A non-contact laser online measurement method for the thickness of corrugated board

Abstract

The present application relates to the field of industrial automation detection, and particularly relates to a non-contact laser online measurement method for the thickness of corrugated board. The method comprises: collecting the vertical distance signal of the surface of the corrugated board and the real-time pulse signal of the conveying belt, and constructing the equidistant spatial domain sequence; calculating the instantaneous gradient and local energy features of the sequence, and then calculating the separation coefficient by using the nonlinear classification function and updating the vibration reference line by using the recursive filtering, so as to obtain the geometric thickness; based on the viscoelastic mechanics model, the vertical acceleration is calculated by using the second-order difference of the vibration reference line, and the compression deformation variable caused by the inertial force is calculated by combining the surface density of the corrugated board and the dynamic compression modulus; and the geometric thickness and the compression deformation variable are superimposed to obtain the real physical thickness. The present application analyzes the vibration and texture overlapped in the frequency domain through phase space projection, and compensates the dynamic compression error from the physical level, so as to improve the measurement accuracy under high-speed working conditions.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of industrial automation detection, and in particular to a non-contact laser online measurement method for the thickness of corrugated paperboard. BACKGROUND

[0002] In the production process of corrugated paperboard, the thickness of the paperboard is a core indicator for measuring product quality, which directly determines the compression strength and cushioning performance of the final carton; in order to improve production efficiency, modern corrugated paperboard production lines usually run at a speed of 150 meters to 300 meters per minute, and there are complex working conditions such as mechanical vibration, which poses a challenge to the online measurement of the thickness of corrugated paperboard.

[0003] The existing online measurement technology for the thickness of corrugated paperboard mainly has the following two types of mainstream solutions: the first type is contact measurement, which uses a roller to press on the surface of the paperboard for measurement. This method has the risk of causing physical indentation on the paperboard, and the mechanical wear is fast when running at high speed; the second type is non-contact measurement based on the laser triangulation reflection principle, which avoids physical damage, but in practical application, it faces two technical problems:

[0004] The first is the difficulty of signal separation caused by frequency domain overlap; the mechanical vibration frequency of the conveyor belt is usually in the range of 10 to 20 hertz, and the texture characteristics of the surface of the corrugated paperboard also exhibit frequency signals of dozens of hertz under high-speed motion, and the two are overlapped in the frequency domain, and the conventional low-pass filter is difficult to distinguish; if the texture details are retained, it is difficult to remove the vibration noise, and if the vibration is removed, the corrugated paperboard peak may be smoothed out, resulting in a smaller measurement value.

[0005] The second is the physical error caused by dynamic compression; the existing technology usually regards the corrugated paperboard as a rigid body, considering that the distance change measured by the laser sensor is only the movement of the position; however, the corrugated paperboard has the characteristics of viscoelastic material, when the conveyor belt vibrates, the vertical acceleration will generate an inertial force, under the action of the force, the corrugated paperboard will be micron-level elastically compressed, which not only changes the spatial position of the corrugated paperboard, but also changes the physical thickness of the corrugated paperboard; the existing technology only deducts the displacement without compensating for the compression amount, resulting in a measurement result lower than the static true value in the strong vibration section, affecting the accuracy of quality control. SUMMARY

[0006] In order to solve the problems of existing technology that vibration noise is difficult to separate due to frequency domain overlap, and dynamic compression error cannot be compensated due to not considering the viscoelasticity of the material, the present application provides a non-contact laser online measurement method for the thickness of corrugated paperboard, which comprises:

[0007] The distance time sequence of the corrugated board surface and the real-time pulse sequence of the conveying belt are collected by using a preset distance acquisition unit and a displacement recording unit, the real-time pulse sequence is used for interpolation processing of the distance time sequence to obtain a spatial domain sequence, the spatial domain sequence is composed of a series of spatial discrete sampling points, and each sampling point has a distance measurement value; the instantaneous gradient feature and the local energy feature of the sampling point are calculated based on the distance measurement value, and then a separation coefficient of the sampling point is calculated by using a preset nonlinear classification function; a recursive filter is constructed based on the separation coefficient to iteratively update the vibration reference line of each sampling point, and then the geometric thickness of each sampling point is obtained; based on a preset viscoelastic dynamics model, the vertical acceleration of each sampling point is calculated by using the vibration reference line, and in combination with a preset dynamic compression modulus, the compression deformation variable of each sampling point under the action of inertial force is obtained; the geometric thickness is compensated by using the compression deformation variable to obtain the real physical thickness of each sampling point.

