Data acquisition and monitoring control system for corrugated board production line

Abstract

This invention relates to the field of advanced metal materials and surface engineering, specifically a data acquisition and monitoring control system for a corrugated cardboard production line. It includes a data acquisition terminal, a control server, and an execution terminal. The system acquires work orders and sensor data, constructs an ideal coupled benchmark model, and generates a set of real residual data. Its core is to compare the real residuals with theoretical residuals based on a fault mechanism library, and execute a graded judgment logic: high similarity indicates equipment failure triggering shutdown and maintenance; low similarity but large statistical fluctuations indicate raw material fluctuations triggering process compensation. This invention effectively distinguishes between mechanical component defects and raw material quality problems, avoiding indiscriminate shutdowns and significantly improving the operation and maintenance efficiency of the production line.

Description

Technical Field

[0001] This invention relates to the field of advanced metal materials and surface engineering, specifically to a data acquisition and monitoring control system for corrugated cardboard production lines. Background Technology

[0002] In automated corrugated cardboard production environments, production lines involve complex thermodynamic and dynamic multi-physics coupling. Each network element needs to frequently adjust its speed and process parameters according to work order instructions, and the sensor network generates continuous time-series data in real time, including multiple dimensions such as position, temperature, tension, and current. To monitor the status of this data, existing solutions generally adopt passive monitoring modes based on static thresholds or simple statistical characteristics. This involves directly collecting real-time sensor data and comparing it with preset fixed upper and lower limits, or judging production line anomalies solely through single-dimensional fluctuation analysis. While this approach is suitable for steady-state production scenarios... While the existing system possesses certain monitoring capabilities, the corrugated cardboard production process exhibits significant nonlinear and time-varying characteristics. Dynamic acceleration and deceleration of the machine speed or changes in work orders often cause large fluctuations in normal data. Furthermore, natural fluctuations in raw material properties such as basis weight and moisture content introduce background noise. This makes it difficult for the existing solution to effectively separate the subtle fault characteristics of the equipment itself from the fluctuation characteristics of raw materials and normal process dynamics. Consequently, false alarms frequently occur when operating conditions change, or the root cause of the anomaly cannot be accurately distinguished as a defect in mechanical parts or a problem with the quality of raw materials. Consequently, it is difficult to support accurate shutdown maintenance decisions and closed-loop process compensation.

[0003] Therefore, how to construct a benchmark model that adapts to dynamic operating conditions and achieve precise decoupling and differentiated control of equipment failures and raw material fluctuations, so as to improve the accuracy of production line fault diagnosis and the efficiency of automated operation and maintenance, has become an urgent technical problem to be solved. Summary of the Invention

[0004] In view of the shortcomings of the prior art, the purpose of this invention is to provide a data acquisition and monitoring control system for corrugated cardboard production lines, which can solve the technical problems existing in the prior art. Specifically, the technical solution of this invention includes:

[0005] Data acquisition terminal, control server, and execution terminal;

[0006] The data acquisition terminal is configured to acquire work order instruction data and real-time sensor data of the corrugated cardboard production line, and send the work order instruction data and the real-time sensor data to the control server.

[0007] The control server is configured to construct an ideal coupled benchmark model representing the standard operating state of the production line based on the work order instruction data, and generate a set of real residual data based on the ideal coupled benchmark model and the real-time sensing data.

[0008] The control server is also configured to generate a theoretical residual data set based on a preset fault mechanism library, and to calculate the similarity value between the actual residual data set and the theoretical residual data set.

[0009] The control server is also configured to execute the following determination logic: if the similarity value is greater than or equal to a preset similarity threshold, determine that there is a equipment failure in the corrugated cardboard production line and send a shutdown maintenance instruction to the execution terminal; if the similarity value is less than the preset similarity threshold and the statistical fluctuation value of the actual residual data set is greater than a preset fluctuation threshold, determine that there is a raw material fluctuation in the corrugated cardboard production line and send a process compensation instruction to the execution terminal.

[0010] The execution terminal is configured to execute the shutdown maintenance command or the process compensation command.

[0011] Preferably, the control server is further used for:

[0012] Based on the paper grade parameters, weight parameters, and predetermined speed parameters in the work order instruction data, the heat absorption curve and tension transfer function of the paper under ideal conditions are calculated using preset thermodynamic and kinetic formulas.

[0013] The heat absorption curve and the tension transfer function are used as the ideal coupling reference model.

[0014] Preferably, the control server is further used for:

[0015] Extract fault mechanism parameters from the fault mechanism library;

[0016] The fault mechanism parameters are injected into the ideal coupled benchmark model to generate simulation data containing specific fault characteristics;

[0017] The difference between the simulation data and the ideal coupled benchmark model is calculated to generate the theoretical residual data set.

