Textile printing and dyeing process parameter adaptive control method and system, and storage medium

Abstract

The present application relates to the technical field of textile automation control, and discloses a textile printing and dyeing process parameter adaptive control method and system and a storage medium, which comprises the following steps: outputting a torque instruction of a superimposed sine waveform to a frequency conversion servo guide roller to excite fabric warp tension fluctuation, synchronously collecting transient absorbance data of a working liquid boundary layer through a sensor array; calculating the phase delay of the absorbance data relative to the torque instruction based on orthogonal coherent demodulation operation to generate a transient spatial phase matrix; after low-pass filtering the matrix, converting the phase trend deviation into an original pressure compensation vector of a corresponding segmented bending-proof roller hydraulic cylinder by using a control stiffness conversion matrix; and combining the cloth running transmission delay and the mechanical inherent time constant to calculate a pre-trigger lead time, and sending a target pressure setting instruction in advance at a trigger node to adjust the independent hydraulic cylinder. The present application realizes online perception of fabric microscopic liquid absorption rate difference and eliminates the dynamic control dead zone caused by mechanical execution hysteresis.

Description

Technical Field

[0001] This invention relates to the field of textile automation control technology, specifically to an adaptive control method, system, and storage medium for textile printing and dyeing process parameters. Background Technology

[0002] In continuous pad-dyeing processes in textiles, fabrics are continuously impregnated through dye baths and then squeezed by padding machines to achieve a specific liquid carry-over rate. This process directly determines the color depth and color fixation uniformity of the final dyed product. However, after pretreatment, the microscopic physical properties of the fabric's porous media are often difficult to maintain absolute uniformity, frequently exhibiting subtle differences in capillary effects and porosity in the transverse or longitudinal directions. These microscopic differences cause localized variations in the fluid diffusion resistance of the working liquid as it penetrates the yarn network, resulting in fluctuations in the actual transient liquid carry-over rate of the fabric even under identical macroscopic mechanical extrusion conditions, ultimately leading to color differences on the fabric surface.

[0003] Current dyeing and printing control systems typically employ a hysteresis feedback mechanism. For example, this involves installing online colorimeters at the end of the production line to monitor color deviations in the finished product, or deploying macroscopic concentration sensors within the dye bath to maintain the overall balance of the working solution. When these end-point devices detect color or concentration shifts and trigger the control loop to adjust the rolling pressure, the fabric segment containing defects has already passed through the rolling zone due to the long physical span of the production line. Essentially, the feedback control is applying blind compensation to subsequent, unmeasured fabric.

[0004] Furthermore, the padding machines in continuous dyeing equipment are typically equipped with heavy-duty hydraulic servo systems to provide strong linear pressure. However, limited by the bulk modulus of elasticity of the hydraulic oil, pipeline damping, and the mechanical rotational inertia of the anti-bending rollers themselves, the actuators require an inherent physical pressure build-up time to reach the target output after the control system issues a pressure adjustment command. This mechanical lag at the execution level, combined with the high-speed fabric movement, renders existing systems fundamentally incapable of implementing spatiotemporally matched online compensation control for sudden changes in the microscopic physicochemical properties of the fabric. Summary of the Invention

[0005] The technical problem solved by this invention is that in existing continuous dyeing and finishing processes in textiles, closed-loop control systems typically rely on color measuring equipment or macroscopic concentration sensors at the end of the production line for feedback adjustment. When there are microscopic differences in the fabric pretreatment, the constant control of macroscopic process parameters cannot compensate for the dynamic fluctuations in the actual transient liquid absorption rate of the fabric, resulting in color differences in the finished product. Simultaneously, due to the mechanical inertia and fluid damping of the padding machine actuator, conventional feedback control systems suffer from physical transmission lag, making it impossible to implement real-time matching and pressure compensation for localized changes in the dyeability of the running fabric.

[0006] To address the above problems, the present invention provides the following technical solution: The first aspect of this invention provides an adaptive control method for textile printing and dyeing process parameters, comprising the following steps: The torque command with a sinusoidal waveform signal is output to the variable frequency servo guide roller, causing the fabric to produce dynamic tension fluctuations in the warp direction. Multiple photoelectric sampling nodes of the near-field photoelectric sensor array are controlled to synchronously acquire transient absorbance data and generate a one-dimensional transient concentration response signal vector; Based on the one-dimensional transient concentration response signal vector and the reference frequency of the sinusoidal waveform signal, a digital quadrature coherent demodulation operation is performed to calculate the phase delay value of the transient absorbance data relative to the torque command and generate a transient spatial phase matrix. A first-order inertial low-pass filter operation is performed on the transient spatial phase matrix to extract the phase trend deviation parameter, and the phase trend deviation parameter is converted into the original pressure compensation vector using the control stiffness transformation matrix. The pre-trigger lead time is calculated by combining the transmission delay time parameter of the fabric operation with the mechanical inherent time constant of the segmented anti-bending roller. When the trigger node is reached, a target pressure setting command with the original pressure compensation vector is sent to the segmented anti-bending roller to adjust the pressure output value of the independent hydraulic cylinder.