[0008] The present application is different from the conventional frequency domain filtering method, selects to construct a phase space in the time domain, realizes vibration separation by using the difference between the gradient and the energy, aims to make the measurement data correspond to the physical position of the corrugated board, reduces the systematic error caused by high-speed vibration, and improves the measurement accuracy.

[0009] Further, the interpolation processing specifically includes: counting the total number of pulses output by the displacement recording unit, converting the total number of pulses into the real-time transmission distance of the conveying belt relative to the starting position according to the preset pulse equivalent; setting a fixed spatial sampling interval; performing linear interpolation operation on the distance time sequence in the coordinate system defined by the real-time transmission distance to obtain the spatial domain sequence.

[0010] The present application solves the problem of uneven signal distribution under the variable frequency production line by converting the time-based non-uniform sampling into uniform sampling based on physical distance through constructing an equidistant spatial domain mapping, and ensures the physical consistency of subsequent calculation.

[0011] Further, the calculation of the instantaneous gradient feature and the local energy feature of the sampling point specifically includes: for a certain sampling point, the absolute value of the difference between the distance measurement value of the sampling point and the distance measurement value of the previous sampling point is calculated to obtain the instantaneous gradient feature of the sampling point; a sliding window is preset, the length of the sliding window covers the physical distance of a corrugated peak and a corrugated valley of the corrugated board; for a certain sampling point, the sliding window centered on the sampling point is selected, and the root mean square value of the distance measurement value in the sliding window is calculated to obtain the local energy feature of the sampling point.

[0012] Further, if the serial number of the sampling point in the spatial domain sequence is represented as n, then the first sampling point is n = 1, the second sampling point is n = 2, and so on. ​ The separation coefficients of the sampling points satisfy the following relationship:

[0013]

[0014] in, For the first The separation coefficient of each of the sampling points. For boundary sharpness factor, For the first The instantaneous gradient features of the sampling points For the first The local energy characteristics of the sampling points. The preset normalized gradient threshold, The preset normalized energy threshold, It is a natural exponential function.

[0015] This invention utilizes the nonlinear saturation characteristics of the Sigmoid function to construct a relational formula. Compared with the traditional method of judging by a fixed threshold, this invention can adaptively adjust in areas where texture features such as warping and vibration of corrugated cardboard are mixed, thereby ensuring the accuracy of measurement under complex working conditions.

[0016] Further, the iterative update of the vibration baseline of each sampling point specifically includes: for a sampling point, setting the separation coefficient of the sampling point as a weight, and performing a weighted summation of the vibration baseline of the previous sampling point and the distance measurement value of the sampling point to obtain the vibration baseline of the sampling point.

[0017] This invention implements an adaptive vibration baseline tracking strategy by constructing a recursive filter based on the separation coefficient, which avoids measurement errors and noise residue caused by traditional methods and improves measurement accuracy.

[0018] Furthermore, if using The index of the sampling point in the spatial domain sequence is represented by the number of the sampling point. The actual physical thickness of each of the sampling points satisfies the following relationship:

[0019]

[0020] in, For the first The actual physical thickness of each of the aforementioned sampling points. For the first The distance measurement value of each of the sampling points For the first The vibration baseline of the sampling points. a face density of the corrugated paperboard, a vertical acceleration of the first a vertical acceleration of the first a dynamic compressive modulus.

[0021] The present application comprehensively considers the deformation of corrugated paperboard as a viscoelastic material on the conveyor belt, constructs a compensation relationship including an inertial force term and a dynamic modulus term, and restores the thickness of the corrugated paperboard elastically compressed due to inertial force.

[0022] Further, the vertical acceleration is calculated by the second-order difference of the vibration reference line combined with the speed of the conveyor belt.

[0023] Further, the boundary sharpness factor is obtained through a phase space boundary continuous calibration experiment.

[0024] Further, the dynamic compressive modulus is comprehensively calibrated through a true value reverse fitting experiment.

