A method and system for monitoring and adaptively calibrating the blade condition of a hawthorn slicer based on multi-sensor fusion

Abstract

This invention provides a method and system for monitoring and adaptively calibrating the blade condition of a hawthorn slicer based on multi-sensor fusion. It collects spindle motor current signals, blade holder vibration acceleration signals, and slice images, and extracts feature values. When the vibration feature value changes abruptly and the current feature value decreases simultaneously, blade breakage is detected, triggering an emergency stop. If breakage is not detected, the average current trend is monitored to generate a wear warning. Visual acquisition is initiated according to preset conditions to obtain the actual slice thickness. Using a pre-calibrated mapping relationship between slice thickness and current and vibration feature values, the threshold compensation amount is calculated in reverse based on the thickness deviation, adaptively correcting the detection threshold. This invention reduces the false blade breakage rate, achieves proactive wear warning, compensates for sensor drift, and ensures a high slice yield.

Description

Technical Field

[0001] This invention relates to the field of agricultural product processing machinery technology, specifically to a method and system for monitoring and adaptively calibrating the blade status of a hawthorn slicer based on multi-sensor fusion. Background Technology

[0002] Hawthorn, as an important agricultural product with both medicinal and edible uses, typically requires slicing during deep processing to prepare end products such as hawthorn slices, dried hawthorn, and hawthorn cake. The uniformity of slice thickness directly affects the drying efficiency, appearance, and yield of the finished product. The hawthorn slicing machine is the core equipment for this slicing process. During long-term, high-intensity operation, its blades are repeatedly impacted by hard foreign objects such as hawthorn pits contained in the hawthorn pulp, making them prone to wear, chipping, and even breakage. If these failures are not detected in time, they can lead to uneven slice thickness and increased scrap rates, or even secondary mechanical damage to critical components such as the blade holder and spindle, resulting in unplanned downtime and significant economic losses. Therefore, real-time and reliable monitoring of the operating status of the hawthorn slicing machine blades is crucial for ensuring the stable operation of the slicing production line.

[0003] Existing technologies for monitoring the condition of cutting tools in cutting equipment can be broadly categorized into two types. The first type is an indirect monitoring scheme based on mechanical structural deformation and electrical signals. Chinese patent application CN120628595A discloses an intelligent fault diagnosis system for a bamboo and wood cutting machine. This system uses multiple sensors arranged in a quadrilateral monitoring area formed by the spindle bearing, transmission gearbox, tool clamping end, and base support to collect structural deformation parameters such as displacement, clearance, and pressure. Combined with motor current and vibration signals, the system dynamically adjusts the current and vibration thresholds based on the hardness of the processed material. Furthermore, it uses geometrically derived parameters such as quadrilateral symmetry attenuation and angular displacement deviation, along with electromagnetic torque, to differentiate fault types and provide graded early warnings. This scheme relies entirely on indirect inference of equipment status based on mechanical vibration, structural deformation, and electrical quantities, without introducing any physical indicators that directly reflect the quality of processed products as closed-loop feedback. Furthermore, the established quadrilateral geometric model is highly dependent on the installation accuracy of the sensors and the rigidity of the machine structure, and the algorithm is quite complex. More importantly, its dynamic threshold adjustment is based solely on forward calculations of the hardness changes of the processed material, without compensating for the zero-point drift and sensitivity decay of the sensors themselves under high dust, high sap, and long-term operating conditions, causing the monitoring benchmark to gradually become distorted over time.

[0004] The second category is monitoring schemes based on direct observation of the cutting tool itself using machine vision. Chinese patent application CN118003385A discloses a food slicer with optical blade assessment. It incorporates an optical imaging device inside the slicer to directly image the outer cutting edge of the blade. Image processing algorithms such as edge detection and linear transformation are used to extract geometric features like the blade contour, reflective band width, and chipping. These features are then compared with pre-stored standard images of sharp blades to determine whether the blade needs sharpening or replacement. While this scheme can provide a relatively intuitive view of the blade's geometry, it only addresses the localized issue of blade sharpness. It doesn't monitor process parameters such as motor load and machine vibration during cutting, nor does it address the crucial product quality indicator of final slice thickness. Furthermore, given the generally harsh working conditions in food and agricultural product slicing environments, including splashes of juice, debris, and moisture, directly aiming the optical imaging device at the blade is prone to image failure due to surface contamination, resulting in poor engineering feasibility.

