Slicer blade wear monitoring method and system

Abstract

The invention belongs to the technical field of equipment intelligent monitoring, and discloses a slicing machine blade wear monitoring method and system, and the method comprises the steps: collecting blade modal data in the same time sequence and blade state description at a corresponding moment; performing multi-scale space embedding on the blade modal data to obtain a multi-scale space diagram, and performing feature extraction on the multi-scale space diagram to obtain modal space features; performing quantum mapping on the modal space features to obtain feature quantum states, and constructing a blade damage quantum library based on the feature quantum states and blade state description; real-time blade modal data are collected, damage evaluation is conducted on the real-time blade modal data based on the blade damage quantum database, the blade state is obtained, and corresponding early warning is given out based on the blade state; and the real-time performance and the accuracy of the wear monitoring of the blade of the slicing machine are greatly improved.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent equipment monitoring, and more specifically, to a method and system for monitoring blade wear of a slicer. Background Art

[0002] In modern industrial production, slicers, as a common equipment, are widely used in food processing, metal processing, electronic component processing and other fields. They are responsible for cutting materials into the required shape or size. However, the wear problem of slicer blades has always been a key factor affecting production efficiency and product quality. Blade wear will not only reduce cutting accuracy, but may also lead to increased material loss and even cause equipment failure; therefore, timely and effective monitoring of blade wear has become an important task to improve production efficiency, extend blade life and reduce production downtime.

[0003] Existing monitoring methods rely on empirical models or analytical models based on physical laws to infer the wear of blades. These methods often ignore complex dynamic factors, such as the high-frequency vibration generated by slicer blades during the actual cutting process. The changes in these factors lead to low prediction accuracy of traditional methods, especially when facing irregular wear patterns or complex working conditions, and cannot accurately reflect the actual wear state of the blades. Traditional blade wear monitoring technology usually cannot provide a fast enough response time. Due to the long data collection, processing and analysis cycle, it is difficult to respond quickly in the early stage of micro damage to the blade. This means that in the critical stage of blade wear, equipment maintenance personnel cannot obtain accurate alarms in time, thereby delaying the maintenance or replacement of the blade, increasing production costs and downtime. Traditional methods also have problems in accuracy and cannot capture the changes in micro damage in real time. For high-precision, high-load tools such as slicer blades, the accumulation of micro damage directly affects its cutting efficiency and processing quality. If there is no effective technical means to accurately capture and predict these micro damages, the blade may still "seem good" and suffer from severe damage such as fracture and cracking, leading to serious production accidents and equipment damage.

[0004] In view of this, the present invention proposes a slicer blade wear monitoring method and system to solve the above problems. Summary of the invention

[0005] In order to overcome the above-mentioned defects of the prior art and to achieve the above-mentioned purpose, the present invention provides the following technical solution: a method for monitoring the wear of a slicer blade, comprising: S1, collecting blade modal data of the same time series and blade state description at the corresponding moment; S2, performing multi-scale spatial embedding on the blade modal data to obtain a multi-scale spatial graph, performing feature extraction on the multi-scale spatial graph to obtain modal space features; S3, perform quantum mapping on the modal space characteristics to obtain characteristic quantum states, and construct a blade damage quantum library based on the characteristic quantum states and blade state description; S4. Collect real-time blade modal data, perform damage assessment on the real-time blade modal data based on the blade damage quantum library, obtain the blade status, and issue a corresponding warning based on the blade status.

[0006] Furthermore, the blade modal data includes: vibration data, current data, acoustic data and temperature data.

[0007] Furthermore, the method of performing multi-scale spatial embedding on the blade modal data includes: A delay timing set is preset, and the delay timing set includes different delay time scales and delay dimensions. The delay timing set is traversed, and the timing of each type of data in the blade modal data is reconstructed based on each traversed delay time scale and delay dimension. The formula for timing reconstruction is: ;in, The representative delay time scale is , the delay dimension is The interval time series data below, represents the delay dimension, represents the delay time scale, Represents the starting time of the blade modal data, represents the number of delay time scales, represents the selection function; the interval time series data obtained by reconstructing the time series of each element in the delayed time series set is used as a time series space, and each data in the interval time series data is used as a spatial node in the time series space; A node connection threshold is preset, and the spatial nodes of the time space are evaluated for association to obtain the degree of association; based on the degree of association between different spatial nodes, the spatial nodes with a degree of association greater than or equal to the node connection threshold are connected, and the degree of association is used as the weight of the connection edge to obtain a time space graph; the spatial node corresponding to the starting time in each time space graph is used as the connection hub of all time space graphs, the connection hub is used as the connection point, and each time space graph is used as a spatial plane, and different time space graphs are connected to the same space to obtain a multidimensional space graph; the spatial nodes at the same time in different spatial planes are used as plane connection points, and different spatial planes are merged based on the plane connection points. The same plane connection points are merged into one spatial node, and the connection relationship between the plane connection points and the spatial nodes in different spatial planes is retained to obtain a multi-scale space graph.

