A paper tube glue semi-finished product stirring tank with a weighing function

Abstract

The application relates to the technical field of high-viscosity material mixing equipment, and discloses a paper tube glue semi-finished product stirring tank with a weighing function, which comprises a tank body, a double-curved-surface rotor precisely matched with a W-shaped groove at the bottom of the tank body, and a planetary differential speed driving structure. The driving rotor rotates at differential speed while revolving, forms a moving linear contact high-pressure grinding area, and eliminates mixing dead angles in the structure in combination with a rotor spiral flow guide groove. The application also comprises a measurement system based on modal analysis, which effectively separates a resonant frequency related to material quality by applying a high-frequency perturbation excitation and collecting vibration responses, and realizes dynamic weighing free from stirring vibration interference by substituting the resonant frequency into a preset mass and frequency mapping model. In addition, the system compares a multimodal fingerprint with a normal working condition fingerprint library, calculates a fingerprint distortion degree, can on-line diagnose abnormalities such as loosening of stirring parts or material caking, and realizes equipment state early warning.

Description

Technical Field

[0001] This invention relates to the field of high-viscosity material mixing equipment, specifically to a semi-finished product mixing tank for paper tube adhesive with weighing function. Background Technology

[0002] In the preparation of high-viscosity, high-solids-content materials (such as paper tube adhesive), achieving uniform mixing and efficient dispersion of materials, as well as precise control of the production process, are key to ensuring the quality of the final product.

[0003] However, existing mixing technologies face challenges when handling such materials. While conventional mixers can move the main material, the high viscosity and poor flowability of the material easily lead to stable laminar flow at the bottom of the tank, near the walls, and in the backflow zone of the mixing blades. This results in some materials not being adequately sheared and exchanged, creating mixing dead zones and uneven dispersion, which directly affects the consistency of product performance.

[0004] Furthermore, real-time monitoring of the material quality inside the tank is crucial for achieving automated control of the production process and precise formulation execution. Currently, the common approach is to place the entire mixing tank on top of an external weighing sensor (such as a piezoelectric weighing module). However, during operation, the powerful motor and the violent tumbling and shearing motion of viscous materials generate strong low-frequency mechanical vibrations. These vibrations are directly transmitted to the weighing sensor, creating noise interference. This results in an extremely low signal-to-noise ratio, drastic fluctuations in readings, and severe distortion. Consequently, accurate dynamic weighing under mixing conditions is impossible, limiting production efficiency and control precision.

[0005] Meanwhile, the reliability of mixing equipment under long-term high-load operation is also an important consideration in production. Existing technologies typically lack real-time online sensing methods for the structural health status of the equipment itself. Potential abnormal operating conditions such as loose connections of mixing components, abnormal material agglomeration or adhesion on the tank walls, etc., can often only be detected through periodic shutdown maintenance or after a failure occurs. There is a lack of technology that can analyze changes in the inherent characteristics of the equipment during normal operation to provide early warnings of potential risks, making it impossible to achieve predictive maintenance and ensure the continuity and safety of production. Summary of the Invention

[0006] To address the shortcomings of existing technologies, this invention provides a semi-finished mixing tank for paper tube adhesive with a weighing function. This solves the problems commonly found in existing mixing tanks for high-viscosity materials, such as uneven dispersion due to mixing dead zones, inability to accurately and dynamically weigh due to severe vibration interference, and inability to provide online early warning of abnormal operating conditions due to a lack of sensing means.

[0007] To achieve the above objectives, the present invention provides the following technical solution: a semi-finished mixing tank for paper tube adhesive with weighing function, comprising a tank body, a top cover provided on the top of the tank body, a motor installed on the top of the top cover, a rotating frame fixedly connected to the output end of the motor, a stirring mechanism provided in the middle of the rotating frame, a planetary gear mechanism provided at the bottom of the top cover, the planetary gear mechanism being used to control the stirring trajectory of the stirring mechanism, and a controller installed on the side of the tank body, the controller having a built-in weighing system;

[0008] The stirring mechanism includes a stirring frame rotatably connected to the middle of the rotating frame. A hyperboloid rotor is fixedly connected to the bottom of the stirring frame. A W-shaped groove is formed inside the tank, and the shape of the W-shaped groove matches the shape of the bottom of the hyperboloid rotor. A spiral guide groove is formed on the outer side of the hyperboloid rotor. The planetary gear mechanism includes an internal gear ring fixedly connected to the bottom of the top cover. A gear one is fixedly connected to the top of the rotating frame, and a gear two is fixedly connected to the top of the stirring frame. The internal gear ring meshes with the gear two, and the gear one meshes with the gear two.

[0009] Preferably, a scraper is fixedly connected to the side of the stirring rack, a cutting blade is fixedly connected to the middle of the stirring rack, and a piezoelectric ceramic actuator and an acceleration sensor are provided on the outside of the tank.

[0010] The weighing system provided by the present invention includes: at least one perturbation excitation unit configured to apply an excitation signal to a tank; a multidimensional vibration response sensor array configured to acquire the vibration response signal of the tank to the excitation signal; and a modal analysis and diagnostic decision unit connected to the perturbation excitation unit and the multidimensional vibration response sensor array.

[0011] The modal analysis and diagnostic decision unit is configured as follows:

[0012] The control perturbation excitation unit sends an excitation signal;

[0013] Process the vibration response signal collected by the multidimensional vibration response sensor array and extract at least one resonant frequency that characterizes the current state of the tank.

[0014] Based on the preset mass-frequency mapping model and the extracted resonant frequency, the mass of the material inside the tank is calculated.