[0025] Further, the distance acquisition unit is a high-frequency line laser profiler, and the displacement recording unit is an incremental encoder installed on the drive shaft of the conveyor belt.

[0026] The present application has the following beneficial effects:

[0027] Compared with the traditional frequency domain filtering, the present application constructs an online measurement system combining edge computing, phase space morphological analysis and rheological mechanics principles, and ensures that the influence of interference can still be inhibited in the case of overlapping of the vibration frequency and the texture frequency of the corrugated paperboard, and ensures the coincidence of the measurement data in the spatial position.

[0028] In the measurement principle, the present application introduces a viscoelastic dynamics compensation mechanism, and for the problem of elastic compression of paperboard caused by vibration, the vertical acceleration is converted into a specific compression deformation variable by using Newton's second law and Hooke's law, and the geometric measurement results are physically corrected. This improvement reduces the measurement error in a strong vibration environment and improves the quality control level in the production process of corrugated paperboard.

[0029] The present application has strong adaptability, and through experimental calibration of the two parameters with physical meaning, the boundary sharpness factor and the dynamic compressive modulus, it can be adapted to various production conditions, reducing the difficulty of on-site deployment and maintenance. BRIEF DESCRIPTION OF DRAWINGS

[0030] Figure 1 is a flowchart of a non-contact laser online measurement method for the thickness of corrugated paperboard provided by an embodiment of the present application;

[0031] Figure 2 is a schematic diagram of signal separation based on gradient-energy phase space provided by the embodiment of the application;

[0032] Figure 3 is a comparative diagram of viscoelastic dynamics compensation effect under strong vibration working condition provided by the embodiment of the application. DETAILED DESCRIPTION

[0033] The application provides a non-contact laser online measurement method for the thickness of corrugated paperboard, referring to Figure 1 , comprising steps S1-S4:

[0034] S1: equal-interval space domain mapping.

[0035] Specifically, the vertical distance signal of the corrugated paperboard surface is collected by using the distance collection unit, the real-time pulse signal of the conveying belt is collected by using the displacement recording unit, and the collected signal is interpolated and resampled to obtain an equal-interval space domain sequence.

[0036] 1. In this embodiment, the Kyans LJ-V7000 series high-frequency laser profiler is selected as the distance collection unit, the sampling frequency is set to 64 kHz, the repeat accuracy is set to 0.5 μm, the distance collection unit is installed directly above the conveying belt for data collection, and a distance-time sequence is obtained. Each data item in the sequence is defined as a distance-time point, and each distance-time point represents the vertical distance between the laser emitting head of the distance collection unit and the surface of the corrugated paperboard at a certain time, which is recorded as a distance measurement value. When the surface of the corrugated paperboard is uneven or the corrugated paperboard vibrates with the conveying belt, the distance measurement value will change over time, so the distance-time sequence reflects the high-low fluctuation state of the surface of the corrugated paperboard.

[0037] 2. In this embodiment, the Omron E6B2 series incremental rotary encoder is selected as the displacement recording unit, the resolution is set to 5000 pulses per revolution, and it is installed on the driving roller shaft of the conveying belt for data collection to obtain a real-time pulse sequence. The sequence records the total number of pulses generated since the start of the conveying belt, which represents the cumulative travel of the conveying belt in the feeding direction.

[0038] 3. Since the production line of the corrugated paperboard is usually in a variable frequency speed regulation state, especially the running speed of the conveying belt dynamically fluctuates between 150 meters per minute and 300 meters per minute, directly analyzing the distance-time sequence will cause the geometric characteristics of the corrugated paperboard surface to be distorted on the time axis. Therefore, it is necessary to eliminate the time variable and establish a unified space coordinate system, which specifically includes:

[0039] (1) The processor reads the total number of accumulated pulses collected by the displacement recording unit in real time, and divides the circumference of the driving roller of the conveying belt by the resolution of the displacement recording unit to obtain the physical travel of the conveying belt represented by a single pulse, i.e., the pulse equivalent;

[0040] (2) The total number of accumulated pulses is multiplied by the pulse equivalent to obtain the real-time transmission distance of the conveying belt relative to the starting time, which reflects the actual physical position of the corrugated board on the conveying belt and is not affected by the speed of the conveying belt;