[0008] Furthermore, the formula for performing association evaluation on the spatial nodes in the time series space is: ;in, Representative The spatial nodes and The correlation degree of spatial nodes, Represents the data category in the spatial node, Representative The spatial node categories of data, Representative The spatial node categories of data, represents the control constant, Represents the index of the data category in the spatial node.

[0009] Furthermore, the method of extracting features from the multi-scale spatial graph includes: Take each spatial node in the multiscale spatial graph as the initial feature point, initialize the multiscale set of each initial feature point to be empty, perform low-dimensional exploration on the initial feature point, find the spatial node with a direct edge connected to the initial exploration point in the multiscale spatial graph as a one-dimensional association point, all one-dimensional association points constitute a one-dimensional association set, and the one-dimensional association set is included in the multiscale set; preset the exploration dimension set and the multidimensional point set, traverse the exploration dimension set, take the result of each traversal as the total exploration step length, perform multi-dimensional exploration on the multiscale spatial graph based on the total exploration step length, take the initial feature point as the starting point, and take the starting point as the starting point. The one-dimensional association set is used as the selection set, and a spatial node is selected from the selection set as the next selection point by using a random selection algorithm, and the next selection point is counted into the multidimensional point set, and the next selection point is used as the new starting point to continue the selection. When the number of elements in the multidimensional point set is equal to the total exploration step, the selection is stopped, and the multidimensional point set at this time is output as the to-be-selected set. The to-be-selected set is connected to the detection, and when the last element in the to-be-selected set is the initial feature point, the to-be-selected set is added to the multi-scale set. When the last element in the to-be-selected set is not the initial feature point, the to-be-selected set is discarded, and the process is repeated until the multi-scale set no longer changes; Based on the multi-scale set, each initial feature point is evaluated for feature prominence. The formula for feature prominence evaluation is: ;in, represents the feature evaluation value, represents the size of the multi-scale set, represents the index of the element in the multi-scale collection, Represents the element length, Represents the first The first element The correlation between the spatial nodes and the initial feature points, Represents the first The first element The time scale of a spatial node, represents the time scale of the initial feature point, Represents the maximum value of the element length; based on the size of the feature evaluation value, the initial feature points are sorted from large to small, and the first N initial feature points are selected as the modal salient features. All modal salient features constitute the modal space features.

[0010] Furthermore, the method of performing quantum mapping on the modal space characteristics includes: A normalization algorithm is used to normalize each type of data in the modal space feature to obtain normalized feature data, and modal analysis is performed on the normalized feature data. The formula for modal analysis is: ;in, Representative Class-normalized feature data The parsed value at the moment, Representative Class-normalized feature data The value of the moment, represents the imaginary unit, represents pi, represents the scale factor, represents infinity; based on the analytical value, the normalized feature data is amplitude mapped to obtain the feature amplitude; based on the analytical value, the normalized feature data is offset mapped to obtain the feature offset; based on the feature amplitude and feature offset, a modal space table is constructed for each modal data, and the modal space table records the feature amplitude and feature offset of the normalized feature data at different times; based on the modal space table, a quantum space table is constructed, and the feature amplitude and feature offset of the normalized feature data in the modal space table at different times are quantum constructed to obtain a composite quantum state; quantum phase calculation is performed on the composite quantum state of different types of normalized feature data to obtain quantum phase; the quantum phase and composite quantum state of different normalized feature data at the same time constitute a characteristic quantum state.

