A non-woven packaging bag welding dynamic detection device and detection method

Abstract

The present invention discloses a non-woven packaging bag welding dynamic detection device and detection method, which belongs to the technical field of non-woven packaging bag welding detection. The present invention collects vibration signals in the welding process, and divides the vibration signals into multiple segments, extracts the main amplitude and main frequency at each moment for each vibration sub-signal, constructs the main amplitude sequence and the main frequency sequence, eliminates the interference of environmental noise and other mechanical vibrations, calculates the neighborhood gap coefficient for each element in the main amplitude sequence and the main frequency sequence, constructs the amplitude change characteristic sequence and the frequency change characteristic sequence, reflects the instantaneous change of amplitude and frequency through the neighborhood gap coefficient, and is convenient for identifying sudden changes or abnormal fluctuations that may occur in the welding process, and obtains the abnormality value of the welding process based on the compensation of the main amplitude offset value and the main frequency offset value according to the amplitude change characteristic sequence and the frequency change characteristic sequence, thereby improving the detection accuracy of the non-woven packaging bag welding process.

Description

Technical Field

[0001] The invention relates to the technical field of non-woven packaging bag welding detection, and in particular to a non-woven packaging bag welding dynamic detection device and detection method. Background Art

[0002] With the enhancement of environmental awareness and the promotion of the concept of sustainable development, non-woven packaging bags have gradually become a widely used packaging material in the market due to their light, durable and recyclable characteristics. However, in the production process of non-woven packaging bags, the welding quality directly affects the performance and safety of the product. Traditional welding quality detection methods mostly rely on manual inspection or simple visual inspection, which is not only inefficient, but also easily affected by human factors, resulting in inaccurate detection results. With the development of automation technology and sensor technology, dynamic detection methods based on vibration signals have gradually attracted attention. By real-time monitoring of the acoustic wave signals generated during welding, abnormal conditions in the welding process can be effectively identified, thereby improving the welding quality. In practical applications, vibration signals are often interfered by environmental noise and other mechanical vibrations. This noise affects the clarity of the signal, resulting in a decrease in the accuracy of feature extraction and analysis, thereby affecting the evaluation of welding quality. Summary of the invention

[0003] In view of the above-mentioned deficiencies in the prior art, a non-woven packaging bag welding dynamic detection device and detection method provided by the present invention solves the problem of low detection accuracy in the existing non-woven packaging bag welding detection.

[0004] In order to achieve the above-mentioned invention object, the technical solution adopted by the present invention is: a non-woven packaging bag welding dynamic detection device, comprising: a collection unit, a signal segmentation unit, an amplitude-frequency feature extraction unit, an amplitude-frequency variation feature extraction unit, an amplitude-frequency offset extraction unit and an abnormality detection unit;

[0005] The acquisition unit is used to collect the vibration signal of the acoustic waveguide rod in the ultrasonic welding machine during the welding process of the non-woven packaging bag;

[0006] The signal segmentation unit is used for dividing the vibration signal into a plurality of vibration sub-signals;

[0007] The amplitude-frequency feature extraction unit is used to extract the main amplitude and main frequency of the vibration sub-signal at each moment in the frequency domain to obtain the main amplitude sequence and the main frequency sequence;

[0008] The amplitude-frequency variation feature extraction unit is used to calculate the neighborhood gap coefficient for each element in the main amplitude sequence and the main frequency sequence, and construct the amplitude variation feature sequence and the frequency variation feature sequence;

[0009] The amplitude-frequency offset extraction unit is used to calculate the main amplitude offset value and the main frequency offset value according to the mean value of each main amplitude sequence and the mean value of each main frequency sequence;

[0010] The abnormality detection unit is used to process the amplitude variation characteristic sequence and the frequency variation characteristic sequence, and obtain the abnormality degree value of the welding process based on the compensation of the main amplitude offset value and the main frequency offset value.