[0018] Preferably, the control server is further used for:

[0019] Extract the coating roller jump disturbance function from the fault mechanism parameters;

[0020] The glue roller bounce disturbance function is injected into the glue quantity model part of the ideal coupled reference model to generate the first type of simulation data.

[0021] Alternatively, extract the thermal conductivity attenuation coefficient from the fault mechanism parameters;

[0022] The thermal conductivity attenuation coefficient is injected into the temperature model part of the ideal coupled reference model to generate the second type of simulation data.

[0023] Preferably, the control server is further used for:

[0024] Calculate the difference between the real-time sensing data and the ideal coupled benchmark model to generate the real residual data set;

[0025] Calculate the cross-correlation function value or dynamic time-warped distance between the actual residual data set and the theoretical residual data set;

[0026] Based on the cross-correlation function value or the dynamic time warping distance, the similarity value between the actual residual data set and the theoretical residual data set is determined.

[0027] Preferably, the execution terminal is further used for:

[0028] In response to receiving the process compensation instruction, the glue application parameters or speed control parameters of the corrugated cardboard production line are adjusted to perform closed-loop process compensation.

[0029] Preferably, the data acquisition terminal includes: a tension sensor, a temperature sensor, a current transformer, and a cardboard flatness detection module;

[0030] The tension sensor is used to collect the tension value of each section of paper.

[0031] The temperature sensor is used to collect the surface temperature of the preheating cylinder;

[0032] The current transformer is used to collect the current waveform of the main motor;

[0033] The cardboard flatness detection module is used to collect flatness readings at the cardboard exit.

[0034] Preferably, the system further includes an expert terminal;

[0035] The expert terminal is used to send new fault mechanism parameters to the control server;

[0036] The control server is also used to receive the new fault mechanism parameters and update the fault mechanism library.

[0037] Preferably, the execution terminal is further used for:

[0038] The equipment fault types corresponding to the shutdown maintenance instructions are displayed visually.

[0039] The parameter adjustment process corresponding to the process compensation command is visualized.

[0040] Compared with the prior art, the present invention has the following improvements and advantages:

[0041] 1. This invention constructs an ideal coupled benchmark model through a control server, and generates a set of real-world residual data by differential analysis between real-time sensor data and this model. This residual data is then compared with a set of theoretical residual data generated based on a fault mechanism library. The system executes a hierarchical judgment logic: when the similarity value is high, it accurately identifies the fault as equipment failure and triggers a shutdown for maintenance; when the similarity value is low but the statistical fluctuation value is large, it is determined to be raw material fluctuation and triggers process compensation. This logic effectively solves the technical problem of existing technologies being unable to distinguish whether the root cause of an anomaly is a defect in mechanical parts or a problem with the quality of raw materials, avoiding blind shutdowns or operation with defects, and improving the operational efficiency of the production line.

[0042] 2. This invention utilizes paper grade, weight, and predetermined speed parameters from work order instruction data, combined with thermodynamic and kinetic formulas to construct a dynamic, ideally coupled benchmark model. This model can reflect the ideal state of the production line in real time during standard acceleration / deceleration or process switching, and as a differential benchmark, it can effectively filter out normal data fluctuations caused by dynamic changes in vehicle speed. By comparing the actual residuals with the theoretical residuals, the system can sensitively capture abnormal signals in complex nonlinear production environments, significantly reducing the false alarm rate caused by changes in operating conditions.

[0043] 3. This invention does not rely on simple statistical data features, but extracts specific fault mechanism parameters from a fault mechanism library, such as the roller vibration disturbance function or the thermal conductivity attenuation coefficient, and injects them into an ideal coupled benchmark model to generate a theoretical residual data set containing specific fault characteristics. By calculating the similarity between the actual residuals and these theoretical residuals, the system can identify specific fault modes like fingerprint comparison. This mechanism enables the system not only to detect anomalies, but also to specifically diagnose whether it is a vibration fault in the mechanical transmission chain or an efficiency attenuation fault in the thermodynamic system, providing maintenance personnel with a clear direction for fault diagnosis.

[0044] 4. After receiving the process compensation instruction, the execution terminal of the present invention can automatically adjust the glue application amount parameter or the speed control parameter. This closed-loop control strategy changes the traditional passive monitoring mode, enabling the production line to dynamically correct the process parameters to offset the impact on the final product quality when it detects natural fluctuations in the raw materials. This ensures the continuous and stable operation of the production line while maintaining the flatness and bonding quality of the cardboard. Attached Figure Description

[0045] Figure 1 This is a structural diagram of the system of the present invention. Detailed Implementation

[0046] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments and accompanying drawings.

[0047] Example 1:

[0048] Please see Figure 1 The data acquisition and monitoring control system for the corrugated cardboard production line includes:

[0049] The system includes a data acquisition terminal, a control server, and an execution terminal. The data acquisition terminal is configured to acquire work order instruction data and real-time sensor data from the corrugated cardboard production line, and send the work order instruction data and real-time sensor data to the control server.