[0011] Furthermore, the formula for calculating the quantum phase of the composite quantum state of different types of normalized characteristic data is: ;in, Representative Class-normalized feature data and Class-normalized feature data The quantum nature of moments, Representative Class-normalized feature data The complex conjugate of the composite quantum state at time instant, Representative Class-normalized feature data The composite quantum state at time Representative Class-normalized feature data The composite quantum state at time Representative Class-normalized feature data The complex conjugate of the composite quantum state at time instant.

[0012] Furthermore, the method of constructing the blade damage quantum library includes: Based on the time scale corresponding to the characteristic quantum state, time sorting is performed to obtain the time series quantum state; based on the blade state description at the time corresponding to the time scale in the time series quantum state, a blade state sequence table is constructed, and the blade state sequence table contains blade state descriptions at different time scales and corresponding time scales; the time series quantum state is feature reconstructed, and the formula for feature reconstruction is: ;in, represents the global quantum state, Represents the number of categories of normalized feature data, represents the quantum gate operation, Represents the first The composite quantum state of the class-normalized characteristic data, Represents the first The composite quantum state of the class-normalized characteristic data, Represents the first Class-normalized feature data and The quantum phase of the class-normalized characteristic data is classified; based on the time scale of the global quantum state, the global quantum state is counted into the blade state sequence table to obtain the blade damage quantum library.

[0013] Furthermore, the method of performing damage assessment on real-time blade modal data based on the blade damage quantum library includes: Collect real-time blade modal data. The data types of the real-time blade modal data and the blade modal data are consistent. Perform quantum mapping on the real-time blade modal data to obtain the real-time quantum state. Perform damage matching with the blade damage quantum library based on the real-time quantum state to obtain the timing label. Perform time matching with the blade damage quantum library based on the timing label. The blade state description in the blade damage quantum library with the same time scale as the timing label is the blade state.

[0014] A slicer blade wear monitoring system, comprising: Data acquisition module: collects blade modal data of the same time series and blade status description at the corresponding moment; Data processing module: including a data embedding unit and a feature extraction unit. The data embedding unit performs multi-scale spatial embedding on the blade modal data to obtain a multi-scale spatial graph. The feature extraction unit performs feature extraction on the multi-scale spatial graph to obtain modal space features. Feature mapping module: quantum mapping of modal space features to obtain characteristic quantum states, and building a blade damage quantum library based on the characteristic quantum states and blade state description; Damage matching module: collects real-time blade modal data, performs damage assessment on the real-time blade modal data based on the blade damage quantum library, obtains the blade status, and issues corresponding warnings based on the blade status.

[0015] The technical effects and advantages of the slicer blade wear monitoring method and system of the present invention are as follows: The present invention can capture subtle changes in the blade state with high precision through advanced technologies such as multi-scale spatial embedding and quantum mapping, which makes it possible to accurately identify different damage states of the blade, such as wear, cracks, fatigue, etc., at an early stage, avoiding problems that traditional methods cannot detect in time; through the fusion of multi-sensor data, it can provide more comprehensive and accurate blade state information, avoiding errors or omissions that may be caused by single sensor data; the blade damage characteristics are extracted through modal space tables and quantum space tables. This method can effectively capture the nonlinear relationship and complex dynamic characteristics of the data. Compared with traditional linear analysis methods, it can better identify potential wear trends and provide early warnings; through the matching of real-time data with the blade damage quantum library, the blade damage state can be evaluated in real time in actual work. This combination of real-time and accuracy can ensure that a warning is issued in time before the blade reaches the damage threshold, prevent serious damage during the cutting process, and ensure the continuity and safety of production; through automated multi-scale spatial embedding and quantum computing analysis, subjective misjudgments and misoperations in the traditional manual detection process are reduced, the system can automatically identify the state changes of the blade, reduce human intervention, and at the same time improve the accuracy and consistency of detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 A schematic diagram of a method for monitoring blade wear of a slicer according to the present invention; Figure 2 The figure is a schematic diagram of a slicer blade wear monitoring system of the present invention. DETAILED DESCRIPTION

