A printing and binding method and a printing and binding machine

Abstract

This invention discloses a printing and binding method and machine using a combination of glue and thread, belonging to the field of printing and binding technology. The method includes a thread-binding step, setting initial tension; a glue coating step, collecting spine displacement data and calculating curvature distribution; a control parameter acquisition step; a dynamic curing control step, establishing a coupled prediction model, predicting the curvature sequence in each control cycle, solving for the optimal tension adjustment amount through a cost function and constraints, and executing the process; and a curing determination and folding step, where curing is completed and folding occurs when the curvature deviation is less than a threshold and the glue temperature reaches the target. This invention also discloses a binding machine for implementing this method, including a printer, a punching device, a composite binding head integrating a hot melt glue spray head, thread guide pins, and an electromagnetic tension controller, a laser displacement sensor array, a glue temperature sensor, a controller, and a folding device. This invention can actively suppress spine curing caused by glue curing, improving binding quality and consistency.

Description

Technical Field

[0001] This invention relates to the field of printing and binding, specifically to a printing and binding method and a printing and binding machine using glue and thread bonding. Background Technology

[0002] Printing and binding refers to the process of arranging, connecting, and binding printed paper into a book. Common binding methods include staple binding, perfect binding, and thread binding. Perfect binding uses hot melt adhesive to bond the pages, which is efficient and low-cost, but the spine of a perfect book is prone to cracking and cannot lay flat at 180°, especially for coated paper, thick books, or instruction manuals and picture albums that are frequently folded, where the spine of a perfect book lacks durability. Thread binding connects the pages with stitches, which is sturdy and can lay flat, but the thread binding process is complex, inefficient, and costly, and dust can easily get into the gaps. In recent years, a hybrid binding method combining perfect binding and thread binding has emerged, in which hot melt adhesive is applied after threading, attempting to combine the efficiency of perfect binding with the sturdiness of thread binding. However, in practical applications, it has been found that hot melt adhesives shrink in volume during the curing process. If tension is applied to the thread before the adhesive cures to keep the pages firm, the shrinkage of the adhesive layer will pull the thread, and the thread will pull the spine through the hole wall, causing the spine to bend inward. This bending is particularly noticeable when the book is thick, the thread tension is high, or the paper stiffness is low, which seriously affects the stacking flatness and appearance quality of the book.

[0003] Therefore, in order to solve the problems existing in the prior art, the present invention proposes a printing and binding method and a printing and binding machine using glue and thread composite. Summary of the Invention

[0004] In view of the shortcomings of the existing technology, the purpose of this invention is to provide a printing and binding method and a printing and binding machine with adhesive thread bonding.

[0005] To achieve the above objectives, the present invention provides the following technical solution: A printing and binding method using adhesive thread lamination includes: The thread-binding process involves drilling holes along the spine edge of the book block and threading the book through them, then setting the initial tension of the binding thread using an electromagnetic tension regulator. In the adhesive coating step, hot melt adhesive is applied to the spine surface after thread binding, and displacement data of several monitoring points along the spine length direction are collected according to a preset sampling frequency. The curvature distribution of the spine surface is calculated based on the displacement data. The control parameter acquisition step involves obtaining the curing shrinkage rate at the current moment based on the adhesive layer temperature and the preset adhesive layer shrinkage rate calibration curve, obtaining the real-time tension of the thread through the feedback result of the electromagnetic tension regulator, obtaining the distance of the thread from the spine neutral axis and obtaining the preset book core equivalent bending stiffness, and using the acquired data set as tension control parameters. The dynamic solidification control steps involve constructing a coupled prediction model between a preset curvature target and the tension control parameters. In each control cycle, the curvature sequence for the subsequent control cycle is calculated using the coupled prediction model based on the current curvature distribution and the real-time tension of the wire. An adjustment cost function is constructed using tension amplitude constraints and rate of change constraints. The optimal tension adjustment amount sequence for the subsequent adjustment cycle is calculated in conjunction with the initial tension. The wire tension is then adjusted based on the optimal tension adjustment amount sequence. In the curing determination and folding step, the average deviation between the current curvature distribution and the target curvature is calculated. When the average deviation is less than a preset threshold for multiple consecutive sampling cycles and the adhesive layer temperature drops below the set value, the curing is determined to be complete and the book is sent to the folding station for folding. Otherwise, the process returns to the next control cycle of the dynamic curing control step and dynamic curing control is re-performed through the tension deviation adjustment strategy.

[0006] As a further improvement of the present invention, the adhesive coating step includes: collecting displacement data through a laser displacement sensor, wherein the laser displacement sensor is arranged in an equidistant array along the spine length direction, and obtaining the displacement value of the point according to the position coordinates of the laser displacement sensor, calculating the curvature of the point according to the displacement value, and setting the average absolute value of the curvature of all laser displacement sensors as the average curvature deviation of the current spine.

[0007] As a further improvement of the present invention, the calibration of the preset adhesive shrinkage rate calibration curve includes: applying the hot melt adhesive to be used to the test substrate, collecting the temperature change and corresponding volume shrinkage rate during the curing process of the adhesive layer in a constant temperature environment, fitting the collected temperature data and shrinkage rate data, establishing a functional relationship between shrinkage rate and temperature and time, generating the fitted shrinkage rate calibration curve and pre-storing it.

[0008] As a further improvement of the present invention, the dynamic curing control step includes: the number of subsequent control cycles is determined by a prediction time domain and a control time domain; the prediction time domain is set according to the total curing time of the adhesive layer and the sampling period; the control time domain is less than or equal to the prediction time domain and is preset by offline simulation; the rate of change of curvature deviation is monitored during the curing process; when the rate of change exceeds a preset threshold, the control time domain is increased; when the curvature deviation tends to stabilize, the control time domain is decreased.