[0015] Preferably, the excitation signal is a high-frequency excitation signal, and the vibration generated by the tank under stirring conditions is low-frequency vibration noise. The modal analysis and diagnostic decision unit is specifically configured as follows: filtering the collected vibration response signal to separate the low-frequency vibration noise and the high-frequency response component containing the resonant frequency; extracting the resonant frequency from the high-frequency response component; and then substituting the resonant frequency extracted from the high-frequency response component into a preset mass-frequency mapping model to calculate the mass of the material inside the tank.

[0016] In one specific embodiment, the mass-frequency mapping model is established through a calibration method, which includes:

[0017] Under static conditions where a series of materials of known mass are loaded into a tank, the resonant frequencies corresponding to each known mass are collected by a perturbation excitation unit and a multidimensional vibration response sensor array, thereby obtaining a data pair of material mass and resonant frequency.

[0018] By fitting the data pairs, a mathematical relationship between material mass and resonant frequency is established, thereby generating a mass-frequency mapping model. This model can be represented by the following polynomial:

[0019] ;

[0020] In the formula: The mass of the material to be measured; For the quality-frequency mapping model function; The main resonant frequency; The summation symbol indicates that the index will be used for summation. From 0 to Add all the terms together; For the first multinomial regression Term coefficient.

[0021] In one specific embodiment, the perturbation excitation unit includes piezoelectric ceramic actuators deployed orthogonally on the outer wall of the tank; the multidimensional vibration response sensing array includes at least four accelerometers deployed in a spatially non-collinear manner on the outer wall of the tank.

[0022] Preferably, the modal analysis and diagnostic decision unit is further configured as follows:

[0023] Based on the vibration response signal, the signal quality is evaluated in real time. The evaluation metrics for signal quality include signal-to-noise ratio and / or quality factor.

[0024] ;

[0025] In the formula: Signal-to-noise ratio; This is the commonly used logarithm symbol; The power spectral density at the peak of the resonant frequency; This represents the average power spectral density of the noise substrate near the peak of the resonant frequency.

[0026] ;

[0027] In the formula: For quality factor; The center frequency of the resonance peak; This represents the frequency bandwidth of the resonant peak at a 3dB drop in peak power.

[0028] When the signal quality is lower than a preset quality threshold, the excitation mode of the excitation signal is adaptively adjusted to improve the signal quality of the subsequently acquired vibration response signal. The quality threshold is a minimum signal quality standard determined in reverse, based on ensuring that the material quality calculation result meets preset accuracy requirements.

[0029] In one specific embodiment, the adaptive adjustment of the excitation signal excitation mode specifically includes: switching the excitation signal from a wideband pulse excitation mode to a narrowband sweep frequency excitation mode, wherein the sweep frequency center frequency of the narrowband sweep frequency excitation mode is set based on the resonant frequency identified before the switching.

[0030] Preferably, the modal analysis and diagnostic decision unit is further configured as follows:

[0031] Multiple resonant frequencies are extracted to form a multimodal fingerprint vector;

[0032] Based on the currently calculated material quality, a standard fingerprint vector is matched or generated from the preset normal operating condition fingerprint library;

[0033] The multimodal fingerprint vector extracted in real time is compared with the standard fingerprint vector.

[0034] The normal operating condition fingerprint database is established by extracting and recording multiple resonant frequencies corresponding to multiple known material masses for each known material mass.

[0035] In one specific embodiment, the modal analysis and diagnostic decision unit is further configured as follows:

[0036] Based on the comparison results, a fingerprint distortion degree is calculated.

[0037] ;

[0038] In the formula: Fingerprint distortion; The L2 norm symbol indicates the calculation of the Euclidean distance between vectors; This is the normalized multimodal fingerprint vector extracted in real time; This is the standard normalized fingerprint vector function in the normal operating condition fingerprint database; This refers to the material mass calculated in real time.

[0039] When the fingerprint distortion exceeds a preset distortion threshold, the tank is determined to have an abnormal operating condition. This distortion threshold is determined based on statistical analysis of fingerprint distortion data collected from the tank under various normal operating conditions; this threshold is used to distinguish between data fluctuations under normal operating conditions and abnormal operating conditions.

[0040] This invention provides a semi-finished product mixing tank for paper tube adhesive with a weighing function. It has the following beneficial effects:

[0041] 1. This invention achieves precise geometric matching between a hyperboloid rotor and a W-shaped groove, and utilizes a planetary differential drive system to realize a composite motion of revolution and differential rotation, forming a moving line-contact high-pressure grinding zone at the bottom of the tank. Combined with the forced conveying effect of the spiral guide groove on the rotor surface, it can structurally eliminate mixing dead zones, achieve efficient grinding-level dispersion and uniform mixing of materials, and solve the problems of mixing dead zones and uneven dispersion in the stirring process of high-viscosity materials.

[0042] 2. This invention applies high-frequency perturbation excitation to the tank and collects vibration response signals. It uses modal analysis technology to effectively separate the resonant frequency containing material mass information from the low-frequency vibration noise generated by stirring. Based on a preset mass-frequency mapping model, it performs calculations to achieve real-time and dynamic measurement of the material mass in the tank under intense stirring conditions. This solves the problem that traditional weighing methods are easily affected by mechanical vibration and cannot accurately measure the mass.