[0005] In summary, existing tool condition monitoring technologies have the following shortcomings: First, the monitoring methods are relatively simple. Traditional methods rely heavily on manual inspections or threshold judgments based on single current or vibration signals. In hawthorn slicing scenarios, it is difficult to distinguish between instantaneous load fluctuations caused by the hardness of the hawthorn pit and signal changes caused by actual tool breakage or wear, resulting in high false alarm and false negative rates. Second, effective self-calibration mechanisms are generally lacking. Vibration and current sensors are prone to zero-point drift and sensitivity reduction after long-term operation under high dust, high vibration, and high humidity slicing conditions. This causes the originally calibrated judgment thresholds to gradually deviate from the actual working conditions, making it difficult to guarantee the long-term reliability of the monitoring results. Third, existing solutions mostly focus on fault diagnosis at the equipment side, failing to incorporate indicators that directly reflect product quality, such as slice thickness, into the monitoring loop. It is impossible to reverse-check the sensor status with processing results, and even more difficult to adaptively compensate for current and vibration thresholds when sensor drift occurs. Fourth, most equipment still adopts a passive maintenance mode that only triggers shutdown after the tool completely breaks, failing to provide proactive warnings based on wear trends before tool breakage, which can easily lead to secondary equipment damage and material waste. Summary of the Invention

[0006] The purpose of this invention is to provide a method and system for monitoring and adaptively calibrating the blade condition of a hawthorn slicer based on multi-sensor fusion. While ensuring the reliability of real-time blade breakage detection, it also takes into account the early warning of blade wear trends and the benchmark stability of long-term sensor operation, and finally achieves closed-loop self-calibration of the monitoring system based on the slice thickness as a product quality indicator.

[0007] To achieve the above objectives, the present invention provides the following technical solution: A method for monitoring and adaptively calibrating the blade condition of a hawthorn slicer based on multi-sensor fusion includes the following steps: S1: Collect the real-time current signal of the spindle motor of the hawthorn slicer, the real-time vibration acceleration signal of the slicer blade holder or machine body, and the image of the hawthorn slice after being cut by the blade; S2: Perform time-domain analysis on the current signal and the vibration acceleration signal, and extract the current feature value of the current signal and the vibration feature value of the vibration acceleration signal respectively. When a sudden change in the vibration feature value is detected and the current feature value drops suddenly at the same time, it is determined that the tool is broken and an emergency stop command is triggered. S3: Under conditions where tool breakage is not determined, continuously monitor the trend of the average current value of the current signal with processing time. When the average current value shows an increasing trend with processing time and meets the preset wear judgment conditions, generate a tool wear warning. S4: Start visual acquisition according to the preset calibration trigger conditions, obtain the actual slice thickness from the hawthorn slice image, and calculate the thickness deviation between the actual slice thickness and the target slice thickness; S5: Based on the thickness deviation, through the pre-calibrated mapping relationship between the slice thickness and the current characteristic value and the vibration characteristic value, the threshold compensation amount of the current characteristic value and the vibration characteristic value is solved in reverse, and the threshold used to determine the tool breakage and tool wear in steps S2 and S3 is adaptively corrected with the threshold compensation amount.

[0008] Further: In step S1, before extracting feature values ​​from the current signal and the vibration acceleration signal, the current signal and the vibration acceleration signal are preprocessed. The preprocessing includes high-pass filtering and amplitude normalization. The high-pass filtering is used to filter out low-frequency interference components that are lower than the fundamental frequency of the slicer spindle.

[0009] Further: In step S2, the current characteristic value includes at least the mean current value of the current signal, and the vibration characteristic value includes at least one or more of the vibration variance, vibration peak factor and vibration kurtosis of the vibration acceleration signal; The sudden change in the vibration characteristic value means that the vibration variance, vibration peak factor or vibration kurtosis exceeds the corresponding sudden change threshold within a preset time window. The sudden drop in current characteristic value means that the decrease in the average current value within the preset time window relative to the average steady-state current value before the fracture determination exceeds the preset sudden drop threshold.

[0010] Further: In step S3, the wear determination condition includes at least one of the following: The rate of change of the average current over a unit processing time exceeds a preset rate of change threshold. The average current exceeds the absolute threshold corresponding to the rated current of the spindle motor. The current wear component, constructed from the difference between the mean current and the initial mean, and the vibration wear component, constructed from the difference between the vibration variance and the initial variance, are weighted according to a preset weight to obtain a comprehensive wear index that exceeds a preset wear threshold.

[0011] Furthermore, in step S3, before generating the tool wear warning, a process parameter stability judgment is also performed. Only when the fluctuation of the feed speed and cutting depth of the hawthorn slicer are within the preset steady-state range will the tool wear warning be triggered based on the changing trend of the average current.

[0012] Further: In step S4, the calibration trigger condition is any one or a combination of the following two: A visual acquisition is initiated after a preset processing time. A visual acquisition is initiated once a preset number of hawthorn slices have been processed.