[0009] As a further improvement of the present invention, the dynamic curing control step further includes establishing a functional mapping relationship between the current curvature, real-time thread tension, curing shrinkage rate, equivalent bending stiffness of the book core, and the distance of the thread from the neutral axis of the spine. The functional mapping relationship is a discrete state equation, which includes a tension influence coefficient and a shrinkage influence coefficient. The tension influence coefficient is predetermined based on the equivalent bending stiffness of the book core and the distance of the thread from the neutral axis of the spine. The shrinkage influence coefficient is determined by offline calibration based on the type of adhesive and the thickness of the adhesive layer. The coupled prediction model is constructed by taking the current curvature and real-time thread tension as the initial state and iteratively predicting the curvature value of each future control cycle according to the discrete state equation. In each iteration, the thread tension is updated according to the tension adjustment amount of the current cycle and substituted into the calculation of the next cycle. The tension adjustment amount is determined by the candidate value generated by the adjustment cost function in each iteration.

[0010] As a further improvement of the present invention, the weight matrix in the adjustment cost function includes a curvature deviation weight matrix and a tension adjustment amount weight matrix, which are pre-tuned through offline simulation combined with experiments.

[0011] As a further improvement of the present invention, the tension deviation adjustment strategy includes: when the average deviation is greater than a preset threshold, decomposing the current curvature deviation into a contraction contribution component, a tension contribution component, and a diffusion contribution component according to the discrete state equation of the coupled prediction model; comparing the magnitude of the absolute values ​​of each component to determine the main source of the deviation; if the contraction contribution component is the largest, increasing the ratio of the curvature deviation weight matrix to the tension adjustment amount weight matrix in the adjustment cost function, and simultaneously pre-compensating the tension adjustment amount for the next two control cycles according to the current rate of change of contraction rate; if the tension contribution component is the largest, keeping the weight ratio unchanged, and directly increasing the upper limit of the tension adjustment amount for the next control cycle; if the diffusion contribution component is the largest, increasing the prediction time domain and decreasing the control time domain, and superimposing the above adjustment actions and outputting them to the electromagnetic tension regulator.

[0012] As a further improvement of the present invention, the control parameter acquisition step further includes applying a small tension pulse of preset amplitude through an electromagnetic tension regulator after the thread binding is completed and before the glue application step begins, recording the transient curvature response of the spine measured by the laser displacement sensor array, calculating the equivalent bending stiffness of the current book block based on the peak ratio of the tension pulse amplitude to the curvature response, combined with the distance of the thread from the neutral axis of the spine, and updating the pre-stored stiffness value in the controller.

[0013] This invention also proposes a printing and binding machine with adhesive thread lamination, comprising: printer; A punching device, located on the side wall of the printer, is used to drill through holes along the spine edge of the book block. A composite binding head is mounted on a mounting rod on the side wall of the printer. The composite binding head integrates a hot melt adhesive spray head, a wire guide pin, and an electromagnetic tension controller. A laser displacement sensor array is set above the binding station and is equidistant along the spine length to collect displacement data on the spine surface. The adhesive temperature sensor is located at the outlet of the hot melt adhesive spray nozzle to collect the temperature of the adhesive layer. The controller is electrically connected to the electromagnetic tension controller, the laser displacement sensor array, the adhesive temperature sensor, and the hot melt adhesive spray head, respectively. A folding device, located on the lower surface of a substrate, is used to fold bound books in half.

[0014] As a further improvement of the present invention, the laser displacement sensor array is used to continuously collect displacement data of each measuring point in the spine length direction at a sampling frequency of 100Hz to 200Hz during the hot melt adhesive curing process; the controller is configured to calculate the curvature of each measuring point by means of the three-point difference method based on the displacement data of each measuring point, and take the average absolute value of the curvature of all measuring points as the average curvature deviation of the current spine. Based on the discrete state equation, using the average curvature deviation and the real-time tension of the wire fed back by the electromagnetic tension controller as the initial state, the curvature sequence within the future prediction time domain steps is iteratively predicted. Based on the adjustment cost function, a quadratic programming problem is solved under the constraints of tension amplitude and rate of change to obtain the optimal tension adjustment sequence. Only the first adjustment is output to the electromagnetic tension controller to change the wire tension. Repeat the above steps of data collection, calculation, prediction, solution, and adjustment until the conditions for the solidification determination step are met.

[0015] The beneficial effects of this invention are as follows: This invention uses a laser displacement sensor array to monitor the spine curvature in real time, combined with an adhesive temperature sensor to obtain the adhesive layer temperature, to establish a coupled predictive model that includes curing shrinkage rate, thread tension, book core stiffness, and curvature spatial diffusion. The model predictive control dynamically adjusts the thread tension, actively suppressing spine bending caused by volume shrinkage during hot melt adhesive curing, ensuring the finished book remains straight and significantly improving binding quality. Furthermore, this invention organically integrates a punching device, a composite binding head, a laser displacement sensor array, an adhesive temperature sensor, a controller, and a folding device. The controller embeds a model predictive control algorithm, with real-time sensor feedback and precise actuator response, forming a closed-loop control system that achieves fully automated binding from punching to folding. Because it can actively suppress bending, there is no need to reduce thread tension based on experience, maintaining a high initial tension and ensuring binding strength. Simultaneously, dynamic adjustment during curing shortens the curing waiting time, improves production efficiency, and reduces the scrap rate caused by bending. Attached Figure Description

[0016] Figure 1 This is a flowchart of a printing and binding method using adhesive thread lamination according to the present invention; Figure 2 This is a schematic diagram of a printing and binding machine with adhesive thread lamination according to the present invention.

[0017] Attached image captions: 1. Printer; 2. Punching device; 3. Composite binding head; 4. Mounting rod; 5. Laser displacement sensor array; Folding device. Detailed Implementation

[0018] The present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Identical components are denoted by the same reference numerals. It should be noted that the terms "front," "rear," "left," "right," "upper," and "lower" used in the following description refer to directions in the accompanying drawings, and the terms "bottom surface," "top surface," "inner," and "outer" refer to directions toward or away from the geometric center of a specific component, respectively.