[0043] 3. This invention extracts a multimodal fingerprint vector composed of multiple resonant frequencies and compares it in real time with the standard fingerprint vector in the normal operating condition fingerprint library to calculate the fingerprint distortion degree. This can quantify the degree of deviation of the equipment's operating status, thereby issuing timely alarms when situations such as loose stirring components, abnormal material agglomeration, or wall adhesion occur. While achieving accurate weighing, it also provides online monitoring and abnormal early warning functions for the operating condition of the mixing tank equipment. Attached Figure Description

[0044] Figure 1 This is a perspective view of the present invention;

[0045] Figure 2 This is a schematic diagram of the side structure of the tank body of the present invention;

[0046] Figure 3 This is a schematic diagram of the structure below the top cover of the present invention;

[0047] Figure 4 This is a schematic diagram of the rotating frame and gear of the present invention;

[0048] Figure 5 This is a schematic diagram of the stirring mechanism of the present invention;

[0049] Figure 6 This is a schematic diagram of the planetary gear mechanism of the present invention;

[0050] Figure 7 This is a schematic cross-sectional view of the tank body of the present invention;

[0051] Figure 8 This is a functional block diagram of a weighing system according to an embodiment of the present invention;

[0052] Figure 9 This is a functional block diagram of the modal analysis and diagnostic decision unit according to an embodiment of the present invention.

[0053] The components include: 1. Tank body; 2. Top cover; 3. Motor; 4. Rotating frame;

[0054] 5. Mixing mechanism; 501. Mixing frame; 502. Hyperboloid rotor; 503. W-shaped groove; 504. Scraper; 505. Cutting blade;

[0055] 6. Planetary gear mechanism; 601. Internal gear ring; 602. Gear one; 603. Gear two;

[0056] 7. Piezoelectric ceramic actuator; 8. Accelerometer; 9. Controller; 100. Perturbation excitation unit; 200. Multidimensional vibration response sensor array; 300. Modal analysis and diagnostic decision unit. Detailed Implementation

[0057] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0058] Please see the appendix Figure 1 - Appendix Figure 7 This invention provides a semi-finished mixing tank for paper tube adhesive with weighing function, including a tank body 1, a top cover 2 on the top of the tank body 1, a motor 3 installed on the top of the top cover 2, a rotating frame 4 fixedly connected to the output end of the motor 3, a stirring mechanism 5 in the middle of the rotating frame 4, a planetary gear mechanism 6 at the bottom of the top cover 2, the planetary gear mechanism 6 being used to control the stirring trajectory of the stirring mechanism 5, and a controller 9 installed on the side of the tank body 1, the controller 9 having a built-in weighing system;

[0059] The stirring mechanism 5 includes a stirring frame 501, which is rotatably connected to the middle of the rotating frame 4. A hyperboloid rotor 502 is fixedly connected to the bottom of the stirring frame 501. A W-shaped groove 503 is provided inside the tank 1. The shape of the W-shaped groove 503 matches the shape of the bottom of the hyperboloid rotor 502. A spiral guide groove is provided on the outer side of the hyperboloid rotor 502. The planetary gear mechanism 6 includes an internal gear ring 601, which is fixedly connected to the bottom of the top cover 2. A gear 602 is fixedly connected to the top of the rotating frame 4. A gear 603 is fixedly connected to the top of the stirring frame 501. The internal gear ring 601 and the gear 603 mesh with each other. The gear 602 and the gear 603 mesh with each other.

[0060] Specifically, the cross-sectional profile of the hyperboloid rotor 502 is generated based on a specific hyperbola. Correspondingly, the inner wall of the W-shaped groove 503 is also composed of a corresponding hyperbola. This precise geometric matching enables a continuous, dynamically moving line contact high-pressure zone to be formed between the outer surface of the hyperboloid rotor 502 and the inner wall of the W-shaped groove 503 during assembly and movement. When the material is brought into this area, it will be subjected to great extrusion and grinding, thereby breaking down stubborn powder agglomerates or gel particles in high-viscosity materials.

[0061] When the motor 3 drives the rotating frame 4 to revolve around the central axis of the tank 1, the second gear 603, which is a planetary gear, moves with the rotating frame 4 on the one hand, and on the other hand, because it meshes with the fixed internal gear ring 601, it is forced to rotate around its own axis. At the same time, by presetting the number of teeth of the first gear 602, the second gear 603 and the internal gear ring 601, there is a set difference between the rotational angular velocity of the second gear 603 and the angular velocity in the pure rolling state. This speed difference causes the hyperboloid rotor 502 to crush the material in the online contact area while also generating high-speed relative sliding friction, thereby applying a strong shearing force and realizing efficient dispersion and kneading of the material.

[0062] The spiral guide grooves machined on the outer surface of the hyperboloid rotor 502 have preset pitch and helix angle. When the hyperboloid rotor 502 rotates, these spiral guide grooves function like a screw pump, generating a downward axial driving force on the viscous material in the tank. This driving force forces the material that has not been fully mixed in the upper and central areas of the mixing tank to the line contact high-pressure grinding zone at the bottom of the tank 1 for processing. The processed material is then turned up from both sides of the grinding zone, thereby establishing a full-volume circulating flow field without dead angles inside the entire tank 1, ensuring that all materials can be homogenized.

[0063] Please see the appendix Figure 1 - Appendix Figure 7In a preferred embodiment of the present invention, a scraper 504 is fixedly connected to the side of the stirring rack 501, a cutting blade 505 is fixedly connected to the middle of the stirring rack 501, and a piezoelectric ceramic actuator 7 and an acceleration sensor 8 are provided on the outside of the tank body 1.

[0064] Specifically, the outer side of the scraper 504 contacts the inner wall of the tank 1 to scrape off the material attached to the inner wall, and the cutting blade 505 can cut during the rotation and stirring process to improve the uniformity of the material.