[0013] Further: In step S5, the mapping relationship is pre-defined in the following manner: With the cutting tools in good working order and the process parameters stable, adjust the operating parameters of the hawthorn slicer to obtain multiple sets of different slice thicknesses. And simultaneously record the thickness of each group of slices. Corresponding average current and vibration variance The slice thickness was established through regression fitting. Regarding the average current and vibration variance Functional relationship This functional relationship is used as the mapping relationship between the current characteristic value, the vibration characteristic value and the slice thickness.

[0014] Further: In step S5, the reverse calculation of the threshold compensation amount and the threshold correction include: When the absolute value of the thickness deviation continuously exceeds the preset process tolerance, and the deviation of the current characteristic value and vibration characteristic value relative to the historical normal operating conditions exceeds the preset drift threshold, it is determined that sensor drift has occurred. According to the mapping relationship Using the thickness deviation as input, the drift compensation amount of the average current is obtained by reverse calculation. And the drift compensation amount of vibration variance ; With the drift compensation amount The threshold for determining the average current is adjusted based on the drift compensation amount. Correct the threshold for determining the vibration variance; After the correction is completed, new current signals, vibration acceleration signals and slice images are acquired, and the thickness deviation is calculated again. If the thickness deviation returns to the process tolerance, the calibration is determined to be complete. Otherwise, the above reverse solution and threshold correction process is repeated until the thickness deviation returns to the process tolerance.

[0015] Further: In step S4, the actual slice thickness is obtained by an industrial camera and a laser displacement sensor. The industrial camera is used to acquire two-dimensional images of hawthorn slices, and the laser displacement sensor is used to obtain the distance from the upper surface of the hawthorn slice to the reference plane. The average of the measured thicknesses of multiple hawthorn slices is taken as the actual slice thickness.

[0016] The present invention also provides a hawthorn slicer blade condition monitoring and adaptive calibration system for implementing the above method, characterized in that it includes: The current detection module is installed in the power supply circuit of the main shaft motor of the hawthorn slicer to collect real-time current signals; The vibration detection module is installed on the blade holder or body of the hawthorn slicer to collect real-time vibration acceleration signals; The visual inspection module is located downstream of the cutting station of the hawthorn slicer and is used to acquire images of hawthorn slices after they have been cut by the blade. The controller is communicatively connected to the current detection module, vibration detection module and vision detection module respectively. The controller is configured to execute steps S2, S3 and S5 in the method of any one of claims 1 to 9, and output an emergency stop command to the actuator of the hawthorn slicer when it is determined that the tool is broken.

[0017] Compared with the prior art, the present invention has the following advantages: First, it significantly reduces the false alarm rate of tool breakage during hawthorn slicing. This invention employs an orthogonal dual-modal fusion judgment mechanism in the tool breakage determination stage, combining "abrupt vibration characteristic value" and "sudden drop in current characteristic value." These two mechanisms correspond to the two physically contradictory phenomena that inevitably occur simultaneously: the sudden mechanical impact and the sudden disappearance of cutting load at the moment of tool breakage. However, the instantaneous condition of hawthorn pits impacting the cutting edge, a common occurrence in hawthorn processing, is essentially a momentary increase in cutting load rather than its disappearance, and therefore does not simultaneously meet the above two judgment conditions. Thus, it can effectively distinguish between genuine tool breakage events and load impacts caused by the hardness of hawthorn pits, overcoming the high false alarm rate of existing technologies using single current or single vibration signal threshold judgment methods, and significantly improving the reliability and robustness of tool breakage monitoring.

[0018] Secondly, it achieves proactive trend warning of tool wear, transforming passive downtime maintenance into proactive planned maintenance. Under normal operating conditions where tool breakage is not determined, this invention continuously tracks the trend of the average current changing with machining time. Combined with process parameter stability verification and an optional current-vibration weighted comprehensive wear index, it can identify the monotonically increasing trend of cutting resistance with increasing wear before the tool completely fails. This generates a wear warning in advance to prompt planned tool replacement, avoiding the unplanned downtime, secondary mechanical damage to the tool holder and spindle, and material waste caused by the common practice of "stopping only after complete tool breakage" in existing technologies, thus extending the effective service life of the tool.

[0019] Third, this invention solves the problem of monitoring benchmark failure caused by zero-point drift and sensitivity decay in process sensors during long-term operation in existing technologies. This invention uses slice thickness, a physical indicator directly reflecting product quality, as the "gold standard," and establishes a pre-calibrated functional mapping relationship between slice thickness and mean current and vibration variance. After periodically activating visual acquisition, the threshold drift compensation amount of current characteristic values ​​and vibration characteristic values ​​is calculated in reverse based on the deviation between the actual slice thickness and the target thickness. This is then used to adaptively correct the thresholds for tool breakage and wear detection, and calibration is completed through iterative convergence verification. This mechanism overcomes the limitations of existing dynamic threshold schemes that rely solely on material hardness for forward calculations and cannot compensate for sensor drift. It enables the monitoring system to maintain long-term stability of the judgment benchmark even under harsh slicing conditions involving high dust, high sap levels, and prolonged operation, fundamentally ensuring the long-term reliability of tool condition monitoring.