[0017] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0018] Example 1 See also Figure 1 As shown, this embodiment provides a method for monitoring blade wear of a slicer, comprising: S1, collecting blade modal data of the same time series and blade state description at the corresponding moment; S2, performing multi-scale spatial embedding on the blade modal data to obtain a multi-scale spatial graph, performing feature extraction on the multi-scale spatial graph to obtain modal space features; S3, perform quantum mapping on the modal space characteristics to obtain characteristic quantum states, and construct a blade damage quantum library based on the characteristic quantum states and blade state description; S4, collecting real-time blade modal data, performing damage assessment on the real-time blade modal data based on the blade damage quantum library, obtaining the blade status, and issuing corresponding warnings based on the blade status; Blade modal data includes: vibration data, current data, acoustic data and temperature data; the blade status description is an overview of the blade status, such as: intact, cracked, worn and fatigue damaged, etc.; vibration data is obtained through a vibration sensor, and the vibration signal can reflect the blade's operating status, vibration mode, resonance phenomenon and the increase in vibration amplitude caused by wear; current data is obtained through a current sensor, and the slicer's current signal can reflect the load condition of the blade. When the blade is worn, the cutting efficiency decreases, and the blade may bear a greater cutting load, resulting in increased current fluctuations in the motor; acoustic data is obtained through a sound sensor. During the blade cutting process, due to blade wear, surface unevenness and changes in cutting force, the friction between the blade and the material will trigger a sound wave signal, and the frequency and amplitude changes of the sound wave can reflect the degree of wear of the blade; temperature data is obtained through a temperature sensor, and the temperature change of the slicer blade is an important indicator of wear. A worn blade usually causes an increase in blade temperature due to increased cutting force and friction.

[0019] The multi-scale spatial embedding methods for blade modal data include: A delay timing set is preset, and the delay timing set includes different delay time scales and delay dimensions. The delay timing set is traversed, and the timing of each type of data in the blade modal data is reconstructed based on each traversed delay time scale and delay dimension. The formula for timing reconstruction is: ;in, The representative delay time scale is , the delay dimension is The interval time series data below, represents the delay dimension, represents the delay time scale, Represents the starting time of the blade modal data, represents the number of delay time scales, represents the selection function, which is used to select the blade modal data of the corresponding time scale; the interval time series data obtained by time series reconstruction of each element in the delayed time series set is used as a time series space, and each data in the interval time series data is used as a spatial node in the time series space; the node connection threshold is preset, and the spatial nodes of the time series space are associated and evaluated. The formula for the association evaluation is: ;in, Representative The spatial nodes and The correlation degree of spatial nodes, Represents the data category in the spatial node, Representative The spatial node categories of data, Representative The spatial node categories of data, represents a control constant to avoid division by zero errors, Represents the index of the data category in the spatial node; based on the correlation between different spatial nodes, connect the spatial nodes whose correlation is greater than or equal to the node connection threshold, and use the correlation as the weight of the connection edge to obtain a temporal spatial graph; use the spatial node corresponding to the starting time in each temporal spatial graph as the connection hub of all temporal spatial graphs, use the connection hub as the connection point, use each temporal spatial graph as a spatial plane, connect different temporal spatial graphs to the same space, and obtain a multi-dimensional spatial graph; use the spatial nodes at the same time in different spatial planes as the plane connection points, merge nodes of different spatial planes based on the plane connection points, merge the same plane connection points into one spatial node, retain the connection relationship between spatial nodes in different spatial planes at the plane connection points, and obtain a multi-scale spatial graph; Assuming that the delay time scale of the traversal is 3, the delay dimension is 4, and the starting time of the blade modal data is 0, the interval time series data is: ;in, represents the blade modal data when the time scale is 0, and the subsequent ones are the blade modal data when the time scales are 3, 6 and 9 respectively.