[0065] Please see the appendix Figure 8 The weighing system provided by the present invention may include: a perturbation excitation unit 100 configured to apply an excitation signal to a tank 1; a multidimensional vibration response sensor array 200 configured to collect the vibration response signal of the tank 1 to the excitation signal; and a modal analysis and diagnostic decision unit 300, which is signal-connected to the perturbation excitation unit 100 and the multidimensional vibration response sensor array 200.

[0066] During system operation, the modal analysis and diagnostic decision unit 300 controls the perturbation excitation unit 100 to apply a preset excitation signal to the tank 1. The tank 1 vibrates under the action of the excitation signal, and this vibration response is collected by a multidimensional vibration response sensor array 200 deployed outside the tank 1, forming a time-domain vibration response signal. The multidimensional vibration response sensor array 200 transmits the collected time-domain vibration response signal to the modal analysis and diagnostic decision unit 300 for processing.

[0067] The modal analysis and diagnostic decision unit 300 is configured to process and analyze the received time-domain vibration response signal. First, the preprocessed time-domain vibration response signal is analyzed using a Fast Fourier Transform. Convert to frequency domain response signal In order to identify the resonant frequency of the tank 1 structure under the current material load.

[0068] ;

[0069] In the formula: It is a frequency domain response signal; For frequency; The integral symbol indicates the integral with respect to the time variable. Integration operations; This is the preprocessed time-domain vibration response signal; is the base of the natural logarithm; It is the imaginary unit.

[0070] After obtaining the frequency domain response signal, the modal analysis and diagnostic decision unit 300 extracts one or more resonant frequencies from it, and selects the main resonant frequency with the highest signal-to-noise ratio and the most stable signal-to-noise ratio. Substitute it into the preset quality and frequency mapping model function In the process, the mass of the material inside tank 1 is calculated. .

[0071] ;

[0072] In the formula: The mass of the material to be measured; For the quality-frequency mapping model function;

[0073] The main resonant frequency; The summation symbol indicates that the index will be used for summation. From 0 to Add all the terms together; For the first multinomial regression Term coefficient.

[0074] To ensure the accuracy of the measurement results, the modal analysis and diagnostic decision unit 300 is also configured to perform real-time evaluation of the quality of the vibration response signal, with evaluation metrics including signal-to-noise ratio. and quality factor .

[0075] ;

[0076] In the formula: Signal-to-noise ratio; This is the commonly used logarithm symbol; The power spectral density at the peak of the resonant frequency; This represents the average power spectral density of the noise substrate near the peak of the resonant frequency.

[0077] ;

[0078] In the formula: For quality factor; The center frequency of the resonance peak; This represents the frequency bandwidth of the resonant peak at a 3dB drop in peak power.

[0079] In addition, to monitor the operating conditions of the mixing tank, the modal analysis and diagnostic decision unit 300 is also configured to extract multiple resonant frequencies to form a real-time normalized multimodal fingerprint vector. By comparing this vector with the corresponding current material mass in the normal operating condition fingerprint database... Standard fingerprint vector Compare and calculate the fingerprint distortion degree This is to determine whether the equipment is operating abnormally.

[0080] ;

[0081] In the formula: Fingerprint distortion; The L2 norm symbol indicates the calculation of the Euclidean distance between vectors; This is the normalized multimodal fingerprint vector extracted in real time; This is the standard normalized fingerprint vector function in the normal operating condition fingerprint database; This refers to the material mass calculated in real time.

[0082] In this process, a normal operating condition fingerprint database is established by extracting and recording multiple resonant frequencies corresponding to multiple known material masses.

[0083] The perturbation excitation unit 100 in this embodiment of the invention is used to apply a controllable excitation signal to the tank 1 to excite the vibration response of the tank 1 structure.

[0084] In one specific embodiment, at least two perturbation excitation units 100 may be installed, and each unit is composed of a piezoelectric ceramic actuator 7. The piezoelectric ceramic actuator 7 is chosen because it has a wide frequency response range and microsecond-level fast response characteristics, and is able to generate a high-frequency excitation signal that is frequency-domain separated from the low-frequency vibration noise generated by the stirring condition.

[0085] When installing two piezoelectric ceramic actuators 7, they are orthogonally deployed to comprehensively excite the multiple vibration modes of the tank 1. For example, in one embodiment, one piezoelectric ceramic actuator 7 can be arranged in the middle of the tank 1, and the other piezoelectric ceramic actuator 7 can be arranged circumferentially (horizontally) along the tank 1 at a distance of 90 degrees from the first piezoelectric ceramic actuator 7. This orthogonal deployment ensures that regardless of whether the main vibration mode of the tank 1 is axial bending, circumferential extension, or more complex coupled modes, it can be effectively excited by at least one actuator, thereby obtaining richer modal information in the subsequent vibration response signal.

[0086] The perturbation excitation unit 100 is electrically connected to the output port of the modal analysis and diagnostic decision unit 300 and is driven and controlled by it. According to the instructions issued by the modal analysis and diagnostic decision unit 300, the perturbation excitation unit 100 can generate at least two different modes of excitation signals. The first is a wideband pulse excitation signal, which is a short-duration pulse in the time domain and corresponds to a flat and wide spectrum in the frequency domain. It can simultaneously excite all resonant frequencies of the tank 1 within a wide frequency band, suitable for initial system exploration or rapid measurement. The second is a narrowband swept-frequency excitation signal, which is a sine wave whose frequency scans linearly or logarithmically around a preset center frequency. It can concentrate the excitation energy within a specific frequency range, suitable for high signal-to-noise ratio accurate measurement of preliminarily identified resonant frequencies.