[0050] The control server is configured to construct an ideal coupled benchmark model representing the standard operating state of the production line based on work order instruction data, and generate a set of real residual data based on the ideal coupled benchmark model and real-time sensor data; the control server is also configured to generate a set of theoretical residual data based on a preset fault mechanism library, and calculate the similarity value between the set of real residual data and the set of theoretical residual data.

[0051] The control server is also configured to execute the following judgment logic: if the similarity value is greater than or equal to the preset similarity threshold, determine that there is a equipment failure in the corrugated cardboard production line, and send a shutdown maintenance command to the execution terminal;

[0052] If the similarity value is less than the preset similarity threshold, and the statistical fluctuation value of the actual residual data set is greater than the preset fluctuation threshold, it is determined that there is raw material fluctuation in the corrugated cardboard production line, and a process compensation instruction is sent to the execution terminal; the execution terminal is configured to execute a shutdown maintenance instruction or a process compensation instruction.

[0053] This embodiment elaborates on the core control logic of the above system, which is based on synthetic analysis to construct a closed loop. The system uses a hardware sensor network, i.e., data acquisition terminals, configured at each key station of the corrugated cardboard production line to sense the operating status of the physical world in real time and acquire real-time sensor data. Simultaneously receive work order instruction data ;

[0054] The control server serves as the computing core, constructing an ideal coupled benchmark model based on the thermodynamic and kinetic equations of the corrugated cardboard forming mechanism. Specifically, the construction process includes the following steps: establishing the differential equation for paper temperature rise:

[0055]

[0056] in, For paper temperature, The convective heat transfer coefficient is obtained by retrieving its value from a pre-set material thermal property database based on the paper grade parameters in the work order. The effective heat exchange area of ​​the paper in the current heating zone is obtained by multiplying the length of the heating plate fixed in the equipment by the paper width specified in the work order. The specific heat capacity of the paper was also obtained based on the paper grade index material library; To determine the real-time quality of the paper within this section; establish the discrete equation for paper web tension transmission:

[0057]

[0058] in, For tension, For Young's modulus, The time step for the simulation calculation is set to 10ms. This value is selected based on Shannon's sampling theorem to ensure that it is more than twice the natural frequency of the system's mechanical components, in order to prevent simulation aliasing. The span between rollers in the tension control section is a fixed mechanical dimension measured during equipment installation; The cross-sectional area of ​​the paper is calculated from the work order weight and the paper width. and These are the theoretical linear velocities at the exit and entrance of the section, respectively. Both are calculated based on the predetermined speed curve in the work order, rather than actual measured values. This represents the tension value flowing into this section;

[0059] The system employs a time-domain iterative solution method, such as the forward Euler method, with a step size of [missing information]. Set to 10ms, solve the above equations simultaneously to obtain the ideal numerical sequence. The model takes work order parameters as input and outputs a theoretical numerical sequence under the assumptions of no equipment failures and perfect raw materials. The system executes dual residual generation logic and outputs the standard output. With real-time sensor data Differentiation is performed to obtain a mixed residual signal that includes equipment failure characteristics, raw material fluctuation characteristics, and environmental noise, i.e., the actual residual data set. The calculation formula is:

[0060]

[0061] Based on this, the control server extracts specific fault parameters from a pre-set fault mechanism library and injects them into an ideally coupled benchmark model to generate simulation data. Then, the difference between it and the ideal benchmark is calculated to obtain a pure residual signal containing only a specific single fault characteristic, i.e., the theoretical residual data set. The calculation formula is:

[0062]

[0063] In particular, considering that the fault mechanism library contains Different types of preset fault modes, such as bearing wear, roller eccentricity, heater open circuit, etc., and indexed as follows: The process of calculating similarity scores is configured as an iterative optimization matching process: the control server traverses and generates... Group of theoretical residual datasets Calculate the actual residuals respectively With each group similarity , here Calculated using the normalized cross-correlation function:

[0064]

[0065] To eliminate the influence of signal amplitude differences and perform max pooling, select... The final similarity score will be used in subsequent determinations, and the corresponding first similarity score will be used in the subsequent determinations. The fault type is marked as a candidate fault type; this step ensures that the system can identify the specific fault that best matches the current signal characteristics from a variety of possible theoretical fault types.

[0066] The system is based on this optimal similarity value Perform hierarchical determination: Simultaneously, the system calculates the actual residual data set. Statistical fluctuation value Its specific definition is The standard deviation within the current sliding window is used to characterize the discrete energy of the signal, and a preset fluctuation threshold is set based on the background noise level of the device under no-load conditions. The method for determining the background noise level is as follows: While the production line is idling and the machine speed is maintained at a constant low speed, sensor data is continuously collected for more than 5 minutes. The standard deviation is calculated as the baseline noise value, and then... Set to 3 to 5 times the reference noise value to cover the effects of environmental micro-vibrations;

[0067] For the preset similarity threshold The method for setting it is to select the lower limit of the confidence interval based on the statistical distribution of the similarity between historically confirmed fault samples and theoretical residuals, such as a value of 0.8, or to use the optimal critical point determined by the receiver operating characteristic curve, in order to balance the fault detection rate and the false alarm rate.