[0020] Fourth, a closed-loop assurance system from equipment status monitoring to product quality control has been constructed. This invention uses vibration and current signals as "process gatekeepers" with millisecond-level real-time response to protect against blade breakage and provide early warning of wear, and uses visual slice thickness as a periodically triggered "product arbitrator" to handle benchmark calibration. This forms a three-layer progressive architecture that combines fast and slow processing and coarse and fine processing. It not only ensures the safe operation of the equipment and the service life of the blades, but also directly ensures the consistency of hawthorn slice thickness through a closed-loop slice thickness monitoring system. This elevates the traditional simple equipment fault diagnosis to a comprehensive assurance system that takes into account both equipment health and product quality, ultimately maintaining a high yield rate and production stability in hawthorn slice processing.

[0021] Fifth, this invention employs a preprocessing method combining high-pass filtering and amplitude normalization for the process sensor, effectively filtering out low-frequency interference such as the spindle fundamental frequency; for visual thickness measurement, it uses a processing method that combines an industrial camera with a laser displacement sensor and averages multiple slices. Compared with existing visual solutions that directly observe the blade surface, this significantly reduces the impact of juice splashing and debris adhesion on image quality, making it more feasible for engineering and more suitable for the actual working conditions of hawthorn and similar agricultural product slicing production lines. Attached Figure Description

[0022] Figure 1 This is a schematic diagram of the overall structure of the hawthorn slicer tool condition monitoring and adaptive calibration system provided in an embodiment of the present invention; Figure 2 This is an overall flowchart of the hawthorn slicer tool condition monitoring and adaptive calibration method based on multi-sensor fusion provided in an embodiment of the present invention; Figure 3 This is a schematic diagram of the visual adaptive calibration closed loop provided in an embodiment of the present invention. Detailed Implementation

[0023] The technical solution of the present invention will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. 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.

[0024] In the description of this invention, it should be noted that the terms "center," "upper," "lower," "left," "right," "vertical," "horizontal," "inner," and "outer," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are used only for the convenience of describing the invention and for simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on the invention. Furthermore, the terms "first," "second," and "third" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0025] This invention provides a method for monitoring and adaptively calibrating the blade condition of a hawthorn slicer based on multi-sensor fusion. Applied to intelligent monitoring and adaptive calibration of the blade condition in a hawthorn slicer, it aims to solve problems such as insufficient reliability in real-time blade breakage detection, difficulty in providing trend-based early warning of blade wear, and zero-point drift of sensors leading to monitoring benchmark failure during long-term operation by fusing sensor information from vibration, current, and vision modes. The entire method consists of five main stages: multi-source data acquisition and preprocessing, real-time blade breakage detection, blade wear trend analysis, periodic calibration of the vision sensor, and production stability assurance. These stages are described in detail below.

[0026] Regarding the hardware deployment for implementing this method, it is based on a hawthorn slicer tool condition monitoring and adaptive calibration system. This system includes a current detection module, a vibration detection module, a vision inspection module, a controller, a computer, and a slicing device. The current detection module is connected in series to the power supply circuit of the hawthorn slicer spindle motor, typically using a Hall current sensor or a Rogowski coil, to collect the current signal during the spindle motor's operation in real time. The vibration detection module typically uses a piezoelectric accelerometer, fixed to the slicer's tool holder or near the spindle bearing housing using a magnetic base or bolts, to collect the vibration acceleration signal of the tool during the cutting process. The vision inspection module is located downstream of the cutting station of the hawthorn slicer, above or to one side of the conveying channel for the sliced ​​hawthorn slices, to collect images of the sliced ​​hawthorn slices after cutting. The vision inspection module further includes an industrial camera and a laser displacement sensor. The industrial camera is used to acquire two-dimensional images of the hawthorn slices to identify the slice outline, and the laser displacement sensor is used to measure the distance between the upper surface of the hawthorn slice and a preset reference plane to obtain the slice thickness. The controller typically employs an industrial-grade programmable logic controller (PLC) or a dedicated embedded control unit. It communicates with the current detection module, vibration detection module, and vision inspection module to receive data from these three types of sensors and perform calculations such as blade breakage detection, wear detection, visual calibration, and threshold correction as described in this method. The controller is electrically connected to the actuator of the hawthorn slicer to output an emergency stop command when blade breakage is detected. The computer communicates with the vision inspection module and the controller to perform computationally intensive tasks such as image processing, mapping calibration, and data storage. Furthermore, in the pitting process, the controller can also be electrically connected to the pitting device to collaboratively complete the pitting and slicing processes of the hawthorn.