[0019] This invention provides a printing and binding method using adhesive thread lamination, comprising: The thread-binding process involves drilling holes along the spine edge of the book block and threading the book through them, then setting the initial tension of the binding thread using an electromagnetic tension regulator. In the adhesive coating step, hot melt adhesive is applied to the spine surface after thread binding, and displacement data of several monitoring points along the spine length direction are collected according to a preset sampling frequency. The curvature distribution of the spine surface is calculated based on the displacement data. The control parameter acquisition step involves obtaining the curing shrinkage rate at the current moment based on the adhesive layer temperature and the preset adhesive layer shrinkage rate calibration curve, obtaining the real-time tension of the thread through the feedback result of the electromagnetic tension regulator, obtaining the distance of the thread from the spine neutral axis and obtaining the preset book core equivalent bending stiffness, and using the acquired data set as tension control parameters. The dynamic solidification control steps involve constructing a coupled prediction model between a preset curvature target and the tension control parameters. In each control cycle, the curvature sequence for the subsequent control cycle is calculated using the coupled prediction model based on the current curvature distribution and the real-time tension of the wire. An adjustment cost function is constructed using tension amplitude constraints and rate of change constraints. The optimal tension adjustment amount sequence for the subsequent adjustment cycle is calculated in conjunction with the initial tension. The wire tension is then adjusted based on the optimal tension adjustment amount sequence. In the curing determination and folding step, the average deviation between the current curvature distribution and the target curvature is calculated. When the average deviation is less than a preset threshold for multiple consecutive sampling cycles and the adhesive layer temperature drops below the set value, the curing is determined to be complete and the book is sent to the folding station for folding. Otherwise, the process returns to the next control cycle of the dynamic curing control step and dynamic curing control is re-performed through the tension deviation adjustment strategy.

[0020] Specifically, such as Figures 1 to 2 As shown, the adhesive coating step includes arranging 5 to 9 laser displacement sensors at equal intervals along the spine length of the book. Too many sensors result in high cost and large data processing volume; too few sensors lead to inaccurate spine curvature plotting, especially when uneven curvature occurs. The preset sampling frequency is 100Hz to 200Hz. The hot melt adhesive curing process typically completes within 10 to 30 seconds, and while the shrinkage rate is non-linear, it still changes at the millisecond level. Therefore, controlling the sampling frequency above 100Hz captures the details of tension fluctuations caused by adhesive shrinkage, while avoiding excessive bending due to excessively long control cycles. A frequency below 200Hz ensures real-time solving of the MPC problem by the embedded controller. The laser displacement sensors are arranged at equal intervals along the spine length, and the displacement value at each point is obtained based on the sensor's position coordinates. The curvature at that point is calculated from the displacement value. Since the sensors are arranged along the spine length, there are no other coordinate axes; therefore, the position coordinate of the i-th sensor is set to x. i For equally spaced measuring points, by The three-point difference method can be used to calculate the curvature at that location. ,in, h i (t) represents the displacement value of the i-th sensor. Similarly, i-1 and i+1 represent the (i-1)-th and (i+1)-th sensors, respectively. Δx is the distance between adjacent sensors. Under the assumption of small deformation, the spine curvature is approximately d²h / dx², where d is assumed to be 1 and can be ignored. The boundary points of the curvature, i.e., when i=1 or i=N, are obtained through one-sided difference. Using the displacement values ​​of the first three points, a forward difference approximation of the second derivative is constructed. The fourth-order precision forward difference formula obtained through Taylor expansion can be obtained as follows: When the number of sensors is smaller, the average absolute value of the curvature of all laser displacement sensors can be set as the average curvature deviation of the current spine through low-order forward differential calculation, thus obtaining the average curvature deviation of the current spine. .

[0021] Specifically, such as Figures 1 to 2 As shown, the calibration of the preset adhesive shrinkage rate calibration curve is performed during equipment debugging or when the adhesive type is changed. The calibration environment is set to constant temperature and humidity conditions, with the temperature controlled at 20℃±2℃ and the relative humidity controlled at 50%±5%, to avoid interference from environmental fluctuations on the measurement of adhesive curing shrinkage.

[0022] Apply the hot melt adhesive to a smooth test substrate with the same thickness as the actual binding (typically 0.5mm to 1.0mm). The substrate material should be aluminum or stainless steel to ensure uniform thermal conductivity. Embed a 0.2mm diameter K-type thermocouple at the center of the applied adhesive layer to continuously collect the adhesive layer temperature θ(t) at a sampling frequency of 10Hz until the adhesive layer temperature drops to room temperature.

[0023] Volumetric shrinkage was measured using a non-contact laser displacement meter array, with three measuring points arranged along the length of the adhesive layer to measure the height change of the adhesive layer surface during curing. Since the adhesive layer shrinks uniformly in the width direction, the volumetric shrinkage rate... Approximately equal to the shrinkage rate in the thickness direction, calculated using the following formula: ,in (0) represents the initial thickness after coating. θ(t) represents the thickness at time t. The average value calculated independently for each measuring point is taken as the shrinkage rate at that time. The collected temperature data θ(t) and shrinkage rate are then compared. After performing synchronization alignment and removing outliers, the least squares method is used to fit the result. The functional relationship between shrinkage and θ and t is as follows: For EVA hot melt adhesive, the shrinkage rate and temperature fitting curve exhibit nonlinear characteristics and can be fitted using an exponential function or a cubic polynomial. For PUR hot melt adhesive, shrinkage mainly occurs during the curing reaction stage and can be fitted using a double exponential model. The fitting results are pre-stored in the controller memory in the form of a lookup table, with a temperature resolution of 1℃ and a time resolution of 0.1 seconds. In the actual binding process, the control parameter acquisition step obtains the curing shrinkage rate at the current moment by linear interpolation and looking up the table based on the current adhesive layer temperature θ(t) and curing time t. If a different brand or type of hot melt adhesive is used, the above calibration process must be repeated to update the lookup table or fitted curve in the controller to ensure the accuracy of the input to the coupled prediction model.

[0024] Specifically, such as Figures 1 to 2 As shown, the number of subsequent control cycles is determined by the prediction time domain and the control time domain, with N in the prediction time domain. p N represents the number of prediction steps for the future curvature trajectory during each optimization, controlling the time domain. c This represents the number of steps required to solve for the future tension adjustment in each optimization iteration. Prediction time domain N p Based on the total curing time T of the adhesive layer cure and sampling period T s set up, Among them, the total curing time T cure The offline calibration experiments revealed that the time required is typically 10 to 20 seconds for EVA hot melt adhesive and 20 to 30 seconds for PUR hot melt adhesive. Sampling period T sThe sampling frequency is the reciprocal of the sampling frequency. In this embodiment, the sampling frequency is 100Hz to 200Hz, corresponding to T. s The range is from 5ms to 10ms. (T) cure =15 seconds, T s Taking 10ms as an example, N p The number of steps is 1500. To avoid excessive computation, N is set to... p The maximum number of steps is 500, which corresponds to a prediction time of approximately 5 seconds, covering the early stage when glue shrinkage is most significant.