[0041] (3) Since the distance collection unit collects data at a fixed sampling frequency, but the speed of the conveying belt often fluctuates, which leads to uneven physical distribution of the collected distance time series on the corrugated board. When the conveying belt is fast, the physical spacing of the distance time points on the corrugated board is large and sparse, and when the conveying belt is slow, the physical spacing of the distance time points on the corrugated board is small and dense. This unevenness will seriously interfere with the subsequent recognition of the shape of the corrugated board. In order to solve this problem, a fixed spatial sampling interval is set , i.e., a distance measurement value is output every 0.5 mm, and the real-time transmission distance is taken as the abscissa to divide the coordinate axis into equally spaced grid nodes with as the step size;

[0042] (4) The distance measurement value and the real-time transmission distance are one-to-one bound by time to construct a data set containing several discrete data pairs, denoted as data set, where is the real-time transmission distance at time , and is the distance measurement value at time ;

[0043] (5) For the coordinates of a grid node, denoted as , the closest previous and next coordinate pairs in the data set are found, the previous coordinate pair is denoted as , and the next coordinate pair is denoted as , and linear interpolation calculation is performed:

[0044]

[0045] where is the difference distance of the grid node, and the difference distances of all grid nodes are arranged in spatial order to obtain a series of spatially discrete points, defined as sampling points. These sampling points form a spatial domain sequence with constant sampling interval, and each data item in the spatial domain sequence is the distance measurement value of the corresponding sampling point.

[0046] During the production of corrugated cardboard, the speed of the conveyor belt often changes due to processes such as order changes and paper splicing, causing Doppler effect waveform distortion in the collected signals on the time axis. This step constructs a spatial domain sequence through pulse counting and linear interpolation, ensuring that the physical distance between two adjacent sampling points in the sequence remains constant at 0.5mm regardless of the change in conveyor belt speed. This not only restores the true geometric appearance of the corrugated cardboard but also ensures the consistency of the physical meaning of subsequent calculations.

[0047] S2: Orthogonal separation based on gradient-energy phase space.

[0048] Specifically, the instantaneous gradient features and local energy features of each sampling point are extracted from the spatial domain sequence. A gradient-energy phase space is constructed based on the instantaneous gradient features and local energy features. The separation coefficient is calculated using a nonlinear classification function to achieve signal decoupling.

[0049] Traditional frequency domain filtering cannot solve the problem of overlapping vibration frequencies and texture frequencies in the production process of corrugated cardboard. Based on signal feature analysis, it can be seen that although the two have similar frequencies, they have essential differences in local morphology: corrugated texture is characterized by abrupt changes in geometric position, i.e., it has high gradient characteristics; while mechanical vibration is characterized by smooth undulations throughout, i.e., it has high energy characteristics. Based on this feature analysis, this invention constructs a two-dimensional feature space with instantaneous gradient characteristics as the vertical axis and local energy characteristics as the horizontal axis, hereinafter referred to as phase space. The competitive relationship between these two features is used to determine the geometric thickness of the current sampling point.

[0050] 1. Feature extraction and phase space construction, using... This represents the index of the sampling point in the spatial domain sequence. For each sampling point in the spatial domain sequence, its instantaneous gradient features and local energy features are calculated, specifically including:

[0051] (1) For the first There are 1 sampling point, and the distance measurement value is recorded as . , Distance measurement from the previous sampling point The absolute value of the difference is the instantaneous gradient feature of that sampling point, denoted as . It characterizes the steepness of the corrugated cardboard surface at the microscopic scale, such as the high gradient in the transition zone between the flute peaks and valleys of the corrugated cardboard texture.

[0052] (2) In order to smooth out the high-frequency fluctuations caused by the texture undulations of the corrugated cardboard in subsequent calculations and focus on capturing the low-frequency mechanical vibration trend, a length of [missing information] needs to be set. Sliding window, It must cover at least one complete corrugation cycle of the corrugated cardboard being tested, i.e. The distance is greater than the physical distance between two adjacent flute peaks. This embodiment is for C-flute corrugated cardboard with a common standard flute pitch of 7.8mm. ;

[0053] For each sampling point in the spatial domain sequence, a sliding window centered on that sampling point is selected. The root mean square value of the distance measurements of all sampling points within the sliding window is calculated and used as the local energy feature of the current sampling point, denoted as . This characterizes the intensity of mechanical vibration; it should be noted that... The value is not fixed; in practical applications, it can be adaptively adjusted according to the flute profile of the corrugated cardboard. The values ​​of any number of corrugated elements, as long as they meet the condition of covering at least one corrugated cycle, are within the scope of protection of this invention.