[0027] In the multi-source data acquisition and preprocessing stage, the real-time current signal of the spindle motor is denoted as... The real-time vibration acceleration signal of the tool holder or machine body is recorded as Images of hawthorn slices obtained after the hawthorn feed is cut are periodically acquired by the vision inspection module. The sampling frequency of the current signal and vibration acceleration signal should match the bandwidth of their respective signals. Those skilled in the art can determine a reasonable sampling frequency based on the rotational speed and number of cutters of the specific equipment. Before extracting feature values ​​from the acquired current signal and vibration acceleration signal, they are first preprocessed. The preprocessing includes high-pass filtering and amplitude normalization. High-pass filtering is used to filter out low-frequency interference components below the fundamental frequency of the slicer spindle, such as power grid frequency interference and equipment resonance, to improve the signal-to-noise ratio of subsequent time-domain feature extraction. Amplitude normalization is used to eliminate the differences in amplitude dimensions between different sensors and different signals, which facilitates subsequent multi-feature fusion judgment.

[0028] In the real-time tool fracture detection stage, time-domain analysis is performed on the preprocessed current signal and vibration acceleration signal to extract current feature values ​​and vibration feature values, respectively. The current feature values ​​include at least the mean current value of the current signal. The vibration characteristic values ​​include at least the vibration variance of the vibration acceleration signal. One or more of the following: peak vibration factor and kurtosis. The mean current reflects the DC component of the signal, and its calculation method is to apply the current within a window... sampling points Summing and dividing by ,Right now:

[0029] Vibration variance reflects the intensity of signal fluctuations and is the average of the squared deviations of each sampling point within the window from the mean, i.e.:

[0030] Peak factor is the peak value of the signal. With root mean square value The ratio is used to characterize the sudden impact component in a signal, namely:

[0031] Where the root mean square value The formula for calculation is:

[0032] The vibration kurtosis is the ratio of the fourth-order central moment to the square of the variance, and is used to measure the sharpness of the signal distribution, i.e.:

[0033] Furthermore, to further characterize the signal energy features, time-domain energy and frequency-domain energy can also be calculated. Time-domain energy is the sum of the squares of the signal amplitude within a window, i.e.:

[0034] Frequency domain energy is converted to the frequency domain by performing a Fast Fourier Transform on the time domain signal, and then the cutting frequency is... and its second harmonic Third harmonic Amplitude at corresponding frequency points in a specific frequency band We obtain the result by summing the squares of the two terms, i.e.:

[0035] Tool wear leads to a decrease in the proportion of energy in the cutting main frequency and an increase in the proportion of harmonic energy. This characteristic can be used as an auxiliary criterion for judgment.

[0036] In the actual tool breakage detection process, when a sudden change in vibration characteristic value and a simultaneous sharp drop in current characteristic value are detected, tool breakage is determined, and an emergency stop command is immediately triggered. Specifically, a sudden change in vibration characteristic value means that the vibration variance, vibration peak factor, or vibration kurtosis exceeds the corresponding change threshold within a preset time window. This change threshold can be determined based on experimental data, for example, by taking a number of times the steady-state mean of the characteristic value under healthy conditions. A sharp drop in current characteristic value means that the mean current value within the preset time window is relative to the steady-state mean current value before the breakage determination. The magnitude of the decrease satisfies ,in The threshold for sudden drop is preset. The inherent physical logic of this dual-modal judgment mechanism is as follows: the moment the tool breaks, it is inevitably accompanied by a sudden change in mechanical vibration caused by the impact of the tool body breaking and a sudden drop in current caused by the instantaneous disappearance of the cutting load. The two occur simultaneously. However, the interference condition of the hawthorn pit hitting the cutting edge, which is common in hawthorn processing, is physically a sudden increase in the cutting load rather than its disappearance. It is manifested as an instantaneous increase in current rather than a sudden drop. Therefore, it will not meet the above two judgment conditions at the same time, thus effectively avoiding misjudging the impact of the hawthorn pit as tool breakage.

[0037] In the tool wear trend analysis, under normal operating conditions where tool breakage is not identified, the system continuously monitors the change trend of the average current signal over machining time. When the tool wears, the cutting edge becomes blunt, cutting resistance increases, and the spindle motor will draw a larger current to maintain the set speed, thus increasing the average current. With processing time It exhibits a monotonically increasing trend, that is... Wear determination criteria can be achieved in at least one of the following ways: one is to determine the rate of change of trend, that is, to require that the rate of change of the average current over a unit processing time exceeds a preset rate of change threshold. ,Right now:

[0038] Second, the absolute threshold determination, that is, when the average current exceeds the absolute threshold corresponding to the rated current of the spindle motor. When an alert is triggered, that is:

[0039] Third, multi-feature fusion judgment, namely, based on the average current. Compared with the initial mean The difference in the construction current wear component Due to vibration variance With initial variance The difference in the construction of vibration wear components Their definitions are as follows:

[0040] Then according to the preset weight and Weighted average wear index ,Right now:

[0041] when ( An alert is triggered when the wear threshold is preset, where the weight is... and satisfy And determined through experimental data calibration, for example, the following can be taken. To highlight the dominant position of current trends.