[0011] Furthermore, the amplitude-frequency feature extraction unit includes: a time-frequency transformation module, a main amplitude extraction module, a main frequency extraction module, a main amplitude sequence construction module and a main frequency sequence construction module;

[0012] The time-frequency transformation module is used to perform short-time Fourier transformation on each vibration sub-signal to obtain the spectrum information at multiple moments;

[0013] The main amplitude extraction module is used to calculate the average value of each amplitude in the spectrum information at each moment to obtain the amplitude mean value, extract the amplitude greater than the amplitude mean value in the spectrum information at one moment as the characteristic amplitude, and add the characteristic amplitudes of the spectrum information at one moment to obtain the main amplitude;

[0014] The main frequency extraction module is used to extract the frequency corresponding to the characteristic amplitude in the spectrum information at a moment as the characteristic frequency, and add the characteristic frequencies of the spectrum information at a moment to obtain the main frequency;

[0015] The main amplitude sequence building module is used to build the main amplitude sequence by taking the main amplitude at each moment as an element;

[0016] The main frequency sequence construction module is used to construct the main frequency sequence by taking the main frequency at each moment as an element.

[0017] Furthermore, the amplitude-frequency variation feature extraction unit includes: an amplitude neighborhood gap coefficient calculation module, a frequency neighborhood gap coefficient calculation module, an amplitude variation feature sequence construction module, and a frequency variation feature sequence construction module;

[0018] The amplitude neighborhood gap coefficient calculation module is used to calculate the neighborhood gap coefficient of each moment element in the main amplitude sequence;

[0019] The frequency neighborhood gap coefficient calculation module is used to calculate the neighborhood gap coefficient of each moment element in the main frequency sequence;

[0020] The amplitude variation characteristic sequence construction module is used to extract the maximum value of all neighborhood gap coefficients corresponding to a main amplitude sequence, obtain the amplitude variation characteristic value of a section of vibration sub-signal, and use the amplitude variation characteristic value of each vibration sub-signal as an element to construct the amplitude variation characteristic sequence;

[0021] The frequency change feature sequence construction module is used to extract the maximum value of all neighborhood gap coefficients corresponding to a main frequency sequence, obtain the frequency change feature value of a section of oscillator signal, and use the frequency change feature value of each oscillator signal as an element to construct the frequency change feature sequence.

[0022] Furthermore, the formulas for calculating the neighborhood gap coefficient in the amplitude neighborhood gap coefficient calculation module or the frequency neighborhood gap coefficient calculation module are: , where ε is the neighborhood gap coefficient, x t is the t-th element in the sequence, x t+1 is the element at time t+1 in the sequence, x t-1 is the element at time t-1 in the sequence, e is a natural constant, || is the absolute value, and t is the number of the time.

[0023] Further, the amplitude-frequency offset extraction unit includes: an amplitude mean value calculation module, a frequency mean value calculation module, an amplitude offset calculation module and a frequency offset calculation module;

[0024] The amplitude mean calculation module is used to calculate the mean of the main amplitude sequence to obtain the main amplitude mean;

[0025] The frequency mean calculation module is used to calculate the mean of the main frequency sequence to obtain the main frequency mean;

[0026] The amplitude offset calculation module is used to calculate the main amplitude offset value according to the difference between the main amplitude mean value of each vibrator signal and the stored main amplitude mean value;

[0027] The frequency offset calculation module is used to calculate the main frequency offset value according to the difference between the main frequency mean value of each vibrator signal and the stored main frequency mean value.

[0028] Furthermore, the expression of the amplitude offset calculation module is: , where γ F is the main amplitude offset value, arctan is the inverse tangent function, F i is the main amplitude mean of the i-th oscillator signal, F s To store the main amplitude mean, i is the number of the oscillator signal, and N is the number of oscillator signals;

[0029] The expression of the frequency offset calculation module is: , where γ f is the main frequency offset value, f i is the main frequency mean of the i-th oscillator signal, f s To store the main frequency mean.