[0054] 2. Calculate the separation coefficient for each sampling point based on instantaneous gradient characteristics and local energy characteristics. The relationship is as follows:

[0055]

[0056] in, For the first Separation coefficient of each sampling point For boundary sharpness factor, For the first Instantaneous gradient features of each sampling point For the first Local energy characteristics of each sampling point The preset normalized gradient threshold, The preset normalized energy threshold, It is a natural exponential function.

[0057] Characterizes the probability that the current sampling point belongs to the corrugated texture, when A value close to 1 indicates a higher probability that the sampling point has a corrugated texture. A value close to 0 indicates that the probability of the sampling point being a corrugated texture is relatively low, and it may be due to mechanical vibration.

[0058] It should be noted that, and It was obtained through a pre-designed calibration experiment, specifically:

[0059] (1) Using a distance acquisition unit, scan the corrugated cardboard with a flat surface and intact corrugation in a stationary state, and calculate the arithmetic mean of the instantaneous gradient characteristics of all sampling points on its surface, defined as ;

[0060] (2) Control the conveying belt to run at normal production speed in the empty state without placing corrugated board, collect the fluctuation data of the conveying belt surface by the distance acquisition unit, set the sampling time to 5 min, calculate the arithmetic mean of the local energy features of all the collected sampling points, and define it as .

[0061] Boundary sharpness factor for controlling the sensitivity to mechanical vibration, specifically, control the rate of change from 0 to 1, the present application calibrates the through a phase space boundary continuous calibration experiment, the specific steps are as follows:

[0062] (1) Select a length of 500 mm C-corrugated paperboard as the sample for the experiment, the front 250 mm surface is flat and the corrugated shape is intact, the rear 250 mm is artificially manufactured 2 mm warping and 15 Hz mechanical vibration is applied;

[0063] (2) Scan the sample by using the distance acquisition unit to obtain the spatial domain sequence, and then obtain the instantaneous gradient feature, local energy feature and other data of each sampling point, recorded as test data, and the test data of a sampling point is recorded as a group; mark the data of the front 250 mm of the sample in the spatial domain sequence, recorded as the stable area, and the data of the rear 250 mm, recorded as the mutation area, define the range of 50 mm before and after the junction of the two areas as the transition window;

[0064] (3) The preset traversal interval is [2, 10], the step is 0.5, and the corresponding separation coefficient is calculated for each , each group of test data;

[0065] (4) In the stable area, the ideal separation coefficient should tend to 1, so the separation coefficient of the sampling points in the transition window is taken, the mean value of the separation coefficient of the stable area is calculated, and the values less than 0.95 are removed; for the remaining , take the remaining separation coefficients of the stable area and the mutation area, respectively, calculate the absolute value of the difference between the current sampling point and the previous sampling point, and record it as , take the value corresponding to the lowest as the value of the boundary sharpness factor, and the present embodiment takes 5.

[0066] Figure 2 ​​This is a schematic diagram of signal separation based on gradient-energy phase space provided by an embodiment of the present invention. It can be seen that the sampling points exhibit obvious orthogonal clustering characteristics. The blue point group in the upper left region shows high gradient and low energy, corresponding to the effective texture features of the corrugated cardboard surface, namely the geometric abrupt changes between flute peaks and valleys. These sampling points are retained. The red point group in the lower right region shows low gradient and high energy, corresponding to the mechanical vibration features of the conveyor belt, namely the smooth motion with a large amplitude. These sampling points are identified as noise. The black dashed line in the middle of the figure is the nonlinear separation boundary. The present invention distinguishes the two types of features through this boundary. The change in color depth shows the smooth transition of the separation coefficient from 0 to 1.