[0042] To further prevent false alarms in tool wear warnings, the system also performs a stability check on process parameters before generating the warning. This is because machining parameters such as feed rate... Depth of cut Changes in the feed rate and depth of cut can also cause an increase in spindle current. If not distinguished, these fluctuations can easily be misinterpreted as tool wear. Therefore, this method only triggers a tool wear warning based on the trend of the average current when the fluctuations in both feed rate and depth of cut are within a preset steady-state range. In other words, the following conditions must be met simultaneously:

[0043] in , These are the set values ​​for feed rate and depth of cut, respectively. , This sets the corresponding upper limit for fluctuations. Furthermore, the system can also establish and store the average current during machining with a new tool under healthy tool conditions. and vibration variance As a health baseline, the deviation between the current signal and the baseline is compared in real time to improve the accuracy of wear assessment under the influence of various equipment aging factors.

[0044] During the periodic calibration of the vision sensor, the system initiates visual acquisition according to preset calibration trigger conditions. The calibration trigger conditions are any one or a combination of the following two: one is after a preset processing time. That is, to initiate a visual acquisition, for example, to take... Second, for each cumulative completed preset quantity The processing of hawthorn slices triggers a visual acquisition. This periodic triggering mechanism allows the vision module to participate in the monitoring loop as a benchmark calibration method without increasing the burden on the regular process.

[0045] After the vision acquisition is initiated, the industrial camera images the cut hawthorn slices, while the laser displacement sensor simultaneously measures the... Distance from the top surface of a sliced ​​hawthorn to a preset reference plane This yields the actual thickness measurement of each hawthorn slice; the average of the measured thicknesses of multiple hawthorn slices is then used as the actual slice thickness. ,Right now:

[0046] in The number of slices to be averaged; then compared with the target slice thickness set in the process. Compare and calculate the thickness deviation:

[0047] This multi-slice averaging method can significantly reduce random errors caused by surface irregularities or environmental disturbances in individual slices, while also reducing the impact of on-site splattering of juice and debris on the results of a single measurement.

[0048] The core of the visual calibration process lies in using thickness deviation. The drift compensation for current and vibration threshold is calculated in reverse. To achieve this reverse calculation, the mapping relationship needs to be calibrated beforehand under the condition that the tool is in good condition and the process parameters are stable: multiple sets of different slice thicknesses are obtained by adjusting the operating parameters of the hawthorn slicer. And simultaneously record the thickness of each group of slices. Corresponding average current and vibration variance The slice thickness was established by multivariate regression fitting. Regarding the average current and vibration variance Functional relationship:

[0049] The simplest form of this functional relationship can be expressed using linear regression, that is:

[0050] in The calibration coefficients are obtained by least-squares fitting of multiple sets of experimental data. For working conditions with strong nonlinearity, polynomial regression or support vector regression can also be used to achieve this. The functional relationship is used as the mapping relationship between current characteristic values, vibration characteristic values ​​and slice thickness.

[0051] In the actual calibration process, sensor drift identification is performed first: if there is a thickness deviation... The absolute value continuously exceeds the preset process tolerance (For example, take) At the same time, the deviations of the current characteristic value and vibration characteristic value from the historical normal operating conditions exceed the preset drift threshold, that is, the following conditions are met:

[0052] in This is the historical steady-state current average. For example, to preset the drift threshold. If this is detected, it is determined that sensor drift has occurred, such as zero-point drift of a current sensor or decreased sensitivity of a vibration sensor. Then, based on the mapping relationship... With the thickness deviation Using this as input, the drift compensation amount of the mean current is obtained by inverse calculation. And the drift compensation amount of vibration variance When solving, the thickness deviation can be distributed according to a pre-defined partial derivative relationship. and The approach based on two independent variables, namely, the first-order Taylor expansion:

[0053] Alternatively, one of the historical drift values ​​can be fixed. Inverse solution In the form of a linear mapping, the following methods exist:

[0054] Then use the drift compensation amount The threshold for determining the average current is adjusted based on the drift compensation amount. The threshold for determining the vibration variance is corrected as follows:

[0055] in , These are the threshold values ​​for determining the mean current and the vibration variance before correction, respectively. , The corrected threshold is then applied to the subsequent tool breakage and wear determination processes.