[0030] Further, the abnormality detection unit includes: a first processing module, a first compensation module, a second processing module, a second compensation module and an abnormality degree value calculation module;

[0031] The first processing module is used to process the amplitude change characteristic sequence to obtain the main amplitude change abnormal value;

[0032] The second processing module is used to process the frequency change characteristic sequence to obtain the main frequency change abnormal value;

[0033] The first compensation module is used to compensate the main amplitude variation abnormal value by using the main amplitude offset value to obtain the main amplitude variation compensation value;

[0034] The second compensation module is used to compensate the main frequency variation abnormal value by using the main frequency offset value to obtain the main frequency variation compensation value;

[0035] The abnormality degree value calculation module is used to add the main amplitude variation compensation value and the main frequency variation compensation value to obtain the abnormality degree value of the welding process.

[0036] Furthermore, the expression of the first processing module is: , where y F is the abnormal value of the main amplitude change, F v,n is the nth element in the amplitude variation characteristic sequence, F v,th is the amplitude change threshold, M1 is the number of elements in the amplitude change characteristic sequence, e is a natural constant, and n is the number of the element;

[0037] The expression of the second processing module is: , where y f is the main frequency change abnormal value, f v,n is the nth element in the frequency change characteristic sequence, f v,th is the frequency change threshold, and M2 is the number of elements in the frequency change feature sequence.

[0038] Furthermore, the expression of the first compensation module is: ,in, is the compensation value for the main amplitude variation, γ F is the main amplitude offset value;

[0039] The expression of the second compensation module is: ,in, is the main frequency variation compensation value, γ f Main frequency offset value.

[0040] A non-woven packaging bag welding dynamic detection method comprises the following steps:

[0041] S1, collecting the vibration signal of the acoustic waveguide rod in the ultrasonic welding machine during the welding process of the non-woven packaging bag;

[0042] S2, dividing the vibration signal into multiple vibration sub-signals;

[0043] S3, extracting the main amplitude and main frequency of the oscillator signal at each moment in the frequency domain to obtain a main amplitude sequence and a main frequency sequence;

[0044] S4, calculating the neighborhood gap coefficient for each element in the main amplitude sequence and the main frequency sequence, and constructing the amplitude change feature sequence and the frequency change feature sequence;

[0045] S5. Calculate the main amplitude offset value and the main frequency offset value according to the mean value of each main amplitude sequence and the mean value of each main frequency sequence;

[0046] S6. Process the amplitude variation characteristic sequence and the frequency variation characteristic sequence, and obtain the abnormality degree value of the welding process based on the compensation of the main amplitude offset value and the main frequency offset value.

[0047] The beneficial effects of the present invention are as follows: the present invention collects vibration signals during the welding process, and divides the vibration signals into multiple segments, extracts the main amplitude and main frequency at each moment for each vibration sub-signal, constructs a main amplitude sequence and a main frequency sequence, eliminates the interference of environmental noise and other mechanical vibrations, calculates the neighborhood gap coefficient for each element in the main amplitude sequence and the main frequency sequence, constructs an amplitude change characteristic sequence and a frequency change characteristic sequence, and reflects the instantaneous changes of the amplitude and frequency through the neighborhood gap coefficient, which is convenient for identifying sudden changes or abnormal fluctuations that may occur during the welding process, and obtains the abnormality value of the welding process based on the compensation of the main amplitude offset value and the main frequency offset value according to the amplitude change characteristic sequence and the frequency change characteristic sequence, thereby improving the detection accuracy of the welding process of non-woven packaging bags. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 The system block diagram of a non-woven packaging bag welding dynamic detection device. DETAILED DESCRIPTION

[0049] The specific implementation modes of the present invention are described below so that those skilled in the art can understand the present invention. However, it should be clear that the present invention is not limited to the scope of the specific implementation modes. For those of ordinary skill in the art, as long as various changes are within the spirit and scope of the present invention as defined and determined by the attached claims, these changes are obvious, and all inventions and creations utilizing the concept of the present invention are protected.