[0020] Methods for extracting features from multi-scale spatial graphs include: Take each spatial node in the multiscale spatial graph as the initial feature point, initialize the multiscale set of each initial feature point to be empty, perform low-dimensional exploration on the initial feature point, find the spatial node with a direct edge connected to the initial exploration point in the multiscale spatial graph as a one-dimensional association point, all one-dimensional association points constitute a one-dimensional association set, and the one-dimensional association set is included in the multiscale set; preset the exploration dimension set and the multidimensional point set, traverse the exploration dimension set, take the result of each traversal as the total exploration step length, perform multi-dimensional exploration on the multiscale spatial graph based on the total exploration step length, take the initial feature point as the starting point, and take the starting point as the starting point. The one-dimensional association set is used as the selection set, and a spatial node is selected from the selection set as the next selection point by using a random selection algorithm, and the next selection point is counted into the multidimensional point set, and the next selection point is used as the new starting point to continue the selection. When the number of elements in the multidimensional point set is equal to the total exploration step, the selection is stopped, and the multidimensional point set at this time is output as the to-be-selected set. The to-be-selected set is connected to the detection, and when the last element in the to-be-selected set is the initial feature point, the to-be-selected set is added to the multi-scale set. When the last element in the to-be-selected set is not the initial feature point, the to-be-selected set is discarded, and the process is repeated until the multi-scale set no longer changes; Based on the multi-scale set, each initial feature point is evaluated for feature prominence. The formula for feature prominence evaluation is: ;in, represents the feature evaluation value, represents the size of the multi-scale set, represents the index of the element in the multi-scale collection, Represents the element length, Represents the first The first element The correlation between the spatial nodes and the initial feature points, Represents the first The first element The time scale of a spatial node, represents the time scale of the initial feature point, Represents the maximum value of the element length; based on the size of the feature evaluation value, the initial feature points are sorted from large to small, and the first N initial feature points are selected as the modal salient features. All modal salient features constitute the modal space features; The elements in the multi-scale set are all sets. The element length is the number of elements in the set corresponding to the element in the multi-scale set. The same element length means that the element is of the same dimension.

[0021] Ways to perform quantum mapping of modal space features include: A normalization algorithm is used to normalize each type of data in the modal space feature to obtain normalized feature data. Normalization can eliminate the deviation introduced by differences in sensor characteristics, ensure that data can be directly compared and fused in the same dimension, and perform modal analysis on the normalized feature data. The formula for modal analysis is: ;in, Representative Class-normalized feature data The parsed value at the moment, Representative Class-normalized feature data The value of the moment, represents the imaginary unit, represents pi, represents the scale factor, which is used to control the span of the modal analysis. Represents infinity; based on the analytical value, the normalized feature data is amplitude mapped, and the formula for amplitude mapping is: ;in, Representative Class-normalized feature data The characteristic amplitude at the moment; based on the analytical value, the normalized characteristic data is offset mapped, and the formula for offset mapping is: ;in, Representative Class-normalized feature data The characteristic offset at the moment; based on the characteristic amplitude and characteristic offset, a modal space table is constructed for each modal data, and the characteristic amplitude and characteristic offset of the normalized characteristic data at different moments are recorded in the modal space table; a quantum space table is constructed based on the modal space table, and the characteristic amplitude and characteristic offset of the normalized characteristic data in the modal space table at different moments are quantum constructed, and the formula for quantum construction is: ;in, Representative Class-normalized feature data The composite quantum state at time Represents the imaginary unit; quantum phase calculation is performed on the composite quantum state of different types of normalized characteristic data, and the formula for quantum phase calculation is: ;in, Representative Class-normalized feature data and Class-normalized feature data The quantum nature of moments, Representative Class-normalized feature data The complex conjugate of the composite quantum state at time instant. For a quantum state, its complex conjugate is to reverse the sign of the imaginary part of each complex coefficient. Representative Class-normalized feature data The composite quantum state at time Representative Class-normalized feature data The composite quantum state at time Representative Class-normalized feature data The complex conjugate of the composite quantum state at the same moment; the quantum phase and composite quantum state of different normalized characteristic data at the same moment constitute the characteristic quantum state; The use of quantum mapping and quantum phase calculation not only makes the blade wear monitoring method efficient, but also effectively reduces the computational complexity. Quantum computing provides a novel and efficient way in the construction of characteristic quantum states and the creation of blade damage libraries, which makes the damage assessment process more accurate while avoiding the performance bottleneck problems that may exist in traditional algorithms.

[0022] Ways to build a blade damage quantum library include: Based on the time scale corresponding to the characteristic quantum state, time sorting is performed to obtain the time series quantum state; based on the blade state description at the time corresponding to the time scale in the time series quantum state, a blade state sequence table is constructed, and the blade state sequence table contains blade state descriptions at different time scales and corresponding time scales; the time series quantum state is feature reconstructed, and the formula for feature reconstruction is: ;in, represents the global quantum state, Represents the number of categories of normalized feature data, Represents quantum gate operation. Common quantum gate operations include Hadamard gate and CNOT gate. Represents the first The composite quantum state of the class-normalized characteristic data, Represents the first The composite quantum state of the class-normalized characteristic data, Represents the first Class-normalized feature data and The quantum phase of the class-normalized characteristic data is classified; based on the time scale of the global quantum state, the global quantum state is counted into the blade state sequence table to obtain the blade damage quantum library.