[0056] After the correction is completed, the system continues to acquire new current signals, vibration acceleration signals, and slice images to recalculate the thickness deviation. And determine whether it has returned to within the process tolerance: If If the calibration is successful, the system will resume normal monitoring procedures; otherwise, the thickness deviation has not yet converged. Then repeat the above reverse solution and threshold correction process until... This enables adaptive iterative calibration of the monitoring threshold.

[0057] In ensuring production stability, the aforementioned real-time blade breakage detection, wear trend analysis, and periodic visual calibration together form a complete monitoring and calibration closed loop. Vibration and current signals, as millisecond-level process gatekeepers, are responsible for blade breakage protection and wear warnings, while visual slice thickness, as a periodically triggered product assessment method, is responsible for benchmark calibration. The two, one fast and one slow, one coarse and one precise, work together to continuously maintain the reliability of blade condition monitoring and the consistency of slice thickness throughout the hawthorn slicing process, thereby ensuring long-term production stability and yield.

[0058] The implementation process of this method is further illustrated below with a specific embodiment. In this embodiment, the real-time current signal of the spindle motor is acquired during the multi-source data acquisition stage. Real-time vibration acceleration signal of the tool holder The system monitors real-time image data of the hawthorn feed; during the fracture determination stage, time-domain analysis is performed on the current and vibration signals to calculate their energy characteristic values. When a sudden change in the vibration characteristic value and a sharp drop in the current characteristic value are detected, the tool is determined to be fractured and an emergency stop command is triggered; during the wear determination stage, the root mean square value of the current signal is monitored. ,like The value continues to rise above the preset threshold. (Right now If the wear rate is 100%, it is determined to be tool wear, the wear amount is recorded, and replacement is prompted; during the vision calibration phase, the tool is checked every preset time interval. (For example The vision sensor is activated to acquire standard slice images, and the slice thickness deviation is analyzed based on the images. The system automatically adjusts the judgment thresholds of the vibration and current sensors according to the reverse solution process described above, achieving adaptive calibration. Through the above process, this embodiment can achieve the technical effects of improving detection reliability, extending tool life, and ensuring long-term monitoring stability, thereby guaranteeing the yield of hawthorn slices.

[0059] It should be noted that when this method is applied to non-hawthorn agricultural products or other materials that require slicing, it is only necessary to re-map the relationships according to the actual characteristics of the materials. By calibrating and setting relevant thresholds, tool status monitoring and adaptive calibration under corresponding working conditions can be achieved. Therefore, the application scenarios of this invention are not limited to hawthorn slices, and similar modifications that can be made by those skilled in the art should fall within the protection scope of this invention.

[0060] The above embodiments are only for illustrating the technical concept and features of the present invention, and are intended to enable those skilled in the art to understand the content of the present invention and implement it accordingly. They should not be construed as limiting the scope of protection of the present invention. All equivalent transformations or modifications made in accordance with the spirit and essence of the present invention should be covered within the scope of protection of the present invention.

Claims

1. A method for monitoring and adaptively calibrating the blade condition of a hawthorn slicer based on multi-sensor fusion, characterized in that, Includes the following steps: S1: Collect the real-time current signal of the spindle motor of the hawthorn slicer, the real-time vibration acceleration signal of the slicer blade holder or machine body, and the image of the hawthorn slice after being cut by the blade; S2: Perform time-domain analysis on the current signal and the vibration acceleration signal, and extract the current feature value of the current signal and the vibration feature value of the vibration acceleration signal respectively. When a sudden change in the vibration feature value is detected and the current feature value drops suddenly at the same time, it is determined that the tool is broken and an emergency stop command is triggered. S3: Under conditions where tool breakage is not determined, continuously monitor the trend of the average current value of the current signal with processing time. When the average current value shows an increasing trend with processing time and meets the preset wear judgment conditions, generate a tool wear warning. S4: Start visual acquisition according to the preset calibration trigger conditions, obtain the actual slice thickness from the hawthorn slice image, and calculate the thickness deviation between the actual slice thickness and the target slice thickness; S5: Based on the thickness deviation, through the pre-calibrated mapping relationship between the slice thickness and the current characteristic value and the vibration characteristic value, the threshold compensation amount of the current characteristic value and the vibration characteristic value is solved in reverse, and the threshold used to determine the tool breakage and tool wear in steps S2 and S3 is adaptively corrected with the threshold compensation amount.

2. The method for monitoring and adaptive calibration of hawthorn slicer blade status based on multi-sensor fusion according to claim 1, characterized in that, In step S1, before extracting feature values ​​from the current signal and the vibration acceleration signal, the current signal and the vibration acceleration signal are preprocessed. The preprocessing includes high-pass filtering and amplitude normalization. The high-pass filtering is used to filter out low-frequency interference components that are lower than the fundamental frequency of the slicer spindle.