[0087] The multidimensional vibration response sensing array 200 in this embodiment of the invention is used to collect multi-point vibration response signals generated by the tank 1 under the excitation of the perturbation excitation unit 100, and convert these signals into electrical signals.

[0088] In one specific embodiment, the multidimensional vibration response sensing array 200 may consist of at least four accelerometers 8. Accelerometers 8 with high sensitivity, wide bandwidth response, and high signal-to-noise ratio, such as piezoelectric accelerometers, are selected to ensure accurate pickup of weak vibration responses excited by high-frequency excitation signals, while covering all resonant frequency ranges requiring analysis.

[0089] When selecting four accelerometer sensors 8 for installation, in order to comprehensively capture the overall vibration pattern of the tank 1, the four accelerometer sensors 8 are deployed on the outer wall of the tank 1 in a spatially non-collinear manner. This deployment method is fundamental to ensuring the ability to identify and distinguish translational, rotational, and complex coupled vibration modes. In one embodiment, the four accelerometer sensors 8 are distributed at different heights above and below the tank 1 and arranged at staggered angles in the circumferential direction, together forming a three-dimensional sensing array.

[0090] Each accelerometer 8 in the multidimensional vibration response sensor array 200 is connected to multiple input channels of the modal analysis and diagnostic decision unit 300 via shielded cables. The shielded cables are used to suppress electromagnetic interference from the motor 3 or other external devices. The modal analysis and diagnostic decision unit 300 can simultaneously acquire vibration response signals from these four accelerometers 8, thus providing a data foundation for subsequent multi-channel signal processing and accurate modal parameter identification.

[0091] Please refer to the appendix. Figure 9 The modal analysis and diagnostic decision unit 300 in this embodiment of the invention is a central processing and control component that executes all algorithms, including data acquisition, signal processing, quality calculation, and operating condition diagnosis. The modal analysis and diagnostic decision unit 300 integrates a signal acquisition and preprocessing module, a quality-frequency mapping model establishment module, a real-time material quality calculation module, a signal quality assessment and adaptive excitation control module, and an operating condition anomaly diagnosis module.

[0092] The signal acquisition and preprocessing module integrated within the modal analysis and diagnostic decision unit 300 is responsible for digitizing and purifying the raw vibration signals received from the multidimensional vibration response sensor array 200 to extract effective signal components for modal analysis.

[0093] The signal acquisition and preprocessing module first synchronously samples and digitizes the analog signals from the four accelerometers 8 using its built-in multi-channel high-speed analog-to-digital converter (ADC). The sampling frequency is set according to the Nyquist sampling theorem, and its value is set to more than twice the highest resonant frequency expected to be analyzed. For example, if the upper limit of the resonant frequency to be analyzed is 10kHz, the sampling frequency can be set to 25.6kHz to avoid frequency aliasing and retain complete signal information.

[0094] Digitalized signal It is a mixed signal containing low-frequency strong vibration noise and high-frequency perturbation response. To achieve frequency domain separation, the signal acquisition and preprocessing module uses a digital high-pass filter to process the signal.

[0095] In one embodiment, a Butterworth high-pass filter can be used. The cutoff frequency of this filter... The setting is based on the following: its value is higher than all major vibration harmonic frequencies (e.g., revolution frequency, rotation frequency, and their harmonics) generated when the stirring mechanism 5 is operating, while being lower than the starting frequency of the excitation signal applied by the perturbation excitation unit 100. Through this filtering operation, low-frequency, high-amplitude vibration noise components generated by material tumbling, mechanical transmission, etc., can be filtered out, retaining only high-frequency, low-amplitude response signal components containing the inherent resonance information of the tank 1 structure. .

[0096] Before performing spectrum analysis, to reduce the spectral leakage effect caused by signal truncation, the signal acquisition and preprocessing module will also process the filtered signal. Apply a window function, such as the Hanning window. Applying a window function allows the two ends of the truncated signal segment to transition smoothly to zero, thereby concentrating the energy more at the true frequency peak and improving the accuracy of resonant frequency identification.

[0097] After filtering and windowing, this module calls the Fast Fourier Transform (FFT) algorithm to transform the time-domain signal. Transformed into frequency domain response signal This is so that subsequent modules can extract and analyze the resonant frequency.

[0098] ;

[0099] In the formula: It is a frequency domain response signal; For frequency; The integral symbol indicates the integral with respect to the time variable. Integration operations; This is the preprocessed time-domain vibration response signal; is the base of the natural logarithm; It is the imaginary unit.

[0100] The module for establishing the mass and frequency mapping model integrated in the modal analysis and diagnostic decision unit 300 is designed to establish a precise mathematical relationship between the material mass and the main resonant frequency of the tank body 1 through a one-time calibration process before the mixing tank is officially put into use.

[0101] The calibration process is carried out under static conditions (i.e., the stirring mechanism 5 is not running), and the specific steps are as follows:

[0102] First, a series of materials of known quality (e.g., calibration materials with the same or similar physical properties as the paper tube adhesive semi-finished product used in actual production) are added to the empty tank 1 in batches. For example, starting from an empty tank, the material is gradually added in units of 10% of the rated capacity of the mixing tank until it is fully loaded. Each time, the mass of the material added to the tank 1 is accurately recorded using external measuring equipment such as a high-precision weighbridge. .

[0103] In each given material mass The module controls the perturbation excitation unit 100 to apply broadband pulse excitation to the tank 1, and the multidimensional vibration response sensor array 200 collects the vibration response signal. The signal acquisition and preprocessing module processes the acquired signal to generate a frequency domain response signal spectrum. The mass-frequency mapping model establishment module automatically or with manual assistance identifies the most stable and sharpest peak resonant frequency under the current mass from the spectrum. By repeating this process, a set of data pairs covering the entire weighing range is obtained. ,in , This represents the total number of data points calibrated.