[0023] The damage assessment methods for real-time blade modal data based on the blade damage quantum library include: Collect real-time blade modal data. The data types of real-time blade modal data and blade modal data are consistent. Perform quantum mapping on the real-time blade modal data to obtain the real-time quantum state. Perform damage matching based on the real-time quantum state and the blade damage quantum library. The damage matching formula is: ;in, represents a time series label, represents the quantum distance function. Common quantum distance functions include the inner product in Hilbert space, Represents the blade damage quantum library The global quantum state at time represents the real-time quantum state, Represents the time scale in the blade damage quantum library; based on the timing label, time matching is performed with the blade damage quantum library, and the blade state description with the same time scale as the timing label in the blade damage quantum library is the blade state.

[0024] This embodiment uses advanced technologies such as multi-scale spatial embedding and quantum mapping to capture subtle changes in the blade state with high precision, which makes it possible to accurately identify different damage states of the blade, such as wear, cracks, fatigue, etc., at an early stage, avoiding problems that traditional methods cannot detect in time. The fusion of multi-sensor data can provide more comprehensive and accurate blade state information, avoiding errors or omissions that may be caused by single sensor data. The blade damage characteristics are extracted through modal space tables and quantum space tables. This method can effectively capture the nonlinear relationship and complex dynamic characteristics of the data. Compared with traditional linear analysis methods, it can better identify potential wear trends and provide early warnings. By matching real-time data with the blade damage quantum library, the blade damage state can be evaluated in real time in actual work. This combination of real-time and accuracy can ensure that a warning is issued in time before the blade reaches the damage threshold, prevent serious damage during the cutting process, and ensure the continuity and safety of production. Through automated multi-scale spatial embedding and quantum computing analysis, subjective misjudgments and misoperations in the traditional manual detection process are reduced. The system can automatically identify the state changes of the blade, reduce human intervention, and improve the accuracy and consistency of detection.

[0025] Example 2 See also Figure 2 As shown, the part not described in detail in this embodiment can be seen from the description of embodiment 1, and a slicer blade wear monitoring system is provided, including: Data acquisition module: collects blade modal data of the same time series and blade status description at the corresponding moment; Data processing module: including a data embedding unit and a feature extraction unit. The data embedding unit performs multi-scale spatial embedding on the blade modal data to obtain a multi-scale spatial graph. The feature extraction unit performs feature extraction on the multi-scale spatial graph to obtain modal space features. Feature mapping module: quantum mapping of modal space features to obtain characteristic quantum states, and building a blade damage quantum library based on the characteristic quantum states and blade state description; Damage matching module: collects real-time blade modal data, performs damage assessment on the real-time blade modal data based on the blade damage quantum library, obtains the blade status, and issues corresponding warnings based on the blade status; The modules are connected to each other via wired and / or wireless means to achieve data transmission between modules.

[0026] Example 3 This embodiment discloses an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the operation mode of the above-mentioned slicer blade wear monitoring method is implemented.

[0027] Since the electronic device introduced in this embodiment is an electronic device used to implement a slicer blade wear monitoring method in the embodiment of the present application, based on the slicer blade wear monitoring method introduced in the embodiment of the present application, the technical personnel of the field can understand the specific implementation of the electronic device of the present embodiment and its various variations, so how the electronic device implements the method in the embodiment of the present application is not described in detail here. As long as the technical personnel of the field implement the electronic device used in the slicer blade wear monitoring method in the embodiment of the present application, it belongs to the scope of protection of the present application.

[0028] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters and thresholds in the formula are set by technicians in this field according to actual conditions.

[0029] The above is only a preferred embodiment of the present invention, and the protection scope of the present invention is not limited to the above embodiments. All technical solutions under the concept of the present invention belong to the protection scope of the present invention. It should be pointed out that for ordinary technical users in this technical field, some improvements and modifications without departing from the principle of the present invention should also be regarded as the protection scope of the present invention.