[0025] Control time domain N c Typically less than or equal to the prediction time domain, in this embodiment, offline simulation is used to pre-calibrate the binding process for different types of adhesives and different book thicknesses, aiming to minimize the final curvature upon completion of curing, and selecting the optimal N. c Values, specifically including the adhesive layer shrinkage curve. The equivalent bending stiffness EI of the book block, the distance d from the wire to the neutral axis of the spine, and the coefficient of friction μ between the wire and the hole wall are substituted into the coupled prediction model to form a complete simulation environment for the binding process. The input of the simulation model is the initial tension T0 and the tension adjustment sequence ΔT(t), and the output is the trajectory of the spine curvature κ(t) over time. Several typical working condition combinations are set for actual equipment usage scenarios. The adhesive types include at least EVA and PUR, the book thickness is selected as three typical values ​​of 5mm, 15mm, and 25mm, and the paper weight is selected as two typical values ​​of 120g / m² and 200g / m², for a total of 12 working conditions. Under each working condition, the total curing time is determined according to the calibration curve, and the sampling period T is... s Fixed at 10ms. Set N. c The candidate range is 5 to 30 steps, with a step size of 1. For each working condition, run N... c =5, 6, 7, ..., 30, a total of 26 simulations. In each simulation, other parameters of the model predictive control, such as N, are... p The weight matrices Q and R remain fixed, only N is changed. c The simulation records the final curvature K when the solidification is complete. final And the maximum curvature K during the entire curing process. max The optimal N is selected using a weighted comprehensive scoring method based on the total tension adjustment amplitude Σ|ΔT|. c The evaluation function is: ; in, , , These are the weighting coefficients. The weighting percentage representing the final curvature is usually set to the highest. This represents the weighting percentage of the maximum curvature. Since the maximum curvature reflects the stability of the process, it is set to the second highest. This represents the weighting percentage of the tension adjustment range, which reflects energy consumption and actuator wear. The minimum values ​​of each parameter are respectively for each candidate N. c The optimal value for this operating condition is used for normalization. Calculate each N. c The average comprehensive score under 12 working conditions ,choose The smallest N c As the globally optimal value, this value is substituted into actual binding equipment for verification testing. If the test results are consistent with the simulation trend, this value is pre-stored as the default parameter in the controller. If the glue type is changed or the page specifications are significantly altered later, the above simulation process can be re-executed and N updated. c value.

[0026] Specifically, such as Figures 1 to 2 As shown, the dynamic curing control step further includes, based on the elastic beam bending theory, considering the spine as a continuous elastic beam of length L and equivalent bending stiffness EI, on which two types of loads act: distributed moment and concentrated force. The distributed moment is M. c (x,t) is caused by the curing shrinkage of the adhesive layer. The shrinkage gradient of the adhesive layer in the thickness direction generates a shrinkage stress distribution, characterized by rapid surface shrinkage and slow internal shrinkage. This can be equated to a torque distributed along the spine of the book. ,in This is the shrinkage gradient coefficient, with a value between 0 and 1. E represents the initial thickness of the adhesive layer. g L represents the elastic modulus of the colloidal compound. s The characteristic length of the shrinkage is considered. The wire tension T(t) is transmitted to the spine through friction between the wire and the hole wall. Its position of action is the hole through which the wire passes. A dynamic friction effect exists during tension transmission, and the friction coefficient μ(t) varies with the degree of adhesive curing, expressed as μ(t) = μ0 + k. μ · Where μ0 is the initial friction coefficient, k μ These are empirical coefficients. Under the aforementioned loads, the bending curvature κ(x,t) of the beam satisfies the Euler-Bernoulli beam equation. Discretizing this equation spatially yields a discretized form of the curvature distribution. Further discretizing the time period, i.e., setting a sampling period Δt, results in the following discrete state equation: ; Among them, K i (t) represents the curvature of the i-th measurement point at time t, which is K(xi,t).

[0027] This represents the diffusion of curvature in space, where A is the diffusion coefficient. As EI and paper density increase, A decreases. This represents the effect of the rate of change of tension on curvature, where B is the tension influence coefficient, reflecting the dynamic response to sudden tension changes. This represents the change in curvature directly caused by rubber shrinkage. C is the shrinkage influence coefficient. The coefficient of C is adaptively modified according to the type of rubber. The more easily the rubber shrinks and deforms, the larger C will be. This represents the effect of the rate of change of the shrinkage rate on the curvature. D is the shrinkage acceleration coefficient. When the rate of change of the shrinkage rate increases, the coefficient D also increases.

[0028] In each control cycle, the controller acquires the curvature value of each measuring point at the current moment, the current wire tension value, the tension value of the previous control cycle, the current adhesive layer curing shrinkage rate, and the curing shrinkage rate of the previous control cycle.

[0029] The controller calculates the diffusion effect of curvature along the spine length direction based on the current curvature distribution, i.e., the influence of the curvature difference between adjacent measuring points on the curvature change at the current measuring point. Simultaneously, it calculates the impact of tension abrupt changes on curvature based on the difference between the current tension and the tension of the previous cycle. It calculates the curvature change directly caused by adhesive shrinkage based on the current curing shrinkage rate, and calculates the impact effect of shrinkage acceleration on curvature based on the difference between the current shrinkage rate and the shrinkage rate of the previous cycle. These four effects are superimposed to obtain the predicted curvature of each measuring point at the next moment. After obtaining the predicted curvature for the next moment, the controller uses this predicted value as the curvature input for the next cycle, assumes a candidate tension adjustment amount, and updates the wire tension accordingly. The new tension equals the current tension plus the adjustment amount. Simultaneously, it retrieves the curing shrinkage rate for the next moment from a pre-stored shrinkage rate calibration curve based on the curing time. This iterative process is repeated until a pre-set prediction time-domain step is predicted, thus obtaining the complete curvature sequence of each measuring point within several future control cycles.