[0067] 3. Utilize A recursive filter is constructed to iteratively update the vibration baseline of the sampling points, with the following relationship:

[0068]

[0069] in, For the first Distance measurements at each sampling point; For the first Vibration baseline at each sampling point It is the first The vibration baseline at each sampling point shows that when... When the value approaches 1, the vibration baseline approaches the value of the vibration baseline of the previous sampling point, ignoring the texture abrupt change at the current sampling point and avoiding misjudging high ridges as vibration; when When the value approaches 0, the vibration baseline approaches the distance measurement value of the current sampling point, thus restoring the undulation state of the conveyor belt.

[0070] For each sampling point, the absolute value of the difference between its distance measurement and the vibration baseline is taken to obtain the geometric thickness of that sampling point.

[0071] It should be noted that during the system initialization phase, zero-point calibration needs to be performed. Specifically, the conveyor belt is controlled to run at normal production speed under no-load conditions. Distance measurement values ​​are continuously collected from a section of the conveyor belt surface using a distance acquisition unit. The collection duration is set to cover at least one cycle of conveyor belt operation. The arithmetic mean of these distance measurements is calculated and recorded as the absolute zero-point distance. For the initial vibration baseline, i.e., the vibration baseline before iterative updates, this value is assigned to the absolute zero-point distance. Since the vibration baseline essentially reflects the real-time spatial position of the corrugated cardboard bottom surface, i.e., the conveyor belt surface, under vibration, therefore… and The absolute value of the difference between the two is the geometric thickness of the sampling point.

[0072] S3: Viscoelastic dynamics compensation.

[0073] Since corrugated cardboard is a viscoelastic material, it will undergo micron-level elastic compression under the inertial force generated by the vibration of the conveyor belt. This step aims to correct this physical error. Specifically, a viscoelastic dynamic model is constructed, the vertical acceleration is calculated using the vibration baseline, and the deformation of the corrugated cardboard under the action of inertial force is calculated by combining the areal density of the corrugated cardboard with the preset dynamic compressive modulus, thereby obtaining the true physical thickness of each sampling point.

[0074] 1. Principles of Physical Modeling

[0075] Existing technologies typically treat corrugated cardboard as a rigid body, assuming that the scanning results from the distance acquisition unit are merely changes in the spatial position of the corrugated cardboard. However, since corrugated cardboard is a multi-layered structure composed of face paper, core paper, and liner paper, it exhibits significant viscoelastic characteristics, and its physical properties are similar to a parallel spring-damped system. When the conveyor belt experiences high-frequency mechanical vibration in the vertical direction, the corrugated cardboard will generate a vertical acceleration. According to Newton's second law, this acceleration will generate a vertical inertial force. Under the compression of this inertial force, the corrugated cardboard's peaks will undergo micron-level elastic compressive deformation. At this point, the geometric thickness of each sampling point is actually the thickness of the cardboard after deformation, which is less than the true physical thickness of the cardboard in its free and static state. Therefore, this step compensates for this deformation by constructing a viscoelastic dynamic model to obtain the true physical thickness of each sampling point.

[0076] 2. Vertical acceleration

[0077] Using the central difference method, the vertical acceleration at each sampling point is calculated based on the vibration baseline. For the th... Vertical acceleration at each sampling point : ,in The conveyor belt is in the... The velocity at each sampling point is obtained by differentiating the real-time pulse sequence with respect to time. The preset spatial sampling interval, It is the first Vibration baseline of each sampling point.

[0078] 3. Based on Hooke's Law and the formula for inertial force, a compensation model is constructed to compensate for the geometric thickness of each sampling point, thereby obtaining the true physical thickness of each sampling point. The relationship is as follows:

[0079]

[0080] in, For the first The actual physical thickness of each sampling point For the first Distance measurements of each sampling point For the first Vibration baseline at each sampling point The item is the first Geometric thickness of each sampling point The areal density of the corrugated cardboard is obtained from production work order data. For the first Vertical acceleration at each sampling point This term characterizes the vertical inertial force experienced by corrugated cardboard per unit area. It is the dynamic compressive modulus.

[0081] It should be noted that dynamic compressive modulus The stiffness of corrugated cardboard in the vertical direction against high-frequency impacts is characterized by a comprehensive calibration through a true-value inverse fitting experiment, the steps of which are as follows:

[0082] 1. Select a 500mm long, flat, and intact C-flute corrugated cardboard as the experimental sample. Under static conditions, use a high-precision micrometer to measure its average thickness at multiple points, and record it as the true static thickness.