Claims

1. A method for monitoring blade wear of a slicer, characterized in that: include: S1, collecting blade modal data of the same time series and blade state description at the corresponding moment; S2, performing multi-scale spatial embedding on the blade modal data to obtain a multi-scale spatial graph, performing feature extraction on the multi-scale spatial graph to obtain modal space features; S3, perform quantum mapping on the modal space characteristics to obtain characteristic quantum states, and construct a blade damage quantum library based on the characteristic quantum states and blade state description; S4. Collect real-time blade modal data, perform damage assessment on the real-time blade modal data based on the blade damage quantum library, obtain the blade status, and issue a corresponding warning based on the blade status.

2. The method for monitoring the wear of a slicer blade according to claim 1, characterized in that: The blade modal data includes: vibration data, current data, acoustic data and temperature data.

3. The method for monitoring the wear of a slicer blade according to claim 2, characterized in that: The method of multi-scale spatial embedding of blade modal data includes: A delay timing set is preset, and the delay timing set includes different delay time scales and delay dimensions. The delay timing set is traversed, and the timing of each type of data in the blade modal data is reconstructed based on each traversed delay time scale and delay dimension. The formula for timing reconstruction is: ;in, The representative delay time scale is , the delay dimension is The interval time series data below, represents the delay dimension, represents the delay time scale, Represents the starting time of the blade modal data, represents the number of delay time scales, represents the selection function; the interval time series data obtained by reconstructing the time series of each element in the delayed time series set is used as a time series space, and each data in the interval time series data is used as a spatial node in the time series space; A node connection threshold is preset, and the spatial nodes of the time space are evaluated for association to obtain the degree of association; based on the degree of association between different spatial nodes, the spatial nodes with a degree of association greater than or equal to the node connection threshold are connected, and the degree of association is used as the weight of the connection edge to obtain a time space graph; the spatial node corresponding to the starting time in each time space graph is used as the connection hub of all time space graphs, the connection hub is used as the connection point, and each time space graph is used as a spatial plane, and different time space graphs are connected to the same space to obtain a multidimensional space graph; the spatial nodes at the same time in different spatial planes are used as plane connection points, and different spatial planes are merged based on the plane connection points. The same plane connection points are merged into one spatial node, and the connection relationship between spatial nodes in different spatial planes is retained to obtain a multi-scale space graph.

4. The method for monitoring the wear of a slicer blade according to claim 3, characterized in that: The formula for performing association evaluation on the spatial nodes in the time series space is: ;in, Representative The spatial nodes and The correlation degree of spatial nodes, Represents the data category in the spatial node, Representative The spatial node categories of data, Representative The spatial node categories of data, represents the control constant, Represents the index of the data category in the spatial node.

5. The method for monitoring slicer blade wear according to claim 4, characterized in that: The method of extracting features from the multi-scale spatial graph includes: Take each spatial node in the multiscale spatial graph as the initial feature point, initialize the multiscale set of each initial feature point to be empty, perform low-dimensional exploration on the initial feature point, find the spatial node with a direct edge connected to the initial exploration point in the multiscale spatial graph as a one-dimensional association point, all one-dimensional association points constitute a one-dimensional association set, and the one-dimensional association set is included in the multiscale set; preset the exploration dimension set and the multidimensional point set, traverse the exploration dimension set, take the result of each traversal as the total exploration step length, perform multi-dimensional exploration on the multiscale spatial graph based on the total exploration step length, take the initial feature point as the starting point, and take the starting point as the starting point. The one-dimensional association set is used as the selection set, and a spatial node is selected from the selection set as the next selection point by using a random selection algorithm, and the next selection point is counted into the multidimensional point set, and the next selection point is used as the new starting point to continue the selection. When the number of elements in the multidimensional point set is equal to the total exploration step, the selection is stopped, and the multidimensional point set at this time is output as the to-be-selected set. The to-be-selected set is connected to the detection, and when the last element in the to-be-selected set is the initial feature point, the to-be-selected set is added to the multi-scale set. When the last element in the to-be-selected set is not the initial feature point, the to-be-selected set is discarded, and the process is repeated until the multi-scale set no longer changes; Based on the multi-scale set, each initial feature point is evaluated for feature prominence. The formula for feature prominence evaluation is: ;in, represents the feature evaluation value, represents the size of the multi-scale set, represents the index of the element in the multi-scale collection, Represents the element length, Represents the first The first element The correlation between the spatial nodes and the initial feature points, Represents the first The first element The time scale of a spatial node, represents the time scale of the initial feature point, Represents the maximum value of the element length; based on the size of the feature evaluation value, the initial feature points are sorted from large to small, and the first N initial feature points are selected as the modal salient features. All modal salient features constitute the modal space features.