[0030] Throughout the iteration process, the candidate tension adjustment sequence is generated by the optimizer at each iteration. The optimizer aims to make the predicted curvature sequence as close to zero as possible, while ensuring that the tension value does not exceed the preset amplitude and rate of change constraints.

[0031] Specifically, such as Figures 1 to 2 As shown, the weight matrices in the adjustment cost function include a curvature deviation weight matrix Q and a tension adjustment amount weight matrix R. Q and R are pre-tuned through offline simulation combined with experimental verification. The weight matrices Q and R in the adjustment cost function determine the relative importance between reducing curvature deviation and reducing the tension adjustment amplitude in the control objective. A larger Q indicates that the controller tends to eliminate curvature deviation more quickly, but this may lead to drastic tension fluctuations. A larger R indicates that the controller tends to adjust the tension more smoothly, but this may sacrifice the convergence speed of the curvature.

[0032] During the equipment commissioning phase, weight tuning was performed in an offline simulation environment, and typical operating conditions were set, including adhesive type, book thickness, and paper weight. The prediction time domain N was fixed. p and control time domain N c The simulation iterates through candidate combinations of Q and R, which are typically set as diagonal matrices and, for simplicity, are usually taken in scalar form: Q = q·l and R = r·l. Iterates through the range of the ratio of q to r, running simulations for each set of parameters and recording the final curvature, maximum curvature, and total tension adjustment amplitude upon completion of curing. The primary objective is to minimize the final curvature, while limiting the maximum curvature to a reasonable threshold, such as 0.02 mm. -1 The total tension adjustment range should not exceed 50% of the initial tension. Select the minimum q / r value that meets the conditions, i.e., the minimum allowable R, to obtain the fastest curvature convergence. If multiple ratios meet the conditions, select the set with the smallest tension adjustment range. Pre-store the selected q and r values ​​in the controller. If the type of adhesive is changed or the book specifications change significantly, the above simulation tuning needs to be repeated.

[0033] During the binding process, the controller monitors the average value E of the curvature deviation in real time. K(t) And its rate of change, when E K(t) When the value consistently exceeds the preset threshold and the rate of change is positive, it indicates that the curvature is intensifying and the current control action is insufficient. In this case, the ratio of Q to R is automatically increased to enhance the penalty for curvature deviation and accelerate the response. When E... K(t) When the value is less than the preset threshold and the rate of change is negative, the bending has been effectively suppressed. At this point, the ratio of q to r should be reduced to decrease the intensity of tension adjustment and reduce actuator wear. Online adaptive adjustment is performed once per control cycle, synchronously with MPC optimization. The adjusted q and r are then substituted into the cost function for the current cycle for solution.

[0034] Using the current curvature and real-time wire tension as the initial state, the curvature value for each future control cycle is iteratively predicted according to the discrete state equation. In each iteration, the wire tension is updated based on the tension adjustment amount of the current cycle and substituted into the calculation for the next cycle. The tension adjustment amount is determined by the candidate value generated by the adjustment cost function in each iteration.

[0035] Specifically, such as Figures 1 to 2As shown, the tension deviation adjustment strategy includes the following steps: when the average deviation exceeds a preset threshold, it is determined that curing is not complete and the dynamic curing control step needs to be returned to for tension adjustment again. At this time, the controller executes the tension deviation adjustment strategy, decomposing the current curvature deviation into shrinkage contribution components, tension contribution components, and diffusion contribution components according to the discrete state equation of the coupled prediction model. The controller calculates the curvature deviation between the measured curvature of each measuring point at the current moment and the target curvature of 0, and compares it with the curvature deviation of the previous control cycle to obtain the change in curvature deviation at each measuring point. Based on the physical composition of the discrete state equation, the change is decomposed to obtain the shrinkage contribution component, tension contribution component, diffusion contribution component, and residual. The shrinkage contribution component is caused by the curvature shrinkage rate of the adhesive layer itself and its rate of change. Based on the current curvature shrinkage rate of the adhesive layer and the shrinkage rate at the previous moment, the curvature change caused by the shrinkage rate itself and its rate of change is calculated. Specifically, during the curing process, the volume shrinkage of the colloid generates shrinkage stress. The larger the shrinkage rate or the faster the shrinkage rate changes, the more significant the impact on curvature, reflecting the driving effect of the adhesive layer material properties on bending. The tension contribution component is calculated based on the difference in wire tension between the current and previous moments (i.e., the rate of change of tension), combined with the tension influence coefficient. This calculates the impact of sudden tension changes on curvature changes. When wire tension changes abruptly within a short time, the friction between the wire and the hole wall transmits to the spine, causing a rapid change in curvature, reflecting the dynamic response characteristics of the actuator. The diffusion contribution component is caused by the non-uniform spatial distribution of curvature. Based on the current spatial distribution of curvature, the influence of the curvature difference between adjacent measuring points on the curvature change at the current measuring point is calculated. If the curvature at a certain measuring point is greater than that of adjacent measuring points, the bending will propagate to both sides. This propagation effect is proportional to the second-order spatial difference of curvature, i.e., the weighted difference between the curvature at the center point and the curvature at the two side points. The proportionality coefficient is the diffusion coefficient A. The diffusion contribution reflects the inhibitory effect of the overall stiffness of the spine on the propagation of bending. Subtracting the sum of the above three contributions from the measured curvature deviation change yields the residual, which may be affected by measurement noise, etc. The controller takes the absolute value of each of the four contributions at each measuring point and then averages them across all measuring points to obtain the final components.

[0036] By comparing the absolute values ​​of each component, the main source of deviation is determined, and the controller adjusts accordingly based on the different sources of deviation.

[0037] If the shrinkage contribution component is the largest, it indicates that the current increased bending is mainly caused by excessively rapid curing shrinkage of the adhesive layer or drastic fluctuations in the shrinkage rate. In this case, the ratio of the curvature deviation weight matrix to the tension adjustment weight matrix in the adjustment cost function is increased by one step. This enhances the penalty for curvature deviation, making the controller more aggressive in suppressing bending, and performs feedforward compensation based on the current contraction rate change to predict the contraction rate for the next two control cycles. as well as Apply additional tension adjustment amount in advance , Where the feedforward coefficient k pre The value ranges from 0.5 to 2N / %, determined by offline calibration. This feedforward compensation amount is superimposed with the adjustment amount output by MPC optimization and then output to the electromagnetic tension regulator, temporarily increasing the prediction time domain in the cost function, thereby better predicting the future trend of shrinkage rate.