[0104] After obtaining the complete set of data pairs, this module uses a multinomial regression algorithm to fit the data.

[0105] Specifically, the least squares method is used to determine a coefficients of the terms of the polynomial This ensures that the curve described by the polynomial best fits the aforementioned data pair. The order of the polynomial. This is a preset integer, and its selection principle is to ensure fitting accuracy while avoiding overfitting.

[0106] ;

[0107] In the formula: The mass of the material to be measured; For the quality-frequency mapping model function;

[0108] The main resonant frequency; The summation symbol indicates that the index will be used for summation. From 0 to Add all the terms together; For the first multinomial regression Term coefficient.

[0109] After the calibration process is completed, the determined polynomial order is... and each coefficient ( The data is stored as a complete mass-frequency mapping model in the non-volatile memory of the modal analysis and diagnostic decision unit 300, so that the real-time material mass calculation module can call it in subsequent dynamic measurements.

[0110] The real-time material quality calculation module integrated in the modal analysis and diagnostic decision unit 300 is designed to receive the frequency domain response signal output by the signal acquisition and preprocessing module and calculate the current mass of the material in tank 1 using the calibrated mass-frequency mapping model.

[0111] The real-time calculation module for the material mass receives the frequency domain response signal. Then, a resonant frequency locking algorithm is executed first. In one embodiment, this algorithm searches for peaks within a preset frequency range, which is determined based on the resonant frequency variation range obtained during the calibration phase. The real-time material mass calculation module traverses all frequency points within this range, searching for local maxima of the power spectral density. To improve the accuracy of resonant frequency identification, after locking a discrete frequency peak, the real-time material mass calculation module can also use interpolation algorithms, such as parabolic interpolation or Gaussian interpolation, to fit the peak and its two adjacent frequency points, thereby calculating an accurate principal resonant frequency with a resolution higher than the inherent frequency resolution of the Fast Fourier Transform (FFT). .

[0112] After accurately determining the main resonant frequency Subsequently, the real-time mass calculation module of the material uses this frequency value as an input variable and substitutes it into the mass-frequency mapping model function pre-stored in non-volatile memory. The calculations are performed in the middle.

[0113] ;

[0114] In the formula: The mass of the material to be measured; For the quality-frequency mapping model function;

[0115] The main resonant frequency; The summation symbol indicates that the index will be used for summation. From 0 to Add all the terms together; For the first multinomial regression Term coefficient.

[0116] Calculated material mass This represents the measurement result of the material inside tank 1 at the current moment. This calculation process is executed cyclically at a preset time period (e.g., every 1 second), thereby achieving continuous and dynamic updates to the material quality. The real-time material quality calculation module displays the results of each calculation. The output can be used for analysis by the subsequent abnormal operation diagnosis module, or it can be sent to an external display unit or host computer system through the communication interface.

[0117] The signal quality assessment and adaptive excitation control module integrated in the modal analysis and diagnostic decision unit 300 functions to quantitatively assess the quality of the frequency domain response signal used for quality calculation in real time, and to actively adjust the excitation strategy to ensure the continuity and accuracy of the measurement when the signal quality deteriorates.

[0118] After each resonant frequency lock, the signal quality assessment and adaptive excitation control module will base its assessment on the frequency domain response signal. Calculate at least two signal quality evaluation metrics: signal-to-noise ratio (SNR) ) and quality factor ( The signal-to-noise ratio (SNR) measures the degree of separation between the resonant peak and the noise floor, while the quality factor measures the sharpness of the resonant peak.

[0119] ;

[0120] In the formula: Signal-to-noise ratio; This is the commonly used logarithm symbol; The power spectral density at the peak of the resonant frequency; This represents the average power spectral density of the noise substrate near the peak of the resonant frequency.

[0121] ;

[0122] In the formula: For quality factor; The center frequency of the resonance peak; This represents the frequency bandwidth of the resonant peak at a 3dB drop in peak power.

[0123] When calculated value or When the factor falls below a preset quality threshold, the signal quality assessment and adaptive excitation control module determines that the current signal quality is insufficient to support high-precision quality calculation. This quality threshold is the minimum signal quality standard determined in reverse, based on ensuring that the material quality calculation result meets the preset accuracy requirement (e.g., accuracy better than 1%).

[0124] Once it is determined that the signal quality does not meet the requirements, the signal quality assessment and adaptive excitation control module will send an instruction to the perturbation excitation unit 100 to adaptively adjust the excitation mode of the excitation signal.

[0125] In one embodiment, the excitation mode is switched from the default wideband pulse excitation mode to a narrowband sweep excitation mode. This switch concentrates the excitation energy within a limited frequency band at the resonant frequency, thereby significantly increasing the response amplitude at the resonant frequency while maintaining the same total excitation energy, and thus improving the signal-to-noise ratio. The center frequency of the narrowband sweep excitation mode is set based on the main resonant frequency identified before signal quality degradation, and its sweep range covers a small interval around that center frequency, for example... .

[0126] The working condition anomaly diagnosis module integrated in the modal analysis and diagnostic decision unit 300 is designed to perform online monitoring and anomaly diagnosis of the structural health and operating conditions of the mixing tank by analyzing the comprehensive characteristics of the vibration modes of the tank 1 while performing quality measurements.