[0068] Based on this, the system performs a hierarchical determination: in response to Greater than or equal to the preset similarity threshold The system determines that the actual abnormal state highly matches the known fault, identifies it as a device malfunction, and sends a shutdown maintenance command; in response to Less than and Statistical fluctuation value Greater than the preset fluctuation threshold The system deduced that the fluctuation originated from a raw material issue, identified it as a raw material fluctuation, and sent a process compensation instruction.

[0069] This embodiment effectively filters out normal data fluctuations caused by changes in vehicle speed or work orders by introducing an ideal coupled benchmark model as a differential benchmark. By comparing the topological similarity between the actual residual and the theoretical residual, it solves the problem of false alarms caused by poor raw paper quality in the corrugated cardboard production scenario, and realizes precise decoupling and differentiated control of minor equipment faults and natural fluctuations in raw materials.

[0070] Example 2:

[0071] The control server is also used to: calculate the heat absorption curve and tension transfer function of paper under ideal conditions based on the paper grade parameters, weight parameters and predetermined speed parameters in the work order instruction data, using preset thermodynamic and kinetic formulas; and use the heat absorption curve and tension transfer function as an ideal coupling reference model.

[0072] This embodiment details the specific steps of the control server in constructing an ideal coupled benchmark model using physical formulas, focusing on solving the convergence problem of physical field coupling calculations during continuous production; the control server parses work order instruction data and extracts paper grade parameters, weight parameters, and predetermined speed parameters. ;

[0073] Here, the preset vehicle speed parameters are... It is not a single scalar, but rather a time-series trajectory that is parsed to conform to the standard acceleration and deceleration logic of the production line. The standard acceleration / deceleration logic employs an S-shaped velocity curve planning algorithm, generating a smooth speed control command sequence based on preset maximum acceleration and jerk, ensuring that the ideal model can dynamically reflect the physical response of the equipment during standard start-up and shutdown phases. For the heat absorption curve, the system introduces a discretized temperature rise model, solved using the fourth-order Runge-Kutta method.

[0074]

[0075] in, The set constant temperature for the hot plate comes from the process setting parameters in the work order instruction, such as setting it to 170 degrees Celsius;

[0076] The effective heat exchange area of ​​the paper on the hot plate is calculated using the following formula:

[0077]

[0078] To control the real-time quality of the heated paper inside the body, the calculation formula is modified as follows:

[0079]

[0080] in, The effective heating length of the hot plate. The width of the paper; in the formula This refers to the weight parameter extracted from the work order instruction data, in units of... The specific heat capacity in the formula The grade is determined by referring to the preset material property table based on the paper grade parameters in the work order; the formula is divided by... This is to change the units of the calculation results from Convert to To ensure consistency with the standard SI unit of specific heat capacity parameter in subsequent thermodynamic formulas;

[0081] To address the convective heat transfer coefficient in the model The problem of tracing the value of a given value is addressed in this embodiment. Instead of being a fixed constant, it is set as a function that changes dynamically with vehicle speed:

[0082]

[0083] in, That is, the reference heat transfer coefficient, and The air entrainment influence factors are all empirical constants indexed from a pre-defined thermodynamic parameter database based on paper grade parameters, such as surface roughness grade. To ensure that the above empirical formulas satisfy the principle of homogeneity of physical dimensions, given... The unit is m / s. The dimensions are clearly defined as To offset the velocity term The dimensionality of the power is affected, thus ensuring the dimensionless characteristics of the operands within the parentheses, thereby accurately simulating the hindering effect of the air film on heat conduction during high-speed paper feeding.

[0084] Reference heat transfer coefficient and air-borne influencing factors The method for determining the paper web temperature rise data at different steady-state speeds is as follows: During the equipment debugging phase, infrared thermal imagers are used to record paper web temperature rise data at different steady-state speeds to construct a speed-temperature rise dataset. The least squares method is used to perform regression analysis on the above functional relationship to calculate the empirical constants corresponding to specific paper grades and store them in the database.