[0083] 2. Place the sample on a conveyor belt and control the conveyor belt to run at a vibration frequency of 20Hz. Calculate the geometric thickness and vertical acceleration of each sampling point.

[0084] 3. Preset The traversal interval is [50kPa, 500kPa], with a step size of 50kPa. For each Calculate the arithmetic mean of the physical true thickness at all sampling points, and take the arithmetic mean that is closest to the true static thickness. As the value of dynamic compressive modulus, in this embodiment Take 300 kPa.

[0085] Figure 3 This is a comparison chart of the viscoelastic dynamic compensation effect under strong vibration conditions provided by the embodiments of the present invention. It can be seen that in the strong vibration range shown in the figure, that is, when the vertical acceleration is high, the dotted line shows a concave trend, and the measured value is lower than the true value. This is because the cardboard is compressed by inertial force. The solid line represents the true physical thickness after compensation. The overall waveform remains stable. According to actual measurement, the measurement error after compensation is converged to within ±0.03 mm.

[0086] The prior art often ignores the phenomenon that corrugated paperboard as a viscoelastic material will be deformed due to inertial force in high-speed motion, resulting in that the geometric thickness measured in the strong vibration section such as the acceleration section of the conveying belt is lower than the actual thickness of the corrugated paperboard, thereby causing quality misjudgment; the present step compensates for the micron-level thickness compressed due to vibration by introducing the surface density and the vertical acceleration, and ensures that the output measurement result is still accurate even on a bumpy production line.

[0087] S4: thickness statistical output.

[0088] Specifically, the real physical thickness of each compensated sampling point is searched, the sampling points representing the corrugated peaks are extracted, the abnormal values are removed through a statistical process control algorithm, and the final thickness measurement result is generated.

[0089] 1. Data extraction

[0090] Since the surface of the corrugated paperboard presents a periodic wavy structure, the nominal thickness corresponds to the corrugated peak of the corrugated wave; therefore, the processor traverses the real physical thickness of each sampling point, and for any sampling point, if the value of its real physical thickness is greater than the values of the real physical thicknesses of the previous sampling point and the next sampling point, it is determined that the sampling point is a local maximum point, i.e. a corrugated peak; the real physical thicknesses of all sampling points belonging to the corrugated peak are extracted to construct an original corrugated peak thickness set.

[0091] 2. Abnormal value removal based on statistical distribution

[0092] In order to prevent false data caused by local damage, flash, ink dirt and other interference on the surface of the corrugated paperboard, the original corrugated peak thickness set cannot be directly processed, and a statistical process control algorithm needs to be introduced to clean it up.

[0093] In order to balance the measurement real-time and data stability, the statistical period is set to 3 seconds, the arithmetic mean of all real physical thicknesses in each period is calculated and the standard deviation , based on the 3 principle in statistics, the effective data screening interval is constructed; the original corrugated peak thickness set is traversed, and the data falling within the interval is removed to obtain an effective corrugated peak thickness set.

[0094] 3. Final output

[0095] The arithmetic mean of all real physical thicknesses in the set of effective corrugated peak thicknesses is the final measured thickness in the statistical period; in this embodiment, a Siemens S7-1500 series programmable logic controller is used as the core controller, the final measured thickness is sent to the PLC in real time through the Profinet industrial bus, the final measured thickness received by the PLC is compared with the set standard thickness, and the pressure of the ballast roller at the paste box of the double-sided machine is adjusted by using a PID algorithm; if the final measured thickness is too small, the PLC controls the air cylinder to slightly lift the ballast roller; if the final measured thickness is too large, the PLC controls the air cylinder to increase the pressure.

[0096] As can be seen, this processing method not only ensures that the signal output to the PLC is smooth and real, but also prevents errors caused by noise, effectively prolongs the service life of mechanical parts, and ensures the quality of corrugated paperboard.