[0050] Embodiment 1, as Figure 1 As shown, a non-woven packaging bag welding dynamic detection device includes: a collection unit, a signal segmentation unit, an amplitude-frequency feature extraction unit, an amplitude-frequency variation feature extraction unit, an amplitude-frequency offset extraction unit and an abnormality detection unit;

[0051] The acquisition unit is used to collect the vibration signal of the acoustic waveguide rod in the ultrasonic welding machine during the welding process of the non-woven packaging bag;

[0052] The signal segmentation unit is used for dividing the vibration signal into a plurality of vibration sub-signals;

[0053] The amplitude-frequency feature extraction unit is used to extract the main amplitude and main frequency of the vibration sub-signal at each moment in the frequency domain to obtain the main amplitude sequence and the main frequency sequence;

[0054] The amplitude-frequency variation feature extraction unit is used to calculate the neighborhood gap coefficient for each element in the main amplitude sequence and the main frequency sequence, and construct the amplitude variation feature sequence and the frequency variation feature sequence;

[0055] The amplitude-frequency offset extraction unit is used to calculate the main amplitude offset value and the main frequency offset value according to the mean value of each main amplitude sequence and the mean value of each main frequency sequence;

[0056] The abnormality detection unit is used to process the amplitude variation characteristic sequence and the frequency variation characteristic sequence, and obtain the abnormality degree value of the welding process based on the compensation of the main amplitude offset value and the main frequency offset value.

[0057] The sonic waveguide is used to conduct ultrasonic energy and is connected to the welding tool to ensure that the ultrasonic wave is effectively transmitted to the welding area.

[0058] The amplitude-frequency feature extraction unit includes: a time-frequency transformation module, a main amplitude extraction module, a main frequency extraction module, a main amplitude sequence construction module and a main frequency sequence construction module;

[0059] The time-frequency transformation module is used to perform short-time Fourier transformation on each vibration sub-signal to obtain the spectrum information at multiple moments;

[0060] The main amplitude extraction module is used to calculate the average value of each amplitude in the spectrum information at each moment to obtain the amplitude mean value, extract the amplitude greater than the amplitude mean value in the spectrum information at one moment as the characteristic amplitude, and add the characteristic amplitudes of the spectrum information at one moment to obtain the main amplitude;

[0061] The main frequency extraction module is used to extract the frequency corresponding to the characteristic amplitude in the spectrum information at a moment as the characteristic frequency, and add the characteristic frequencies of the spectrum information at a moment to obtain the main frequency;

[0062] The main amplitude sequence building module is used to build the main amplitude sequence by taking the main amplitude at each moment as an element;

[0063] The main frequency sequence construction module is used to construct the main frequency sequence by taking the main frequency at each moment as an element.

[0064] The present invention extracts the amplitude greater than the amplitude mean value of the spectrum information at a moment as the characteristic amplitude, and extracts the characteristic frequency corresponding to the characteristic amplitude, which can effectively identify the significant amplitude and significant frequency, reflecting the real vibration signal applied during the welding process, and then adds the characteristic amplitude of the spectrum information at a moment to obtain the main amplitude, reflecting the welding strength at that moment, and adds the characteristic frequencies of the spectrum information at a moment to obtain the main frequency, reflecting the main mode of the vibration signal applied at that moment. Constructing the main amplitude sequence and the main frequency sequence reflects the main amplitude and main frequency of a section of the vibration sub-signal.

[0065] The amplitude-frequency variation feature extraction unit includes: an amplitude neighborhood gap coefficient calculation module, a frequency neighborhood gap coefficient calculation module, an amplitude variation feature sequence construction module and a frequency variation feature sequence construction module;

[0066] The amplitude neighborhood gap coefficient calculation module is used to calculate the neighborhood gap coefficient of each moment element in the main amplitude sequence;

[0067] The frequency neighborhood gap coefficient calculation module is used to calculate the neighborhood gap coefficient of each moment element in the main frequency sequence;

[0068] The amplitude variation characteristic sequence construction module is used to extract the maximum value of all neighborhood gap coefficients corresponding to a main amplitude sequence, obtain the amplitude variation characteristic value of a section of vibration sub-signal, and use the amplitude variation characteristic value of each vibration sub-signal as an element to construct the amplitude variation characteristic sequence;

[0069] The frequency change feature sequence construction module is used to extract the maximum value of all neighborhood gap coefficients corresponding to a main frequency sequence, obtain the frequency change feature value of a section of oscillator signal, and use the frequency change feature value of each oscillator signal as an element to construct the frequency change feature sequence.