[0085] As an example of specific implementation parameters, for conventional Class B corrugated paper, The range of values ​​is to , The range of values ​​is to Furthermore, the method for determining the aforementioned constants is as follows: During the equipment debugging phase, paper web temperature rise data at different steady-state machine speeds are recorded using an infrared thermal imager, and the above functional relationship is fitted and calibrated using the least squares method; Regarding the tension transfer function, in order to overcome the numerical divergence problem caused by the accumulation of speed differences in the traditional static integral model under continuous production scenarios, the system adopts discretization modeling based on the differential equation of moving paper web tension with mass conservation:

[0086]

[0087] in, Current time step The ideal tension output value, in Newtons; The tension state at the previous time step reflects the system's memory characteristics; Simulation time step, for example, 0.01 seconds; The current tension control section's inter-roller span length, i.e., the physical distance between the shafts of the two drive rollers, is expressed in meters. : Depending on the current temperature The real-time changing thermally coupled Young's modulus is expressed by the following formula:

[0088]

[0089] Among them, standard Young's modulus With thermal softening coefficient All were obtained by indexing the material property database based on the paper grade parameters in the work order; here, the thermal softening coefficient... The dimensions are set as The method for determining it is as follows: In a laboratory environment, multiple sets of tensile tests are conducted on the same paper sample at different temperatures, and the Young's modulus is obtained by fitting the linear slope of the change with temperature.

[0090] Paper cross-sectional area, system based on basis weight parameter Density corresponding to paper grade The density was calculated. The dimensions are The method for determining this is to measure the paper thickness using a standard thickness gauge and then calculate it based on the basis weight parameter.

[0091]

[0092] Calculate the cross-sectional area Note that unit conversions are necessary; the calculation formula is as follows:

[0093]

[0094] This enables the mapping from work order parameters to model geometric parameters; The theoretical linear velocity at the entrance of the current tension control section is not derived from the real-time encoder, but is based on the vehicle speed parameters predetermined by the work order. The calculation is as follows:

[0095]

[0096] in, This is the mechanical transmission ratio coefficient of the inlet roller in this section; The theoretical linear velocity at the exit of the current tension control section is also calculated based on the work order parameters, and the formula is:

[0097]

[0098] in, This refers to the output roller drive ratio. A standard tensile rate is set for the process, such as 0.2%, to ensure that the ideal model only reflects the tension state under standard process conditions and is decoupled from actual sensor noise;

[0099] : Tension outflow term, which physically means the strain energy carried away by the movement of the paper. This is a key damping term to ensure the convergence of the model in steady state. The tension value input from the upstream section is initially set to 0; the system will iteratively calculate the obtained tension value. sequence sum The sequence was established as an ideal coupled baseline model;

[0100] This embodiment achieves accurate and stable numerical simulation of the thermo-mechanical-velocity multiphysics coupling effect during continuous paper feeding by introducing a differential equation containing convection transport terms.

[0101] Example 3:

[0102] The control server is also used to: extract fault mechanism parameters from the fault mechanism library; inject the fault mechanism parameters into the ideal coupled benchmark model to generate simulation data containing specific fault characteristics; and calculate the difference between the simulation data and the ideal coupled benchmark model to generate a theoretical residual data set.

[0103] The control server is also used to: extract the glue roller runout disturbance function from the fault mechanism parameters; inject the glue roller runout disturbance function into the glue quantity model part of the ideal coupled benchmark model to generate the first type of simulation data; or, extract the heat conduction efficiency attenuation coefficient from the fault mechanism parameters; inject the heat conduction efficiency attenuation coefficient into the temperature model part of the ideal coupled benchmark model to generate the second type of simulation data.

[0104] This embodiment details the digital twin fault simulation process that generates theoretical residuals through fault mechanism parameter injection. The control server extracts mathematical operators or functions, i.e., fault mechanism parameters, from the fault mechanism library after abstracting mechanical or electrical faults. For the eccentric runout fault of the coating roller caused by bearing wear, the system extracts the coating roller runout disturbance function. The glue volume model is then injected, and the calculation formula is as follows:

[0105]

[0106] in, The simulated glue application amount under fault conditions is derived from calculation output and its physical meaning is the glue application amount waveform containing fluctuation characteristics. Ideal glue application amount, derived from the benchmark model, physically means the glue application amount set under standard process; : Runout amplitude coefficient, which is derived from empirical values ​​in the fault mechanism library, and its physical meaning is a quantitative representation of the degree of bearing wear; The disturbance angular frequency originates from the real-time vehicle speed. Linear dependence, the specific calculation logic is as follows:

[0107]

[0108] in, The preset radius parameter for the coating roller is derived from the equipment mechanical design drawings or spare parts specifications to ensure that the frequency characteristics are strictly coupled with the mechanical speed. Its physical meaning is the frequency characteristics of the fault occurrence, and the unit is radians per second. Initial phase angle, which is either randomly generated or preset based on the roller installation angle, physically represents the offset of the fault occurrence time relative to the reference time, and is expressed in radians. This is the vibration amplitude coefficient, whose value is determined according to the ISO 10816 mechanical vibration standard. Different vibration intensities are mapped to dimensionless percentage fluctuations in adhesive volume, such as 0.05 to 0.30, and stored in the fault mechanism library. Alternatively, for faults caused by hot plate water accumulation due to drain valve blockage, the system extracts the heat transfer efficiency attenuation coefficient. And inject it into the temperature model, and calculate as follows:

[0109]

[0110] in, The heat transfer coefficient under fault conditions is derived from calculation output and its physical meaning is the heat transfer capacity after damage. The reference heat transfer coefficient is derived from the fixed design parameters of the equipment under calibrated operating conditions or the theoretical value under normal conditions. Its physical meaning is the standard heat transfer efficiency when it is not affected by faults. The unit is watts per square kelvin, which distinguishes it from the dynamically changing comprehensive heat transfer coefficient. Efficiency decay factor, sourced from a fault mechanism library, with a value range of... The physical meaning is the percentage increase in thermal resistance;

[0111] The system calculates the difference between the simulation data and the ideal baseline, for example... or This generates waveform data containing only specific fault frequency characteristics. In this embodiment, standard fingerprints for various faults are generated by actively injecting fault parameters. Instead of relying on scarce real fault history data, fault samples are synthesized through positive physical simulation, which effectively solves the problem of insufficient negative samples in industrial big data, which makes model training difficult.

[0112] Example 4:

[0113] The control server is also used for:

[0114] Calculate the difference between real-time sensing data and the ideal coupled baseline model to generate a set of real residual data; calculate the cross-correlation function value or dynamic time warping distance between the set of real residual data and the set of theoretical residual data; and determine the similarity value between the set of real residual data and the set of theoretical residual data based on the cross-correlation function value or dynamic time warping distance.

[0115] This embodiment illustrates the specific steps for quantifying the similarity between actual and theoretical residuals using statistical signal processing algorithms, with a particular emphasis on adding zero-variance protection logic for steady-state bias faults. To eliminate the nonlinear distortion of the time-domain signal caused by frequent acceleration and deceleration during corrugated cardboard production, the system performs time-space mapping transformation preprocessing: using encoder pulses mounted on the speed measuring roller as the sampling trigger source, a spatial index sequence is constructed.

[0116] ( )

[0117] The system processes real-time sensing data in the time domain. and through the integral ideal velocity curve The ideal model output obtained by mapping Conduct based on Interpolation resampling generates spatial domain sequences. and Perform differential calculation:

[0118]

[0119] This step ensures strict alignment of the residual calculations in physical location; to address the issue that traditional cross-correlation functions, when handling DC faults such as constant temperature differences caused by thermal efficiency decay, have theoretical residual variances approaching zero, leading to numerical calculation crashes with a denominator of 0, the system executes spectral energy diversion judgment logic:

[0120] Calculate the theoretical residual data set of the current traversal. Spatial domain variance:

[0121]

[0122] like Instead, an improved normalized Euclidean similarity method was adopted. :

[0123]

[0124] in, and These are the arithmetic means of the actual residual data set and the theoretical residual data set, respectively. and These are the standard deviations of the two, respectively. For quasi-static DC faults, In this embodiment, the symbol is defined. Characterizing statistical standard deviation, defining the symbol The numerical values ​​representing similarity have different physical meanings. and These are the preset mean bias weight and volatility penalty weight, respectively, for example, both set to 0.5;

[0125] The specific values ​​of the weights were determined using the analytic hierarchy process (AHP), calculated based on a scoring matrix of process experts' assessments of the importance of average deviation and fluctuation deviation in fault diagnosis; and Dimensional , Dimensional To ensure that the denominator is dimensionless; and and The numerical value includes a dimensionless conversion factor that makes the denominator dimensionless, for example, the units are respectively or This is to ensure that the terms in the denominator of the formula satisfy the principle of homogeneity of physical dimensions.

[0126] This formula introduces a fluctuation penalty term. This ensures that when the real signal exhibits drastic fluctuations while the theoretical signal is a stable DC signal, even if their means are equal, the calculation results will be consistent. It will also decrease significantly as the denominator increases, thus avoiding misjudging high-frequency noise as a DC fault and achieving keen identification of signal morphology differences; if The fault model was determined to be a dynamic AC fault, and a modified normalized cross-correlation function was used:

[0127]

[0128] in, To prevent the division by zero minimum value, for example The system covers the window with the largest roller circumference. Inner sliding calculation The maximum value is taken as the similarity score.

[0129] If dynamic time warping is used, the system calculates the cumulative distance. Then, the specific calculation logic is as follows: construct the cumulative distance matrix. , of which elements Represents the th in the real residual sequence The point and the theoretical residual sequence in the th The optimal path distance to a point is solved using the following recursive equation:

[0130]

[0131] in, For local Euclidean distance, the boundary conditions are set as follows: Ultimately As the total cumulative distance; based on this, using the function Map distance to The interval is defined to ensure consistency in the units of measurement for all similarity metrics; the control server outputs the maximum score calculated above as the final similarity value. .