Claims

1. A non-contact laser on-line measurement method of corrugated board thickness, characterized by, The application relates to a method for measuring the thickness of corrugated paperboard. The method comprises the following steps: Collecting a distance-time sequence of the surface of the corrugated paperboard and a real-time pulse sequence of the conveying belt by using a preset distance acquisition unit and a displacement recording unit, and performing interpolation processing on the distance-time sequence by using the real-time pulse sequence to obtain a spatial domain sequence, wherein the spatial domain sequence is composed of a series of spatially discrete sampling points, and each sampling point has a distance measurement value; Calculating the instantaneous gradient feature and the local energy feature of the sampling point based on the distance measurement value, and then calculating the separation coefficient of the sampling point by using a preset nonlinear classification function; The method for calculating the instantaneous gradient feature and the local energy feature of the sampling point comprises the following steps: For a certain sampling point, the absolute value of the difference between the distance measurement value of the sampling point and that of the previous sampling point is calculated to obtain the instantaneous gradient feature of the sampling point; If denotes the sequence number of the sampling point in the spatial domain sequence, then the separation coefficient of the th sampling point satisfies the relationship: For the first Separation coefficient of each sampling point For boundary sharpness factor, For the first Instantaneous gradient features of each sampling point For the first Local energy characteristics of each sampling point The preset normalized gradient threshold, The preset normalized energy threshold, It is a natural exponential function; A sliding window is preset, and the length of the sliding window covers the physical distance between a corrugation peak and a corrugation valley of the corrugated paperboard; for a certain sampling point, a sliding window with the sampling point as the center is selected, and the root mean square value of the distance measurement values in the sliding window is calculated to obtain the local energy feature of the sampling point; A recursive filter is constructed based on the separation coefficient to iteratively update the vibration reference line of each sampling point, and then the geometric thickness of each sampling point is obtained; 2. A non-contact laser online measurement method of the thickness of corrugated paperboard according to claim 1, characterized in that, Based on a preset viscoelastic dynamic model, the vertical acceleration of each sampling point is calculated by using the vibration reference line, and the compression deformation of each sampling point under the action of inertial force is obtained by combining a preset dynamic compressive modulus; the geometric thickness is compensated by using the compression deformation to obtain the true physical thickness of each sampling point. The interpolation processing specifically comprises the following steps: The total number of pulses output by the displacement recording unit is counted, and the total number of pulses is converted into the real-time transmission distance of the conveying belt relative to the starting position according to a preset pulse equivalent; A fixed spatial sampling interval is set; 3. A non-contact laser on-line measurement method of the thickness of corrugated board according to claim 1, characterized in that, In the coordinate system defined by the real-time transmission distance, linear interpolation operation is performed on the distance-time sequence to obtain the spatial domain sequence.

4. The method of claim 1, wherein the method is a non-contact laser online measurement method of a thickness of corrugated paperboard. If denotes the sequence number of the sampling point in the spatial domain sequence, then the real physical thickness of the first sampling point satisfies the relationship: in, For the first The actual physical thickness of each of the aforementioned sampling points. For the first The distance measurement value of each of the sampling points For the first The vibration baseline of the sampling points. The areal density of the corrugated cardboard. For the first The vertical acceleration of each of the sampling points The dynamic compressive modulus is given.

5. A non-contact laser on-line measurement method of the thickness of corrugated board according to claim 4, characterized in that, The iterative updating of the vibration reference line of each sampling point specifically comprises the following steps: for a certain sampling point, the separation coefficient of the sampling point is set as a weight, the vibration reference line of the previous sampling point of the sampling point and the distance measurement value of the sampling point are weighted and summed to obtain the vibration reference line of the sampling point.

6. A non-contact laser on-line measurement method of the thickness of corrugated board according to claim 1, characterized in that, The vertical acceleration is calculated by the second-order difference of the vibration reference line and the speed of the conveying belt.

7. A non-contact laser on-line measurement method of the thickness of corrugated board according to claim 4, characterized in that, The boundary sharpness factor is obtained through a phase space boundary continuous calibration experiment.

8. A non-contact laser online measurement method of the thickness of corrugated paperboard according to claim 1, characterized in that, The dynamic compressive modulus is comprehensively calibrated through a true value reverse fitting experiment. The distance acquisition unit is a high-frequency line laser profiler, and the displacement recording unit is an incremental encoder installed on the driving shaft of the conveying belt.

Images