range of (100Hz, 300Hz) for: in, f s The sampling frequency.

[0046] In step 4, it is determined whether the signal characteristics meet the excavator intrusion signal characteristics, which are as follows: in,K , G 1. G2 They are respectively K , G 1. G 2 are signal characteristics STA / LTA value R f ( l The floating threshold coefficients for frequency domain energy within the range of (1Hz, 100Hz) and frequency domain energy within the range of (100Hz, 300Hz). & This means that both conditions are met simultaneously.

[0047] Step 5: If the excavator intrusion signal characteristics described in Step 4 are met, an alarm will be triggered and it will be determined whether the abnormal parameter update conditions are met.

[0048] The abnormal parameter update conditions in step 5 are as follows: in, δ The range coefficient for parameter updates.

[0049] Step 6: If the abnormal parameter update conditions in Step 5 are not met, then do not update the parameters; if the abnormal parameter update conditions in Step 5 are met, then update the corresponding parameter upper limit; then proceed to Step 9.

[0050] Once the abnormal parameter update condition is met, the steps in step 6 to update the corresponding parameter upper limit are as follows: in, α m This represents the upper limit smoothing factor for all parameters.

[0051] Step 7: If the signal characteristics do not meet the excavator intrusion signal characteristics in Step 5, do not alarm, and determine whether the normal parameter update conditions are met.

[0052] The normal parameter update conditions in step 7 are as follows: in, δ For the range coefficients of parameter updates, K , G 1. G 2 are signal characteristics STA / LTA value R f ( l Frequency domain energy within the frequency domain range (1Hz, 100Hz) Frequency domain energy within the range of (100Hz, 300Hz) The floating threshold coefficient.

[0053] Step 8: If the normal parameter update conditions in Step 7 are not met, then do not update the parameters; if the normal parameter update conditions are met, then update the corresponding lower limit of the parameters; then proceed to Step 9.

[0054] The steps for updating the corresponding lower limit parameter in step 8 are as follows: in, α R for R f ( l ) Parameter lower limit smoothing factor, α e1 for Lower bound smoothing factor of parameters α e2 for Lower limit smoothing factor for parameters.

[0055] Step 9: Determine the vibration signal x ( t Check if all frame signals have finished updating; if not, calculate the vibration signal. x ( t The signal characteristics of the next frame signal are determined, and the process returns to step 3; if the process ends, then the frame signal... x l ( t If there is a parameter upper limit update, then the vibration signal... x ( t All lower limits of parameters for all framed signals are canceled from being updated, and the process returns to step 2 to re-enter the next segment of vibration signal to be identified, and then uses the previous segment of vibration signal as a reference. x ( t The final parameters are used as initialization parameters for identification and updating until all vibration signals to be identified have been identified and updated.

[0056] This invention divides continuous vibration signals into multiple segments (frames). Through frame-based processing, the characteristics of the signal at different time periods can be observed in greater detail, improving processing efficiency and recognition accuracy. Three key signal features are extracted from the vibration signal. STA / LTA Energy ratio, frequency domain energy in the range of 1-100Hz and frequency domain energy in the range of (100Hz, 300Hz) can quickly and accurately identify and distinguish different types of vibration signals, improving the identification accuracy of intrusive vibration signals from excavators and other equipment.

[0057] To illustrate the performance of the method proposed in this invention in detail, specific comparative examples are provided below for verification. This embodiment verifies the performance of identifying excavator signals using real data collected by sensors. A total of 421 data points were used, with each data segment having a sampling time of 5 seconds and a sampling rate of 8kHz. To prevent signal characteristics from being corrupted, a rectangular window was used in the frame processing, with a 25% overlap.

[0058] This invention employs three features, namely, framing signals. x l ( t ) signal characteristics STA / LTA value R f ( l Frequency domain energy within the frequency domain range (1Hz, 100Hz) Frequency domain energy within the range of (100Hz, 300Hz) This embodiment compares the different characteristics of vibration signals from non-excavator and excavator machines, and compares the results under different numbers of characteristics. Probability of algorithm misalignment detection P : in, N f For the number of misidentifications, N This represents the total number of signals.

[0059] from Figure 2 As can be seen, the frequency energy distribution of non-excavator vibration signals is relatively stable in the frequency domain range (1Hz, 100Hz) and frequency domain range (100Hz, 300Hz), with values ​​fluctuating within a certain range and no large peaks. In contrast, excavator excavation signals exhibit particularly large energy fluctuations in both the time domain and various frequency domains, with rapid increases and decreases in a short period of time, exhibiting clear and significant signal characteristics. This invention utilizes this characteristic to identify excavator signal features, enabling the identification and early warning of excavator signal characteristics.