[0132] Example 5:

[0133] The execution terminal is also used for:

[0134] In response to receiving a process compensation instruction, the glue application parameters or speed control parameters of the corrugated cardboard production line are adjusted to perform closed-loop process compensation.

[0135] This embodiment describes the closed-loop compensation strategy of the execution terminal when it determines that there is a fluctuation in raw materials, and focuses on supplementing the logic for preventing integral saturation and verifying control effectiveness; the system receives process compensation instructions containing residual characteristics and uses a sensor identifier (ID) mapping table to divide the residuals into thermodynamic subsets. and dynamic subset And calculate the statistical fluctuation value. and ;

[0136] To prevent the closed-loop control system from falling into an infinitely saturated integral state due to unavoidable raw material defects, the system introduces an incremental accumulation limit and performance backtracking mechanism: setting a dominant factor. Execute the following branches:

[0137] Branch 1, dominated by adhesion risk, namely Calculate the increase in adhesive application amount:

[0138]

[0139] in, To control the gain of the adhesive coating, the dimension is... This refers to the amount of adhesive compensation corresponding to a unit tension fluctuation deviation. The method for determining this is as follows: Apply a step adhesive interference in the open-loop state of the production line, measure the tension response, take its inverse response ratio as the initial value, and optimize it using the Ziegler-Nichols closed-loop tuning method: gradually increase... Record the critical gain until the system exhibits constant-amplitude oscillations. ,set up To ensure the robustness and rapid response capability of the control system;

[0140] Based on this, the system checks the cumulative compensation amount. :

[0141] like , Forced clamping is applied to the maximum permissible bias, such as 20% of the standard value. And send an alarm to the expert terminal that the raw material defect exceeds the compensation limit;

[0142] Otherwise, issue an adjustment command and start a performance backtracking timer: after the transmission lag time has elapsed. Then, resample and calculate the new fluctuation value. ;like If the fluctuation does not decrease, the compensation is deemed invalid, and the system automatically performs a parameter rollback operation to prevent invalid excessive glue application from causing the cardboard to soften.

[0143] Branch Two: Heat Absorption Risk Dominates, i.e. Calculate vehicle speed increment:

[0144]

[0145] in, For vehicle speed control gain, the dimension is The critical gain, i.e., the speed adjustment corresponding to a unit temperature fluctuation deviation, is determined as follows: based on the first-order inertial hysteresis model of hot plate heat transfer, the critical gain is calculated using the following formula. :

[0146]

[0147] in, The thermal inertia time constant, in seconds; For the specific heat capacity of paper, Paper weight per unit area This is the rated power gain of the heater. To detect pure time delay; the system is set To ensure system margin; the same cumulative limit and performance backtracking logic is executed to ensure that the vehicle speed is not infinitely reduced below the stop line due to thermal fluctuations;

[0148] The final execution instructions are written to the underlying driver via the programmable logic controller bus. The system repeats the above closed-loop process of measurement-determination-compensation-verification in each control cycle, realizing safe, bounded and reversible adaptive process compensation for raw material fluctuations.

[0149] Example 6:

[0150] The data acquisition terminal includes: a tension sensor, a temperature sensor, a current transformer, and a paperboard flatness detection module; among them, the tension sensor is used to collect the tension value of each section of paper; the temperature sensor is used to collect the surface temperature of the preheating cylinder; the current transformer is used to collect the current waveform of the main motor; and the paperboard flatness detection module is used to collect the flatness reading at the paperboard exit.

[0151] This embodiment details the hardware configuration of the data acquisition terminal and its acquisition purpose. The system is equipped with a tension sensor between the paper holder and the preheating cylinder to collect the tension values ​​of each section of paper in real time, which serves as the calibration input for the dynamic model. The system is preferably equipped with an infrared non-contact temperature sensor, which is aimed at the surface of the preheating cylinder and the hot plate to collect the surface temperature for monitoring thermodynamic boundary conditions. At the same time, a current transformer is connected in series in the main motor power supply circuit to collect the current waveform and use the high-frequency harmonic components of the current to analyze the wear information of the mechanical transmission chain.

[0152] A cardboard flatness detection module is installed at the exit of the cross-cutting machine. A laser displacement sensor array is used to collect the flatness reading of the cardboard, i.e., the warpage, as a feedback indicator of the final product quality. This embodiment uses multi-dimensional sensor fusion to perceive the status of the production line from four physical dimensions: heat, force, electricity, and light, providing complete and multi-source heterogeneous data support for the construction of the upper-level physical model.

[0153] Example 7:

[0154] The system also includes an expert terminal; the expert terminal is used to send new fault mechanism parameters to the control server; the control server is also used to receive new fault mechanism parameters and update the fault mechanism library; the execution terminal is also used to: visualize the equipment fault types corresponding to shutdown maintenance commands; and visualize the parameter adjustment process corresponding to process compensation commands.