[0038] If the tension contribution component is the largest, the weight ratio remains unchanged, and the upper limit of the tension adjustment amount in the next control cycle is directly increased. The upper limit of the tension adjustment amount in the next control cycle is increased, and the adjustment range can be 30% to 50%. At the same time, the change rate constraint remains unchanged to avoid tension impact from damaging the spine. The tension adjustment amount weight matrix in the cost function is temporarily reduced to 0.7 to 0.8 times the original value, thereby allowing for a larger range of tension adjustment.

[0039] If the diffusion contribution component is the largest, it indicates that the bending is propagating along the spine length. This usually occurs when a local bending has already formed and is spreading to both sides. In this case, it is necessary to increase the prediction time domain to enhance the ability to predict the future bending propagation trend, while reducing the control time domain to make the controller focus more on the near-term response and avoid over-adjusting the long-term control quantity.

[0040] When two sources of bias are significantly greater than the third, adjustments are made using a superimposed adjustment method. For example, if both contraction and diffusion contributions are significant, then the three adjustments in the contraction-dominant strategy and the diffusion-dominant strategy are executed simultaneously. All adjusted parameters are immediately substituted into the MPC optimization of the next control cycle, and after the cycle ends, they are restored to their default values ​​or readjusted based on new bias sources.

[0041] Specifically, in the control parameter acquisition step, the equivalent bending stiffness EI of the book block directly affects the values ​​of the tension influence coefficient α and the diffusion coefficient A in the discrete state equation. However, the elastic modulus, moisture content, fiber orientation, and book thickness of different batches of paper vary, and the offline preset EI value may deviate significantly from the actual value. To improve the accuracy of the coupled prediction model, this invention inserts an online identification sub-step after the thread binding is completed and before the gluing step begins.

[0042] The threaded but un-glued book block is considered as a cantilever beam model, fixed at one end and clamped by a clamping mechanism, with the other end free. The thread passes through the spine holes, and the distance of the tension application point from the spine's neutral axis is a known constant d, pre-determined by the equipment's geometry. When the electromagnetic tension regulator applies a tiny tension pulse ΔT, this pulse is transmitted through the thread to the spine, generating a transient bending moment M = ΔT·d. Under this moment, the spine produces a transient curvature response. According to the beam bending formula in elasticity, curvature is the ratio of bending moment to bending stiffness. Therefore, the bending stiffness can be calculated by measuring the peak value of the curvature response. Simultaneously with sending the tension pulse, a laser displacement sensor array is activated to continuously collect displacement data at each measuring point along the spine's length at the highest sampling frequency, and the curvature at each measuring point is calculated in real time. Due to the extremely short pulse duration, the controller only records the curvature response within 200ms after the pulse begins. The maximum curvature value of each measuring point is extracted from the recorded curvature response, and the average value of all measuring points is taken as the overall peak value of the spine's curvature response. The equivalent bending stiffness of the current book block can be calculated using the beam bending formula.

[0043] The second embodiment of the present invention also proposes a printing and binding machine with adhesive thread lamination, comprising: Printer 1 is used to print the book blocks to be bound and automatically feeds the printed paper to the subsequent workstation. The output end of printer 1 is connected to punching device 2 to ensure that the book blocks can continuously enter the binding process after printing without manual intervention.

[0044] A punching device 2, located on the side wall of printer 1, is used to drill through holes along the spine edge of the book block. This punching device 2 can employ the drill bit, pressure box, suction pump, and anti-clogging structure described in the first embodiment. The punching position, hole diameter, and hole spacing are preset according to the diameter of the binding thread and the book thickness. After punching, the through holes provide a channel for subsequent threading steps, while the dust suction structure promptly removes paper scraps generated during drilling, keeping the spine surface clean and preventing any impact on the subsequent gluing quality.

[0045] The composite binding head 3 is mounted on the mounting rod 4 on the side wall of the printer 1, replacing the traditional staple binding head. The composite binding head 3 integrates three key components: a hot melt adhesive spray head for applying hot melt adhesive to the spine surface after threading. The spray head has a temperature control function, allowing adjustment of the adhesive layer thickness and temperature according to the type of adhesive and curing requirements. A thread guide pin is used to accurately pass the binding thread through the pre-drilled holes in the punching device 2. The guide pin can move along the spine direction, cooperating with the electromagnetic tension controller to complete threading and initial tension setting. The electromagnetic tension controller is used to clamp the end of the binding thread and adjust the thread tension in real time according to controller commands. The tension controller can output tension from 0 to a preset maximum value, has a short response time, and can dynamically follow the adjustment amount given by the model predictive control algorithm during the curing process.

[0046] A laser displacement sensor array 5, positioned above the binding station and equidistantly arranged along the spine length, is used to collect displacement data of the spine surface. During the hot melt adhesive curing process, the sensor array continuously measures the distance from each measuring point to the sensor at a preset sampling frequency, thereby obtaining the vertical displacement distribution of the spine surface along its length. This displacement data forms the basis for calculating the curvature distribution, providing raw data for curvature calculation. Simultaneously, in the online identification step, this sensor array is also used to record the transient curvature response caused by tension pulses.

[0047] An adhesive temperature sensor, located at the outlet of the hot melt adhesive spray nozzle, is used to collect the temperature of the adhesive layer. During and after adhesive application, the sensor continuously transmits temperature signals to the controller. The temperature data is used to look up the shrinkage rate calibration curve and also to determine whether the adhesive layer temperature has dropped below a set value during the curing determination folding step. The adhesive temperature sensor is a key component for achieving online estimation of curing shrinkage rate.