6. The method for monitoring blade wear of a slicer according to claim 5, characterized in that: The method of performing quantum mapping on the modal space characteristics includes: A normalization algorithm is used to normalize each type of data in the modal space feature to obtain normalized feature data, and modal analysis is performed on the normalized feature data. The formula for modal analysis is: ;in, Representative Class-normalized feature data The parsed value at the moment, Representative Class-normalized feature data The value of the moment, represents the imaginary unit, represents pi, represents the scale factor, represents infinity; based on the analytical value, the normalized feature data is amplitude mapped to obtain the feature amplitude; based on the analytical value, the normalized feature data is offset mapped to obtain the feature offset; based on the feature amplitude and feature offset, a modal space table is constructed for each modal data, and the modal space table records the feature amplitude and feature offset of the normalized feature data at different times; based on the modal space table, a quantum space table is constructed, and the feature amplitude and feature offset of the normalized feature data in the modal space table at different times are quantum constructed to obtain a composite quantum state; quantum phase calculation is performed on the composite quantum state of different types of normalized feature data to obtain quantum phase; the quantum phase and composite quantum state of different normalized feature data at the same time constitute a characteristic quantum state.

7. The method for monitoring blade wear of a slicer according to claim 6, characterized in that: The formula for calculating the quantum phase of the composite quantum state of different types of normalized characteristic data is: ;in, Representative Class-normalized feature data and Class-normalized feature data The quantum nature of moments, Representative Class-normalized feature data The complex conjugate of the composite quantum state at time instant, Representative Class-normalized feature data The composite quantum state at time Representative Class-normalized feature data The composite quantum state at time Representative Class-normalized feature data The complex conjugate of the composite quantum state at time instant.

8. The method for monitoring the wear of a slicer blade according to claim 7, characterized in that: The method of constructing the blade damage quantum library includes: Based on the time scale corresponding to the characteristic quantum state, time sorting is performed to obtain the time series quantum state; based on the blade state description at the time corresponding to the time scale in the time series quantum state, a blade state sequence table is constructed, and the blade state sequence table contains blade state descriptions at different time scales and corresponding time scales; the time series quantum state is feature reconstructed, and the formula for feature reconstruction is: ;in, represents the global quantum state, Represents the number of categories of normalized feature data, represents the quantum gate operation, Represents the first The composite quantum state of the class-normalized characteristic data, Represents the first The composite quantum state of the class-normalized characteristic data, Represents the first Class-normalized feature data and The quantum phase of the class-normalized characteristic data is classified; based on the time scale of the global quantum state, the global quantum state is counted into the blade state sequence table to obtain the blade damage quantum library.

9. The method for monitoring the wear of a slicer blade according to claim 8, characterized in that: The method of performing damage assessment on real-time blade modal data based on the blade damage quantum library includes: Collect real-time blade modal data. The data types of the real-time blade modal data and the blade modal data are consistent. Perform quantum mapping on the real-time blade modal data to obtain the real-time quantum state. Perform damage matching with the blade damage quantum library based on the real-time quantum state to obtain the timing label. Perform time matching with the blade damage quantum library based on the timing label. The blade state description in the blade damage quantum library with the same time scale as the timing label is the blade state.

10. A slicer blade wear monitoring system, used to implement the slicer blade wear monitoring method according to any one of claims 1 to 9, characterized in that: include: Data acquisition module: collects blade modal data of the same time series and blade status description at the corresponding moment; Feature extraction module: perform multi-scale spatial embedding on blade modal data to obtain a multi-scale spatial graph, perform feature extraction on the multi-scale spatial graph to obtain modal space features; Feature mapping module: quantum mapping of modal space features to obtain characteristic quantum states, and building a blade damage quantum library based on the characteristic quantum states and blade state description; Damage matching module: collects real-time blade modal data, performs damage assessment on the real-time blade modal data based on the blade damage quantum library, obtains the blade status, and issues corresponding warnings based on the blade status.

Images