[0070] The present invention can accurately capture the dynamic changes of amplitude and frequency during welding by calculating the neighborhood gap coefficient of each moment element in the main amplitude sequence and the main frequency sequence. This accurate capture helps to timely identify abnormal conditions during welding. The maximum value of all neighborhood gap coefficients corresponding to a main amplitude sequence is extracted, and the maximum value reflects the drastic change of the main amplitude of the vibrator signal of this segment; the maximum value of all neighborhood gap coefficients corresponding to a main amplitude sequence is extracted, and the maximum value reflects the drastic change of the main frequency of the vibrator signal of this segment. The more drastic the changes in the main amplitude and main frequency are, the worse the stability during welding.

[0071] The formulas for calculating the neighborhood gap coefficient in the amplitude neighborhood gap coefficient calculation module or the frequency neighborhood gap coefficient calculation module are: , where ε is the neighborhood gap coefficient, x t is the t-th element in the sequence, x t+1is the element at time t+1 in the sequence, x t-1 is the element at time t-1 in the sequence, e is a natural constant, || is the absolute value, and t is the number of the time.

[0072] The present invention is based on x t and adjacent elements x t+1 and x t-1 By comparison, the absolute value of the difference is extracted to reflect the instantaneous fluctuation of the main amplitude or main frequency, and the exponential function is used to enhance the difference, so that there are significant differences between the gap coefficients of different neighborhoods.

[0073] The amplitude-frequency offset extraction unit includes: an amplitude mean value calculation module, a frequency mean value calculation module, an amplitude offset calculation module and a frequency offset calculation module;

[0074] The amplitude mean calculation module is used to calculate the mean of the main amplitude sequence to obtain the main amplitude mean;

[0075] The frequency mean calculation module is used to calculate the mean of the main frequency sequence to obtain the main frequency mean;

[0076] The amplitude offset calculation module is used to calculate the main amplitude offset value according to the difference between the main amplitude mean value of each vibrator signal and the stored main amplitude mean value;

[0077] The frequency offset calculation module is used to calculate the main frequency offset value according to the difference between the main frequency mean value of each vibrator signal and the stored main frequency mean value.

[0078] The expression of the amplitude offset calculation module is: , where γ F is the main amplitude offset value, arctan is the inverse tangent function, F i is the main amplitude mean of the i-th oscillator signal, F s To store the main amplitude mean, i is the number of the oscillator signal, and N is the number of oscillator signals;

[0079] The expression of the frequency offset calculation module is: , where γ f is the main frequency offset value, f i is the main frequency mean of the i-th oscillator signal, f s To store the main frequency mean.

[0080] The present invention calculates the mean of the main amplitude sequence to obtain the main amplitude mean of a section of vibration sub-signal, calculates the mean of the main frequency sequence to obtain the main frequency mean of a section of vibration sub-signal, obtains the main amplitude offset value according to the difference between the main amplitude mean of each section of vibration sub-signal and the stored main amplitude mean, and obtains the main frequency offset value according to the difference between the main frequency mean of each section of vibration sub-signal and the stored main frequency mean, reflecting the deviation of the main amplitude and main frequency from the normal value.

[0081] In the present invention, the stored main amplitude mean and main frequency mean are based on the main amplitude mean and main frequency mean obtained by analyzing the vibration signal of the acoustic waveguide in the ultrasonic welding machine under normal working conditions. The stored main amplitude mean and main frequency mean represent the vibration characteristics under standard operating conditions and provide a reliable benchmark for subsequent deviation detection.