[0048] The controller is electrically connected to the electromagnetic tension controller, the laser displacement sensor array 5, the adhesive temperature sensor, and the hot melt adhesive spray head. The controller includes an embedded processor, a storage unit, and input / output interfaces, and pre-stores adhesive shrinkage rate calibration curves, prediction time-domain and control time-domain parameters, discrete state equation coefficients, and initial values ​​of the weight matrix. The controller collects sensor data in real time according to the method steps, executes the model predictive control algorithm, and outputs tension adjustment commands to the electromagnetic tension controller.

[0049] The folding device is located on the underside of the substrate and is used to fold the bound books in half. Once the controller determines that curing is complete, the book is automatically transported to the folding station. The folding device folds the book according to the preset crease positions and number of folds, and finally outputs the finished product.

[0050] Specifically, such as Figures 1 to 2 As shown, the laser displacement sensor array 5 is used to continuously collect displacement data of each measuring point along the spine length direction at a preset sampling frequency during the hot melt adhesive curing process. This sampling frequency is preset based on the dynamic characteristics of the curing process, requiring both the ability to capture rapid changes in curvature caused by adhesive shrinkage and the ability to exceed the real-time calculation capabilities of the controller. The displacement data of each measuring point reflects the vertical offset of the spine surface at that position relative to its initial flat state.

[0051] The controller is configured to calculate the curvature of each measuring point using a three-point difference method based on the displacement data of each measuring point. Specifically, for sensors arranged at equal intervals, the controller takes the displacement value of the current measuring point and its left and right adjacent measuring points, calculates the second-order difference, and divides it by the square of the interval to obtain an approximate value of the curvature at that point. For the first and last boundary measuring points, the forward or backward difference method is used for calculation. The absolute values ​​of the curvatures calculated from all measuring points are taken and averaged to obtain the average curvature deviation of the current spine. This average curvature deviation is the core feedback quantity for subsequent control, directly reflecting the degree of curvature of the spine.

[0052] Based on the aforementioned discrete state equation, the controller uses the average curvature deviation and the real-time wire tension fed back by the electromagnetic tension controller as the initial state, and iteratively predicts the curvature sequence within the next prediction time domain step. In each iteration, the controller calculates the influence of each physical effect on the curvature at the next moment based on the spatial diffusion term, tension change rate term, adhesive shrinkage term, and shrinkage acceleration term in the discrete state equation, and sums up the contributions of each term. During the iteration process, the controller simultaneously retrieves the curing shrinkage rate at each future moment from the pre-stored shrinkage rate calibration curve based on the curing time, and updates the wire tension by assuming candidate tension adjustment amounts, until the preset prediction time domain step is predicted, thereby obtaining the complete curvature sequence for multiple future control cycles.

[0053] Based on the adjustment cost function, the controller solves a quadratic programming problem under tension amplitude and rate of change constraints to obtain the optimal tension adjustment sequence. The cost function comprises the sum of squares of future curvature deviations and the sum of squares of the tension adjustment amounts, weighted by the curvature deviation weight matrix and the tension adjustment amount weight matrix, respectively. The tension amplitude constraint ensures that the wire tension does not exceed the safety range of the electromagnetic tension controller and the spine's bearing capacity, while the rate of change constraint prevents sudden tension changes from impacting the pages. By solving the quadratic programming problem, the controller obtains the optimal tension adjustment sequence that minimizes the cost function within the next number of control time steps, but only outputs the first adjustment amount in the sequence to the electromagnetic tension controller to change the real-time wire tension. The remaining adjustments are recalculated in the next control cycle. This rolling optimization strategy can promptly compensate for model errors and external disturbances.

[0054] The above-described steps of data acquisition, calculation, prediction, solution, and adjustment are repeated within each control cycle until the conditions for the curing determination folding step are met. The curing determination conditions include: the average curvature deviation being less than a preset threshold for multiple consecutive sampling cycles, and the adhesive layer temperature detected by the adhesive temperature sensor dropping below a set value. When these conditions are met, the controller determines that curing is complete, locks the current wire tension, and sends a start signal to the folding device; otherwise, the controller automatically enters the next control cycle and continues to execute dynamic curing control.

[0055] With the above configuration, the laser displacement sensor array 5, electromagnetic tension controller, adhesive temperature sensor, and controller form a closed-loop control system. The sensors provide real-time feedback, the controller executes a model predictive control algorithm, and the actuator precisely adjusts the wire tension, thereby actively suppressing spine bending during adhesive curing, achieving a deep integration of the device and method.

[0056] The foregoing has illustrated and described the basic features, principles, and advantages of the present invention. It should be noted that the present invention is not limited to the above embodiments, but only to some embodiments. Any improvements and additions made without departing from the spirit and scope of the present invention are considered to be within the scope of protection of the present invention.

Claims

1. A printing and binding method using adhesive thread lamination, characterized in that, include: The thread-binding process involves drilling holes along the spine edge of the book block and threading the book through them, then setting the initial tension of the binding thread using an electromagnetic tension regulator. In the adhesive coating step, hot melt adhesive is applied to the spine surface after thread binding, and displacement data of several monitoring points along the spine length direction are collected according to a preset sampling frequency. The curvature distribution of the spine surface is calculated based on the displacement data. The control parameter acquisition step involves obtaining the curing shrinkage rate at the current moment based on the adhesive layer temperature and the preset adhesive layer shrinkage rate calibration curve, obtaining the real-time tension of the thread through the feedback result of the electromagnetic tension regulator, obtaining the distance of the thread from the spine neutral axis and obtaining the preset book core equivalent bending stiffness, and using the acquired data set as tension control parameters. The dynamic solidification control steps involve constructing a coupled prediction model between a preset curvature target and the tension control parameters. In each control cycle, the curvature sequence for the subsequent control cycle is calculated using the coupled prediction model based on the current curvature distribution and the real-time tension of the wire. An adjustment cost function is constructed using tension amplitude constraints and rate of change constraints. The optimal tension adjustment amount sequence for the subsequent adjustment cycle is calculated in conjunction with the initial tension. The wire tension is then adjusted based on the optimal tension adjustment amount sequence. In the curing determination and folding step, the average deviation between the current curvature distribution and the target curvature is calculated. When the average deviation is less than a preset threshold for multiple consecutive sampling cycles and the adhesive layer temperature drops below the set value, the curing is determined to be complete and the book is sent to the folding station for folding. Otherwise, the process returns to the next control cycle of the dynamic curing control step and dynamic curing control is re-performed through the tension deviation adjustment strategy.