[0127] The functionality of this abnormal operating condition diagnosis module is based on a pre-established normal operating condition fingerprint database. The establishment of this database is synchronized with the quality calibration process. This involves adding each batch of material of known quality to the tank. When performing resonant frequency measurements, this abnormal operating condition diagnostic module not only records the main resonant frequency, but also identifies and records several other frequencies (e.g., ( ) stable and repeatable resonant frequencies, together forming a dimensional frequency vector To eliminate the unidirectional influence of material quality changes on all frequencies and highlight the changes in the relative relationships between frequencies caused by structural state variations, this abnormal operating condition diagnosis module normalizes the frequency vectors. For example, it divides all frequency components by the principal resonant frequency. To obtain a normalized standard fingerprint The set of standard fingerprint vectors corresponding to all quality points constitutes a normal operating condition fingerprint database stored in non-volatile memory.

[0128] During real-time dynamic measurement, the abnormal condition diagnosis module will also extract a value after each spectrum analysis. 3D real-time frequency vector The same normalization method used when establishing the fingerprint database was employed to calculate the real-time normalized multimodal fingerprint vector. .

[0129] Subsequently, the abnormal operation diagnosis module utilizes the current material mass calculated by the real-time material mass calculation module. By looking up tables and interpolating, a standard fingerprint vector corresponding to the current quality is obtained from the normal operating condition fingerprint database. .

[0130] This abnormal operating condition diagnosis module quantifies the degree of deviation of the current device operating condition from the normal state by calculating the Euclidean distance between the real-time fingerprint vector and the standard fingerprint vector. This distance is defined as the fingerprint distortion degree. .

[0131] ;

[0132] In the formula: Fingerprint distortion; The L2 norm symbol indicates the calculation of the Euclidean distance between vectors; This is the normalized multimodal fingerprint vector extracted in real time; This is the standard normalized fingerprint vector function in the normal operating condition fingerprint database; This refers to the material mass calculated in real time.

[0133] Finally, the abnormal operation diagnosis module will calculate the fingerprint distortion degree. With a preset alarm threshold Compare. If If the threshold is exceeded continuously, it is determined that the mixing tank has an abnormal operating condition, such as loose mixing components or abnormal agglomeration of materials on the tank wall. An alarm signal is then generated and sent to the host computer or control system through the communication interface.

[0134] Working principle: After the equipment is started, the motor 3 starts to run, and its output torque is transmitted to the entire stirring and driving device through the rotating frame 4. The rotating frame 4, as a whole, begins to revolve around the central axis of the tank 1. The revolving motion enables the stirring mechanism 5 to cover all areas of the bottom surface inside the tank 1, thereby performing the first macroscopic tumbling and mixing of the material.

[0135] While the rotating frame 4 revolves, the internal gear ring 601 is fixed to the bottom of the top cover 2 and remains stationary. The gear 603, which revolves with the rotating frame 4, rolls along the internal teeth of the internal gear ring 601, and its external teeth also mesh with the internal gear ring 601. This forces the gear 603 to rotate around its own central axis while revolving around the center of the tank 1. By specifically configuring the number of teeth on the gear 602, gear 603, and internal gear ring 601, the rotation of gear 603 is not pure rolling, but a differential rotation with a preset speed difference. The material is completely transmitted to the hyperboloid rotor 502 through the stirring rack 501, so that the final motion of the hyperboloid rotor 502 is a composite motion of revolution and differential rotation. When the hyperboloid rotor 502 moves to the area where its outer surface forms a line contact with the inner wall of the W-shaped groove 503, the revolution brings huge extrusion force to grind the material, while the differential rotation causes high-speed relative sliding in the line contact area, thereby applying a strong shear force to the material. This composite action of extrusion and grinding and high-speed shearing can effectively break and disperse the agglomerates in the material.

[0136] Meanwhile, the outer surface of the hyperboloid rotor 502 is provided with a spiral guide groove, which generates a continuous, top-down axial thrust on the high-viscosity material, forcibly pumping the material in the tank 1 to the grinding and shearing zone at the bottom of the tank, establishing a material circulation throughout the container, avoiding slippage due to excessively high material viscosity as the agitator rotates, and ensuring that all materials in the tank 1 can be homogenized.

[0137] In the dynamic weighing and working condition diagnosis workflow, the modal analysis and diagnosis decision unit 300 actively controls the perturbation excitation unit 100 to apply a high-frequency excitation signal separated from the stirring noise frequency band to the tank 1. The vibration response of the tank 1 structure under this excitation is collected by the multi-dimensional vibration response sensor array 200, and the electrical signal containing structural modal information is transmitted back to the modal analysis and diagnosis decision unit 300. The modal analysis and diagnosis decision unit 300 first performs high-pass filtering on the signal to filter out the low-frequency strong interference generated by material stirring, and then uses spectrum analysis to lock in specific main parameters strongly correlated with material quality. The modal analysis and diagnostic decision unit 300 calculates the real-time material mass by substituting the main resonant frequency into a preset mass-frequency mapping model. Simultaneously, the modal analysis and diagnostic decision unit 300 extracts multiple resonant frequencies from the spectrum to form a multimodal fingerprint. By comparing this fingerprint with the standard fingerprint in the normal operating condition fingerprint library, the fingerprint distortion degree is calculated, enabling online monitoring and anomaly diagnosis of the equipment's structural health status. Throughout the process, the modal analysis and diagnostic decision unit 300 also evaluates the signal quality and adaptively switches the excitation mode when the signal quality deteriorates to ensure the stability and accuracy of the measurement results.