[0082] The abnormality detection unit includes: a first processing module, a first compensation module, a second processing module, a second compensation module and an abnormality degree value calculation module;

[0083] The first processing module is used to process the amplitude change characteristic sequence to obtain the main amplitude change abnormal value;

[0084] The second processing module is used to process the frequency change characteristic sequence to obtain the main frequency change abnormal value;

[0085] The first compensation module is used to compensate the main amplitude variation abnormal value by using the main amplitude offset value to obtain the main amplitude variation compensation value;

[0086] The second compensation module is used to compensate the main frequency variation abnormal value by using the main frequency offset value to obtain the main frequency variation compensation value;

[0087] The abnormality degree value calculation module is used to add the main amplitude variation compensation value and the main frequency variation compensation value to obtain the abnormality degree value of the welding process.

[0088] The present invention first processes the amplitude variation characteristic sequence and the frequency variation characteristic sequence, evaluates the stability of the main amplitude and main frequency changes, and then further improves the detection accuracy of the abnormality degree value of the welding process based on the compensation of the main amplitude offset value or the main frequency offset value.

[0089] The expression of the first processing module is: , where y F is the abnormal value of the main amplitude change, F v,n is the nth element in the amplitude variation characteristic sequence, F v,th is the amplitude change threshold, M1 is the number of elements in the amplitude change characteristic sequence, e is a natural constant, and n is the number of the element;

[0090] The expression of the second processing module is: , where y f is the main frequency change abnormal value, f v,n is the nth element in the frequency change characteristic sequence, f v,th is the frequency change threshold, and M2 is the number of elements in the frequency change feature sequence.

[0091] The amplitude variation threshold is a threshold set for the amplitude variation characteristic value, and the frequency variation threshold is a threshold set for the frequency variation characteristic value. The amplitude variation threshold and the frequency variation threshold are the amplitude variation characteristic values ​​and the frequency variation characteristic values ​​obtained by analyzing the vibration signal of the acoustic waveguide rod in the ultrasonic welding machine under normal working conditions.

[0092] The present invention is in F v,n With F v,th , f v,n With f v,th The larger the gap, the larger the main amplitude change anomaly and the main frequency change anomaly.

[0093] The expression of the first compensation module is: ,in, is the compensation value for the main amplitude variation, γ F is the main amplitude offset value;

[0094] The expression of the second compensation module is: ,in, is the main frequency variation compensation value, γ f Main frequency offset value.

[0095] Embodiment 2, a non-woven packaging bag welding dynamic detection method, comprising the following steps:

[0096] S1, collecting the vibration signal of the acoustic waveguide rod in the ultrasonic welding machine during the welding process of the non-woven packaging bag;

[0097] S2, dividing the vibration signal into multiple vibration sub-signals;

[0098] S3, extracting the main amplitude and main frequency of the oscillator signal at each moment in the frequency domain to obtain a main amplitude sequence and a main frequency sequence;

[0099] S4, calculating the neighborhood gap coefficient for each element in the main amplitude sequence and the main frequency sequence, and constructing the amplitude change feature sequence and the frequency change feature sequence;

[0100] S5. Calculate the main amplitude offset value and the main frequency offset value according to the mean value of each main amplitude sequence and the mean value of each main frequency sequence;

[0101] S6. Process the amplitude variation characteristic sequence and the frequency variation characteristic sequence, and obtain the abnormality degree value of the welding process based on the compensation of the main amplitude offset value and the main frequency offset value.

[0102] The specific implementation process of Example 2 is the same as that of Example 1.

[0103] The present invention collects vibration signals during the welding process, and divides the vibration signals into multiple segments, extracts the main amplitude and main frequency at each moment for each vibration sub-signal, constructs a main amplitude sequence and a main frequency sequence, eliminates interference from environmental noise and other mechanical vibrations, calculates a neighborhood gap coefficient for each element in the main amplitude sequence and the main frequency sequence, constructs an amplitude change feature sequence and a frequency change feature sequence, and reflects the instantaneous changes of the amplitude and frequency through the neighborhood gap coefficient, so as to facilitate the identification of sudden changes or abnormal fluctuations that may occur during the welding process, and obtains the abnormality value of the welding process based on the compensation of the main amplitude offset value and the main frequency offset value according to the amplitude change feature sequence and the frequency change feature sequence, thereby improving the detection accuracy of the welding process of the non-woven packaging bag.