2. The printing and binding method using adhesive thread lamination according to claim 1, characterized in that, The adhesive coating step includes: collecting displacement data through laser displacement sensors, which are arranged in an equidistant array along the spine length direction; obtaining the displacement value of a point based on the position coordinates of the laser displacement sensors; calculating the curvature of the point based on the displacement value; and setting the average absolute value of the curvature of all laser displacement sensors as the average curvature deviation of the current spine.

3. The printing and binding method using glue and thread lamination according to claim 1, characterized in that, The calibration of the preset adhesive shrinkage rate calibration curve includes: applying the hot melt adhesive to be used to the test substrate, collecting the temperature change and corresponding volume shrinkage rate during the curing process of the adhesive layer in a constant temperature environment, fitting the collected temperature data and shrinkage rate data, establishing a functional relationship between shrinkage rate and temperature and time, generating the fitted shrinkage rate calibration curve and pre-storing it.

4. The printing and binding method using glue and thread lamination according to claim 1, characterized in that, The dynamic curing control steps include: the number of subsequent control cycles is determined by a prediction time domain and a control time domain; the prediction time domain is set according to the total curing time of the adhesive layer and the sampling period; the control time domain is less than or equal to the prediction time domain and is preset by offline simulation; the rate of change of curvature deviation is monitored during the curing process; when the rate of change exceeds a preset threshold, the control time domain is increased; when the curvature deviation tends to stabilize, the control time domain is decreased.

5. The printing and binding method using adhesive thread lamination according to claim 1, characterized in that, The dynamic curing control step further includes establishing a functional mapping relationship between the current curvature, real-time thread tension, curing shrinkage rate, equivalent bending stiffness of the book core, and the distance of the thread from the spine neutral axis. The functional mapping relationship is a discrete state equation, which includes a tension influence coefficient and a shrinkage influence coefficient. The tension influence coefficient is predetermined based on the equivalent bending stiffness of the book core and the distance of the thread from the spine neutral axis. The shrinkage influence coefficient is determined through offline calibration based on the type of adhesive and the thickness of the adhesive layer. The coupled prediction model is constructed by using the current curvature and real-time thread tension as the initial state and iteratively predicting the curvature value for each future control cycle based on the discrete state equation. In each iteration, the thread tension is updated based on the tension adjustment amount of the current cycle and substituted into the calculation for the next cycle. The tension adjustment amount is determined by the candidate value generated by the adjustment cost function in each iteration.

6. The printing and binding method using adhesive thread lamination according to claim 5, characterized in that, The weight matrix in the adjustment cost function includes a curvature deviation weight matrix and a tension adjustment weight matrix, which are pre-tuned through offline simulation combined with experiments.

7. The printing and binding method using glue and thread lamination according to claim 1, characterized in that, The tension deviation adjustment strategy includes the following steps: when the average deviation is greater than a preset threshold, the current curvature deviation is decomposed into a contraction contribution component, a tension contribution component, and a diffusion contribution component according to the discrete state equation of the coupled prediction model. The magnitude of the absolute value of each component is compared to determine the main source of the deviation. If the contraction contribution component is the largest, the ratio of the curvature deviation weight matrix to the tension adjustment amount weight matrix in the adjustment cost function is increased, and the tension adjustment amount for the next two control cycles is pre-compensated according to the current rate of change of contraction rate. If the tension contribution component is the largest, the weight ratio is kept unchanged, and the upper limit of the tension adjustment amount for the next control cycle is directly increased. If the diffusion contribution component is the largest, the prediction time domain is increased and the control time domain is decreased. The above adjustment actions are superimposed and output to the electromagnetic tension regulator.

8. The printing and binding method using glue and thread lamination according to claim 1, characterized in that, The control parameter acquisition step further includes applying a small tension pulse of preset amplitude through an electromagnetic tension regulator after the binding is completed and before the gluing step begins, recording the transient curvature response of the spine measured by the laser displacement sensor array, calculating the equivalent bending stiffness of the current book block based on the peak ratio of the tension pulse amplitude to the curvature response, combined with the distance of the thread from the neutral axis of the spine, and updating the pre-stored stiffness value in the controller.

9. A printing and binding machine using glue and thread lamination, applicable to the printing and binding method using glue and thread lamination as described in any one of claims 1 to 8, characterized in that, include: Printer (1); A punching device (2) is installed on the side wall of the printer (1) for drilling through holes along the spine edge of the book block; The composite binding head (3) is mounted on the mounting rod (4) on the side wall of the printer (1). The composite binding head (3) integrates a hot melt adhesive spray head (31), a wire guide pin (32) and an electromagnetic tension controller. A laser displacement sensor array (5) is set above the binding station and is arranged at equal intervals along the spine length direction to collect displacement data of the spine surface. The adhesive temperature sensor is located at the outlet of the hot melt adhesive spray nozzle to collect the temperature of the adhesive layer. The controller is electrically connected to the electromagnetic tension controller, the laser displacement sensor array, the adhesive temperature sensor, and the hot melt adhesive spray head, respectively. A folding device (6) is installed on the lower surface of the substrate and is used to fold the bound book into shape.

10. A printing and binding machine with adhesive thread lamination according to claim 9, characterized in that, The laser displacement sensor array is used to continuously collect displacement data of each measuring point along the spine length direction at a sampling frequency of 100Hz to 200Hz during the hot melt adhesive curing process; the controller is configured to calculate the curvature of each measuring point using the three-point difference method based on the displacement data of each measuring point, and take the average absolute value of the curvature of all measuring points as the average curvature deviation of the current spine. Based on the discrete state equation, using the average curvature deviation and the real-time tension of the wire fed back by the electromagnetic tension controller as the initial state, the curvature sequence within the future prediction time domain step is iteratively predicted. Based on the adjustment cost function, a quadratic programming problem is solved under the constraints of tension amplitude and rate of change to obtain the optimal tension adjustment sequence. Only the first adjustment is output to the electromagnetic tension controller to change the wire tension. Repeat the above steps of data collection, calculation, prediction, solution, and adjustment until the conditions for the solidification determination step are met.

Images