[0060] from Figure 3 It can be seen that when using these three features for identification, the number of signal features is also highly correlated with the identification effect; as the number of features increases, the probability of false identification decreases. It can be seen that the combination of the three feature recognition methods used in this invention significantly improves the accuracy of excavator signal feature recognition.

[0061] This invention calculates the average signal energy within short and long time windows and then determines its... STA / LTAEnergy ratio accurately reflects signal changes over short time scales, aiding in the identification and detection of abrupt signal changes, thus enabling precise identification of the signal characteristics of excavator excavation. Fourier transform converts the time-domain signal to the frequency domain and calculates the signal energy within a specific frequency band, facilitating the identification of signal characteristics within a specific frequency range and the identification of specific frequency characteristics of excavator excavation. The calculation and identification of these three signal characteristics contribute to improving the accuracy of excavator vibration signal identification. By continuously updating the upper and lower limits of signal parameters and dynamically adjusting the identification threshold to adapt to signal recognition under different conditions, the system can adapt to signal changes in various environments, avoiding misjudgments and omissions caused by fixed thresholds, and improving the adaptability and accuracy of the identification system.

[0062] This invention significantly improves the accuracy of excavator excavation signal recognition through specific vibration signal feature extraction and parameter update methods, reduces reliance on large datasets, and enhances the practicality and efficiency of the method. It can reduce the dependence on large datasets for training, improve the practicality and efficiency of the intrusion prevention diagnosis method of this invention, and can be widely used in anti-excavation monitoring of pipelines, cables, roads, etc.; industrial manufacturing, construction, transportation and environmental monitoring and other fields.

[0063] Although the specification has provided a detailed description, it should be understood that various changes, substitutions, and modifications can be made without departing from the spirit and scope of the invention as defined by the appended claims. Furthermore, the specific embodiments described are not intended to limit the scope of the invention, and those skilled in the art will readily understand based on this invention that existing or future-developed processes, machines, manufactures, compositions of matter, means, methods, or steps can perform substantially the same functions or achieve substantially the same results as the embodiments of the invention. Therefore, the appended claims are intended to include such processes, machines, manufactures, compositions of matter, means, methods, or steps within their scope.

Claims

1. An intrusion prevention diagnostic method based on vibration signal feature recognition, characterized in that: Includes the following steps: Step 1: Assume the vibration signal to be identified is x ( t For parameters that need to be updated STA / LTA The range of values ​​( R min , R max Frequency domain energy within the frequency domain range (1Hz, 100Hz) The range of values Frequency domain energy within the range of (100Hz, 300Hz) The range of values Perform parameter initialization; Step 2: Input the vibration signal to be identified x ( t ), and the vibration signal x ( t Frame segmentation is performed to obtain the frame signal for each frame. x l ( t ); Step 3: Calculate the frame signals from Step 2 respectively. x l ( t ) signal characteristics STA / LTA value R f ( l Frequency domain energy within the frequency domain range (1Hz, 100Hz) Frequency domain energy within the range of (100Hz, 300Hz) ; Step 4: Determine whether the signal characteristics meet the excavator intrusion signal characteristics, which are as follows: in, K , G 1. G2 Signal characteristics STA / LTA value R f ( l The floating threshold coefficients for frequency domain energy within the range of (1Hz, 100Hz) and frequency domain energy within the range of (100Hz, 300Hz). & This means that both conditions must be met simultaneously; Step 5: If the excavator intrusion signal characteristics in Step 4 are met, an alarm will be triggered and it will be determined whether the abnormal parameter update conditions are met. Step 6: If the abnormal parameter update condition in Step 5 is not met, then do not update the parameter; if the abnormal parameter update condition is met, then proceed to Step 9. Step 7: If the signal characteristics do not meet the excavator intrusion signal characteristics in Step 4, do not alarm, and determine whether the normal parameter update conditions are met; Step 8: If the normal parameter update conditions in Step 7 are not met, then do not update the parameters; if the normal parameter update conditions are met, then update the corresponding lower limit of the parameters, and then proceed to Step 9. Step 9: Determine the vibration signal x ( t Check if all frame signals have finished updating; if not, calculate the vibration signal. x ( t The signal characteristics of the next frame signal are determined, and the process returns to step 3; if the process ends, then the frame signal... x l ( t If there is a parameter upper limit update, then the vibration signal... x ( t All lower limits of parameters for all framed signals are canceled from being updated, and the process returns to step 2 to re-enter the next segment of vibration signal to be identified, and then uses the previous segment of vibration signal as a reference. x ( t The final parameters are used as initialization parameters for identification and updating until all vibration signals to be identified have been identified and updated.