[0104] The above are only preferred embodiments of the present invention and are not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A non-woven packaging bag welding dynamic detection device, characterized in that: include: Acquisition unit, signal segmentation unit, amplitude-frequency feature extraction unit, amplitude-frequency variation feature extraction unit, amplitude-frequency offset extraction unit and anomaly detection unit; The collection unit is used to collect the vibration signal of the acoustic waveguide rod in the ultrasonic welding machine during the welding process of the non-woven packaging bag; The signal segmentation unit is used to divide the vibration signal into multiple vibration sub-signals; The amplitude-frequency feature extraction unit is used to extract the main amplitude and main frequency of the vibration sub-signal at each moment in the frequency domain to obtain a main amplitude sequence and a main frequency sequence; The amplitude-frequency variation feature extraction unit is used to calculate the neighborhood gap coefficient for each element in the main amplitude sequence and the main frequency sequence, and construct the amplitude variation feature sequence and the frequency variation feature sequence, wherein the formula for calculating the neighborhood gap coefficient is: , where ε is the neighborhood gap coefficient, x t is the t-th element in the sequence, x t+1 is the element at time t+1 in the sequence, x t-1 is the element at the t-1th moment in the sequence, e is a natural constant, || is the absolute value, and t is the number of the moment; The amplitude-frequency offset extraction unit is used to calculate the main amplitude offset value and the main frequency offset value according to the mean value of each main amplitude sequence and the mean value of each main frequency sequence: , , γ F is the main amplitude offset value, arctan is the inverse tangent function, F i is the main amplitude mean of the i-th oscillator signal, F s is to store the main amplitude mean, i is the number of the oscillator signal, N is the number of oscillator signals, γ f is the main frequency offset value, f i is the main frequency mean of the i-th oscillator signal, f s To store the main frequency mean; The abnormality detection unit is used to process the amplitude change characteristic sequence to obtain the main amplitude change abnormal value; process the frequency change characteristic sequence to obtain the main frequency change abnormal value; use the main amplitude offset value to compensate the main amplitude change abnormal value to obtain the main amplitude change compensation value; use the main frequency offset value to compensate the main frequency change abnormal value to obtain the main frequency change compensation value; add the main amplitude change compensation value and the main frequency change compensation value to obtain the abnormality degree value of the welding process.

2. The non-woven packaging bag welding dynamic detection device according to claim 1 is characterized in that: The amplitude-frequency feature extraction unit includes: a time-frequency transformation module, a main amplitude extraction module, a main frequency extraction module, a main amplitude sequence construction module and a main frequency sequence construction module; The time-frequency transformation module is used to perform short-time Fourier transformation on each vibration sub-signal to obtain frequency spectrum information at multiple moments; The main amplitude extraction module is used to calculate the average value of each amplitude in the spectrum information at each moment to obtain the amplitude mean value, extract the amplitude greater than the amplitude mean value in the spectrum information at one moment as the characteristic amplitude, and add the characteristic amplitudes of the spectrum information at one moment to obtain the main amplitude; The main frequency extraction module is used to extract the frequency corresponding to the characteristic amplitude in the spectrum information at a moment as the characteristic frequency, and add the characteristic frequencies of the spectrum information at a moment to obtain the main frequency; The main amplitude sequence construction module is used to construct a main amplitude sequence by taking the main amplitude at each moment as an element; The main frequency sequence construction module is used to construct a main frequency sequence by taking the main frequency at each moment as an element.

3. The non-woven packaging bag welding dynamic detection device according to claim 1 is characterized in that: The amplitude-frequency variation feature extraction unit comprises: an amplitude neighborhood gap coefficient calculation module, a frequency neighborhood gap coefficient calculation module, an amplitude variation feature sequence construction module and a frequency variation feature sequence construction module; The amplitude neighborhood gap coefficient calculation module is used to calculate the neighborhood gap coefficient of each moment element in the main amplitude sequence; The frequency neighborhood gap coefficient calculation module is used to calculate the neighborhood gap coefficient of each moment element in the main frequency sequence; The amplitude variation characteristic sequence construction module is used to extract the maximum value of all neighborhood gap coefficients corresponding to a main amplitude sequence, obtain the amplitude variation characteristic value of a section of vibrator signal, and use the amplitude variation characteristic value of each vibrator signal as an element to construct the amplitude variation characteristic sequence; The frequency change feature sequence construction module is used to extract the maximum value of all neighborhood gap coefficients corresponding to a main frequency sequence, obtain the frequency change feature value of a section of vibrator signal, and use the frequency change feature value of each vibrator signal as an element to construct a frequency change feature sequence.