[0155] This embodiment details the system's expert interaction and visualization functions; the system is configured with an expert terminal, allowing process experts to abstract new fault cases into mathematical parameters, such as the characteristic frequency of a new type of bearing, and send them to the control server; the control server receives the parameters and updates the fault mechanism library, enabling the system to have continuous learning capabilities;

[0156] Simultaneously, the terminal executes visualization logic: in response to shutdown maintenance commands, it displays the equipment fault type and the confidence level of the fault, i.e., the similarity value, on the screen; in response to process compensation commands, it displays the dynamic process of parameter adjustment, such as showing that the vehicle speed is being reduced from 150m / min to 142m / min to compensate for high moisture content; the system provides transparent decision-making basis to operators through the interface; this embodiment realizes the dynamic evolution of the system knowledge base through expert parameter injection and transparent display of the control process, and transforms complex algorithm decisions into intuitive information that operators can understand, thereby improving the efficiency and trust of human-machine collaboration.

[0157] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A data acquisition and monitoring control system for a corrugated cardboard production line, characterized in that, include: Data acquisition terminal, control server, and execution terminal; The data acquisition terminal is configured to acquire work order instruction data and real-time sensor data of the corrugated cardboard production line, and send the work order instruction data and the real-time sensor data to the control server. The control server is configured to construct an ideal coupled benchmark model representing the standard operating state of the production line based on the work order instruction data, and generate a set of real residual data based on the ideal coupled benchmark model and the real-time sensing data. The control server is also configured to generate a theoretical residual data set based on a preset fault mechanism library, and to calculate the similarity value between the actual residual data set and the theoretical residual data set. The control server is also configured to execute the following determination logic: if the similarity value is greater than or equal to a preset similarity threshold, determine that there is a equipment failure in the corrugated cardboard production line and send a shutdown maintenance instruction to the execution terminal; if the similarity value is less than the preset similarity threshold and the statistical fluctuation value of the actual residual data set is greater than a preset fluctuation threshold, determine that there is a raw material fluctuation in the corrugated cardboard production line and send a process compensation instruction to the execution terminal. The execution terminal is configured to execute the shutdown maintenance command or the process compensation command.

2. The system according to claim 1, characterized in that, The control server is also used for: Based on the paper grade parameters, weight parameters, and predetermined speed parameters in the work order instruction data, the heat absorption curve and tension transfer function of the paper under ideal conditions are calculated using preset thermodynamic and kinetic formulas. The heat absorption curve and the tension transfer function are used as the ideal coupling reference model.

3. The system according to claim 1, characterized in that, The control server is also used for: Extract fault mechanism parameters from the fault mechanism library; The fault mechanism parameters are injected into the ideal coupled benchmark model to generate simulation data containing specific fault characteristics; The difference between the simulation data and the ideal coupled benchmark model is calculated to generate the theoretical residual data set.

4. The system according to claim 3, characterized in that, The control server is also used for: Extract the coating roller jump disturbance function from the fault mechanism parameters; The glue roller jump disturbance function is injected into the glue quantity model part of the ideal coupled reference model to generate the first type of simulation data; Alternatively, extract the thermal conductivity attenuation coefficient from the fault mechanism parameters; The thermal conductivity attenuation coefficient is injected into the temperature model part of the ideal coupled reference model to generate the second type of simulation data.

5. The system according to claim 1, characterized in that, The control server is also used for: Calculate the difference between the real-time sensing data and the ideal coupled benchmark model to generate the real residual data set; Calculate the cross-correlation function value or dynamic time-warped distance between the actual residual data set and the theoretical residual data set; Based on the cross-correlation function value or the dynamic time warping distance, the similarity value between the actual residual data set and the theoretical residual data set is determined.

6. The system according to claim 1, characterized in that, The execution terminal is also used for: In response to receiving the process compensation instruction, the glue application parameters or speed control parameters of the corrugated cardboard production line are adjusted to perform closed-loop process compensation.

7. The system according to claim 1, characterized in that, The data acquisition terminal includes: a tension sensor, a temperature sensor, a current transformer, and a cardboard flatness detection module; The tension sensor is used to collect the tension value of each section of paper. The temperature sensor is used to collect the surface temperature of the preheating cylinder; The current transformer is used to collect the current waveform of the main motor; The cardboard flatness detection module is used to collect flatness readings at the cardboard exit.

8. The system according to claim 1, characterized in that, The system also includes an expert terminal; The expert terminal is used to send new fault mechanism parameters to the control server; The control server is also used to receive the new fault mechanism parameters and update the fault mechanism library.

9. The system according to claim 1, characterized in that, The execution terminal is also used for: The equipment fault types corresponding to the shutdown maintenance instructions are displayed visually. The parameter adjustment process corresponding to the process compensation command is visualized.

Images