2. The intrusion prevention diagnostic method based on vibration signal feature recognition as described in claim 1, characterized in that: The frame segmentation process in step 2 is as follows: in, x ( t () represents the vibration signal to be identified. x l ( t ) represents the framing signal for each frame. t Indicates time, l Indicates the current frame. L Total number of frames w l ( t ) is the rectangular window function for the current frame. inc Indicates frame shift, Len Indicates the frame length.

3. The intrusion prevention diagnostic method based on vibration signal feature recognition as described in claim 1, characterized in that: Signal characteristics in step 3 STA / LTA value R f ( l The calculation steps are as follows: Step S1: Assume the current frame is l Within the current frame, set the position of the short time window ( t s1 , t s2 ), length is T sta ; Set the position of the long window ( t l1 , t l2 ), length is T lta The positions of the long and short windows are distributed as follows: If the short-time window precedes the long-time window, then the range of the short-time window is ( l , l + T sta The long window range is ( l + T sta , l + T sta + T lta ); If the short-time window is in the middle of the long-time window, that is, the range of the short-time window is ( l + T lta / 2, l + T lta / 2+ T sta The long window range is ( l , l + T lta / 2)∪( l + T lta / 2+ T sta , l + T sta + T lta ); If the short-time window follows the long-time window, then the range of the short-time window is ( l + T lta , l + T lta + T sta The long window range is ( l , l + T lta ); Step S2: Calculate the following formula: l Short window energy mean of frames STA ( l ), No. l Long-term window energy mean of frames LTA ( l ), No. l Short window energy mean of frames STA ( l ) and long-term window energy mean LTA ( l The ratio of ) R ( l ): Step S3: Calculate the ratio of the short time window to the long time window respectively. R 1( l The ratio of the short window to the middle of the long window. R 2( l The ratio of the short window to the long window. R 3( l The maximum value among the three is taken as the current frame. STA / LTA final value R f ( l ),Right now 。 4. The intrusion prevention diagnostic method based on vibration signal feature recognition as described in claim 1, characterized in that: In step 3, the frequency domain energy within the range of (1Hz, 100Hz) The calculation steps are as follows: Step S1: For each frame's segmented signal x l ( t Perform a Fourier transform to convert the time-domain signal into a frequency-domain signal. X l ( ω ): in, T Represents the entire time domain. j It is an imaginary number. ω For frequency, Indicates Fourier transform; Step S2: For the frequency domain signal X l ( ω Find the frequency domain range. f low , f high The energy within, of which f low Indicates the lower limit of frequency. f high Indicates the upper limit of frequency; Step S3: Frequency domain energy within the frequency range (1Hz, 100Hz) for: in, f s Sampling frequency, Len Indicates the frame length.

5. The intrusion prevention diagnostic method based on vibration signal feature recognition as described in claim 4, characterized in that: The frequency domain energy within the frequency range (100Hz, 300Hz) for: in, f s The sampling frequency.

6. The intrusion prevention diagnostic method based on vibration signal feature recognition as described in claim 1, characterized in that: The abnormal parameter update conditions in step 5 are as follows: in, δ The range coefficient for parameter updates.

7. The intrusion prevention diagnostic method based on vibration signal feature recognition as described in claim 1, characterized in that: In step 6, the corresponding parameter upper limit is updated as follows: in, α m This represents the upper limit smoothing factor for all parameters.

8. The intrusion prevention diagnostic method based on vibration signal feature recognition as described in claim 1, characterized in that: The normal parameter update conditions in step 7 are as follows: in, δ For the range coefficients of parameter updates, K , G 1. G 2 are signal characteristics STA / LTA value R f ( l Frequency domain energy within the frequency domain range (1Hz, 100Hz) Frequency domain energy within the range of (100Hz, 300Hz) The floating threshold coefficient.

9. The intrusion prevention diagnostic method based on vibration signal feature recognition as described in claim 1, characterized in that: In step 8, the corresponding lower limit of the parameter is updated as follows: in, α R for R f ( l ) Parameter lower limit smoothing factor, α e1 for Lower bound smoothing factor of parameters α e2 for Lower limit smoothing factor for parameters.

Images