4. The non-woven packaging bag welding dynamic detection device according to claim 1 is characterized in that: The amplitude-frequency offset extraction unit includes: an amplitude mean value calculation module, a frequency mean value calculation module, an amplitude offset calculation module and a frequency offset calculation module; The amplitude mean value calculation module is used to calculate the mean value of the main amplitude sequence to obtain the main amplitude mean value; The frequency mean calculation module is used to calculate the mean of the main frequency sequence to obtain the main frequency mean; The amplitude offset calculation module is used to calculate the main amplitude offset value according to the difference between the main amplitude mean value of each vibration sub-signal and the stored main amplitude mean value; The frequency offset calculation module is used to calculate the main frequency offset value according to the difference between the main frequency mean value of each vibrator signal and the stored main frequency mean value.

5. The non-woven packaging bag welding dynamic detection device according to claim 1 is characterized in that: The abnormality detection unit includes: a first processing module, a first compensation module, a second processing module, a second compensation module and an abnormality degree value calculation module; The first processing module is used to process the amplitude variation characteristic sequence to obtain the main amplitude variation abnormal value; The second processing module is used to process the frequency change feature sequence to obtain the main frequency change abnormal value; The first compensation module is used to compensate the main amplitude variation abnormal value by using the main amplitude offset value to obtain the main amplitude variation compensation value; The second compensation module is used to compensate the main frequency variation abnormal value by using the main frequency offset value to obtain the main frequency variation compensation value; The abnormality degree value calculation module is used to add the main amplitude variation compensation value and the main frequency variation compensation value to obtain the abnormality degree value of the welding process.

6. The non-woven packaging bag welding dynamic detection device according to claim 5 is characterized in that: The expression of the first processing module is: , where y F is the abnormal value of the main amplitude change, F v,n is the nth element in the amplitude variation characteristic sequence, F v,th is the amplitude change threshold, M1 is the number of elements in the amplitude change characteristic sequence, e is a natural constant, n is the number of the element, and | | is the absolute value; The expression of the second processing module is: , where y f is the main frequency change abnormal value, f v,n is the nth element in the frequency change characteristic sequence, f v,th is the frequency change threshold, and M2 is the number of elements in the frequency change feature sequence.

7. The non-woven packaging bag welding dynamic detection device according to claim 6 is characterized in that: The expression of the first compensation module is: ,in, is the compensation value for the main amplitude variation, γ F is the main amplitude offset value; The expression of the second compensation module is: ,in, is the main frequency variation compensation value, γ f Main frequency offset value.

8. A non-woven packaging bag welding dynamic detection method, implemented based on the non-woven packaging bag welding dynamic detection device according to any one of claims 1 to 7, characterized in that: The following steps are involved: S1, collecting the vibration signal of the acoustic waveguide rod in the ultrasonic welding machine during the welding process of the non-woven packaging bag; S2, dividing the vibration signal into multiple vibration sub-signals; S3, extracting the main amplitude and main frequency of the oscillator signal at each moment in the frequency domain to obtain a main amplitude sequence and a main frequency sequence; S4, calculating the neighborhood gap coefficient for each element in the main amplitude sequence and the main frequency sequence, and constructing the amplitude change feature sequence and the frequency change feature sequence; S5. Calculate the main amplitude offset value and the main frequency offset value according to the mean value of each main amplitude sequence and the mean value of each main frequency sequence; S6. Process the amplitude variation characteristic sequence and the frequency variation characteristic sequence, and obtain the abnormality degree value of the welding process based on the compensation of the main amplitude offset value and the main frequency offset value.

Images