3. The method for monitoring and adaptively calibrating the cutting tool status of a hawthorn slicer based on multi-sensor fusion according to claim 1, characterized in that, In step S2, the current characteristic value includes at least the mean current value of the current signal, and the vibration characteristic value includes at least one or more of the vibration variance, vibration peak factor, and vibration kurtosis of the vibration acceleration signal. The sudden change in the vibration characteristic value means that the vibration variance, vibration peak factor or vibration kurtosis exceeds the corresponding sudden change threshold within a preset time window. The sudden drop in current characteristic value means that the decrease in the average current value within the preset time window relative to the average steady-state current value before the fracture determination exceeds the preset sudden drop threshold.

4. The method for monitoring and adaptively calibrating the cutting tool status of a hawthorn slicer based on multi-sensor fusion according to claim 1, characterized in that, In step S3, the wear determination criteria include at least one of the following: The rate of change of the average current over a unit processing time exceeds a preset rate of change threshold. The average current exceeds the absolute threshold corresponding to the rated current of the spindle motor. The current wear component, constructed from the difference between the mean current and the initial mean, and the vibration wear component, constructed from the difference between the vibration variance and the initial variance, are weighted according to a preset weight to obtain a comprehensive wear index that exceeds a preset wear threshold.

5. The method for monitoring and adaptively calibrating the cutting tool status of a hawthorn slicer based on multi-sensor fusion according to claim 1, characterized in that, In step S3, before generating the tool wear warning, the stability of process parameters is determined. The tool wear warning is triggered only when the fluctuation of the feed speed and cutting depth of the hawthorn slicer are within the preset steady-state range, based on the changing trend of the average current.

6. The method for monitoring and adaptive calibration of hawthorn slicer blade status based on multi-sensor fusion according to claim 1, characterized in that, In step S4, the calibration trigger condition is any one or a combination of the following two: A visual acquisition is initiated after a preset processing time. A visual acquisition is initiated once a preset number of hawthorn slices have been processed.

7. The method for monitoring and adaptively calibrating the blade condition of a hawthorn slicer based on multi-sensor fusion according to claim 1, characterized in that, In step S5, the mapping relationship is pre-defined in the following way: With the cutting tools in good working order and the process parameters stable, adjust the operating parameters of the hawthorn slicer to obtain multiple sets of different slice thicknesses. And simultaneously record the thickness of each group of slices. Corresponding average current and vibration variance The slice thickness was established through regression fitting. Regarding the average current and vibration variance Functional relationship This functional relationship is used as the mapping relationship between the current characteristic value, the vibration characteristic value and the slice thickness.

8. The method for monitoring and adaptive calibration of hawthorn slicer blade status based on multi-sensor fusion according to claim 7, characterized in that, In step S5, the reverse calculation of the threshold compensation amount and the threshold correction include: When the absolute value of the thickness deviation continuously exceeds the preset process tolerance, and the deviation of the current characteristic value and vibration characteristic value relative to the historical normal operating conditions exceeds the preset drift threshold, sensor drift is determined to have occurred. According to the mapping relationship Using the thickness deviation as input, the drift compensation amount of the average current is obtained by reverse calculation. And the drift compensation amount of vibration variance ; With the drift compensation amount The threshold for determining the average current is adjusted based on the drift compensation amount. Correct the threshold for determining the vibration variance; After the correction is completed, new current signals, vibration acceleration signals and slice images are acquired, and the thickness deviation is calculated again. If the thickness deviation returns to the process tolerance, the calibration is determined to be complete. Otherwise, the above reverse solution and threshold correction process is repeated until the thickness deviation returns to the process tolerance.

9. The method for monitoring and adaptively calibrating the cutting tool status of a hawthorn slicer based on multi-sensor fusion according to claim 1, characterized in that, In step S4, the actual slice thickness is obtained by an industrial camera and a laser displacement sensor. The industrial camera is used to acquire two-dimensional images of hawthorn slices, and the laser displacement sensor is used to obtain the distance from the upper surface of the hawthorn slice to the reference plane. The average of the measured thicknesses of multiple hawthorn slices is taken as the actual slice thickness.

10. A hawthorn slicer blade condition monitoring and adaptive calibration system for implementing the method according to any one of claims 1 to 9, characterized in that, include: The current detection module is installed in the power supply circuit of the main shaft motor of the hawthorn slicer to collect real-time current signals; The vibration detection module is installed on the blade holder or body of the hawthorn slicer to collect real-time vibration acceleration signals; The visual inspection module is located downstream of the cutting station of the hawthorn slicer and is used to acquire images of hawthorn slices after they have been cut by the blade. The controller is communicatively connected to the current detection module, vibration detection module and vision detection module respectively. The controller is configured to execute steps S2, S3 and S5 in the method of any one of claims 1 to 9, and output an emergency stop command to the actuator of the hawthorn slicer when it is determined that the tool is broken.

Images