Claims

1. A semi-finished mixing tank for paper tube adhesive with weighing function, comprising a tank body (1), characterized in that, The tank (1) is provided with a top cover (2), a motor (3) is installed on the top of the top cover (2), a rotating frame (4) is fixedly connected to the output end of the motor (3), a stirring mechanism (5) is provided in the middle of the rotating frame (4), a planetary gear mechanism (6) is provided at the bottom of the top cover (2), the planetary gear mechanism (6) is used to control the stirring trajectory of the stirring mechanism (5), a controller (9) is installed on the side of the tank (1), and the controller (9) has a built-in weighing system. The stirring mechanism (5) includes a stirring frame (501), which is rotatably connected to the middle of the rotating frame (4). A hyperboloid rotor (502) is fixedly connected to the bottom of the stirring frame (501). A W-shaped groove (503) is provided inside the tank (1). The shape of the W-shaped groove (503) matches the shape of the bottom of the hyperboloid rotor (502). A spiral guide groove is provided on the outer side of the hyperboloid rotor (502). The planetary gear mechanism (6) includes an internal gear ring (601), which is fixedly connected to the bottom of the top cover (2). A gear one (602) is fixedly connected to the top of the rotating frame (4). A gear two (603) is fixedly connected to the top of the stirring frame (501). The internal gear ring (601) and the gear two (603) mesh with each other. The gear one (602) and the gear two (603) mesh with each other. A scraper (504) is fixedly connected to the side of the stirring rack (501), a cutting blade (505) is fixedly connected to the middle of the stirring rack (501), and a piezoelectric ceramic actuator (7) and an acceleration sensor (8) are provided on the outer wall of the tank (1). The weighing system includes: At least two perturbation excitation units are configured to apply an excitation signal to the tank (1); A multidimensional vibration response sensing array configured to acquire the vibration response signal of the tank (1) to the excitation signal; A modal analysis and diagnostic decision unit is electrically connected to the perturbation excitation unit and the multidimensional vibration response sensing array. The modal analysis and diagnostic decision unit is configured as follows: The perturbation excitation unit is controlled to emit the excitation signal; Process the vibration response signal collected by the multidimensional vibration response sensing array and extract at least one resonant frequency that characterizes the current state of the tank (1); Based on the preset mass-frequency mapping model and the extracted resonant frequency, the mass of the material inside the tank (1) is calculated; The excitation signal is a high-frequency excitation signal, and the vibration generated by the tank (1) under stirring conditions is low-frequency vibration noise; The modal analysis and diagnostic decision unit is specifically configured as follows: The collected vibration response signal is filtered to separate the low-frequency vibration noise and the high-frequency response component containing the resonant frequency, and the resonant frequency is extracted from the high-frequency response component. The resonant frequency extracted from the high-frequency response component is substituted into the preset mass-frequency mapping model to calculate the mass of the material inside the tank (1).

2. The semi-finished product mixing tank with weighing function for paper tube adhesive according to claim 1, characterized in that, The quality-frequency mapping model is established through a calibration method, which specifically includes: Under static conditions where a series of materials of known mass are loaded in the tank (1), the resonant frequencies corresponding to each known mass are collected by the perturbation excitation unit and the multidimensional vibration response sensing array, thereby obtaining a data pair of material mass and resonant frequency; The data pairs are fitted to establish a mathematical relationship between the material mass and the resonant frequency, thereby generating the mass-frequency mapping model.

3. The semi-finished product mixing tank for paper tube adhesive with weighing function according to claim 1, characterized in that, The perturbation excitation unit includes the piezoelectric ceramic actuator (7) deployed orthogonally on the outer wall of the tank (1). The multidimensional vibration response sensing array includes at least four acceleration sensors (8) deployed in a spatially non-collinear manner on the outer wall of the tank (1).

4. The semi-finished product mixing tank with weighing function for paper tube adhesive according to claim 1, characterized in that, The modal analysis and diagnostic decision unit is further configured as follows: Based on the vibration response signal, the signal quality is evaluated in real time, and the signal quality includes the signal-to-noise ratio and / or quality factor; When the signal quality is lower than a preset quality threshold, the excitation mode of the excitation signal is adaptively adjusted to improve the signal quality of the vibration response signal acquired subsequently. The quality threshold is a minimum signal quality standard determined in reverse to ensure that the material quality calculation results meet the accuracy requirements.

5. The semi-finished product mixing tank for paper tube adhesive with weighing function according to claim 4, characterized in that, The adaptive adjustment of the excitation mode of the excitation signal specifically includes: The excitation signal is switched from a wideband pulse excitation mode to a narrowband sweep frequency excitation mode, wherein the sweep frequency center frequency of the narrowband sweep frequency excitation mode is set based on the resonant frequency identified before the switch.

6. The semi-finished product mixing tank for paper tube adhesive with weighing function according to claim 1, characterized in that, The modal analysis and diagnostic decision unit is further configured as follows: Multiple resonant frequencies are extracted to form a multimodal fingerprint vector; Based on the currently calculated material mass, a standard fingerprint vector is matched or generated from the preset normal operating condition fingerprint database. The multimodal fingerprint vector extracted in real time is compared with the standard fingerprint vector; The normal operating condition fingerprint database is established by extracting and recording multiple resonant frequencies corresponding to multiple known material masses for each known material mass.

7. The semi-finished product mixing tank for paper tube adhesive with weighing function according to claim 6, characterized in that, The modal analysis and diagnostic decision unit is further configured as follows: Based on the results of the comparison, a fingerprint distortion degree is calculated; When the fingerprint distortion exceeds the preset distortion threshold, it is determined that the tank (1) has an abnormal operating condition; The distortion threshold is determined by statistical analysis of the fingerprint distortion data collected from the tank (1) under various normal operating conditions.

Images