[0007] As one implementation of the first aspect, outputting a torque command superimposed with a sinusoidal waveform signal to the frequency-controlled servo guide roller includes: acquiring the basic running linear velocity and steady-state basic tension of the fabric; superimposing a sinusoidal micro-perturbation control signal into the torque control loop of the frequency-controlled servo guide roller, wherein the transient tension applied to the warp direction of the fabric is based on the steady-state basic tension as the reference center, the amplitude of the micro-perturbation tension is used as the amplitude, and the micro-perturbation angular frequency fluctuates sinusoidally with time according to a preset micro-perturbation angular frequency; and calculating the maximum allowable perturbation amount by combining the weight parameters of the processed fabric and the Young's modulus corresponding to the fiber material, and limiting the amplitude of the micro-perturbation tension within the elastic deformation threshold range of the tested fabric.

[0008] As one implementation of the first aspect, controlling multiple photoelectric sampling nodes of a near-field photoelectric sensor array to synchronously collect transient absorbance data and generate a one-dimensional transient concentration response signal vector includes: determining the spacing parameters of multiple photoelectric sampling nodes equidistantly arranged along the weft direction of the fabric based on the segment width of the independent hydraulic cylinder of the segmented anti-bending roller, and maintaining a preset mechanical gap between the optical detection end face of each photoelectric sampling node and the fabric running trajectory; setting a digital sampling frequency that satisfies the Nyquist sampling theorem based on the highest frequency component of the sine waveform signal; collecting the instantaneous absorbance values ​​of all photoelectric sampling nodes at the same discrete sampling time, and generating a one-dimensional transient concentration response signal vector containing multiple data elements according to the spatial physical arrangement order.

[0009] As one implementation of the first aspect, digital quadrature coherent demodulation is performed to calculate the phase delay value of transient absorbance data relative to the torque command, including: binding and associating the one-dimensional transient concentration response signal vector with the global timestamp of the torque command through a distributed clock synchronization mechanism; within an integration time window equal to an integer multiple of the period of the sine waveform signal, multiplying the instantaneous absorbance values ​​of discrete sampling points with standard sine reference values ​​and standard cosine reference values ​​of the same frequency respectively, summing them and obtaining the arithmetic mean, and calculating the in-phase component and quadrature component respectively; using the quadrature component as the dividend and the in-phase component as the divisor, the transient phase delay angle corresponding to each photoelectric sampling node is calculated by using the arctangent function.

[0010] As one implementation of the first aspect, a first-order inertial low-pass filter operation is performed on the transient spatial phase matrix to extract the phase trend deviation parameter, including: dividing the discrete sampling period of the underlying control by the sum of the discrete sampling period and the set filter time constant to calculate the dimensionless digital filter coefficient, and setting the filter time constant to be greater than or equal to the mechanical inherent time constant of the segmented anti-bending roller hydraulic system; within any discrete sampling period, multiplying the phase trend feature matrix cached in the previous period by the difference between the dimensionless digital filter coefficient and the product of the transient spatial phase matrix input in the current period multiplied by the dimensionless digital filter coefficient, and performing a weighted summation operation to obtain the updated phase trend feature matrix; performing a matrix subtraction operation between the updated phase trend feature matrix and the pre-stored standard fabric phase reference matrix to obtain the spatial phase trend deviation matrix as the phase trend deviation parameter.

[0011] As one implementation of the first aspect, the phase trend deviation parameter is converted into the original pressure compensation vector using a control stiffness transformation matrix. This includes: retrieving a pre-stored control stiffness transformation matrix, setting the number of rows in the control stiffness transformation matrix to correspond to the total number of independent hydraulic cylinders, and the number of columns in the control stiffness transformation matrix to correspond to the total number of photoelectric sampling nodes; extracting the spatial phase trend deviation matrix as a column vector, and performing a linear algebra matrix multiplication operation on it with the control stiffness transformation matrix. The resulting column vector is the original pressure compensation vector, thus completing the dimensionality reduction mapping from multiple sensing dimensions to the dimension of the number of hydraulic cylinders.

[0012] As one implementation of the first aspect, the pre-trigger advance time is calculated by combining the transmission delay time parameter of the fabric operation with the mechanical inherent time constant of the segmented anti-bending roller, including: obtaining the equivalent physical fabric travel distance between the near-field photoelectric sensor array and the segmented anti-bending roller line; dividing the equivalent physical fabric travel distance by the actual running line speed of the fabric to calculate the transmission delay time parameter; subtracting the mechanical inherent time constant of the actuator from the transmission delay time parameter, and defining the difference as the pre-trigger advance time; and appending an enqueue timestamp based on the current system real-time clock to a series of continuously generated original pressure compensation vectors, and sequentially pushing them into the first-in-first-out data buffer queue as independent data packets.

[0013] As one implementation of the first aspect, when the trigger node is reached, a target pressure setting command incorporating the original pressure compensation vector is sent to the segmented anti-bending roll, and the pressure output value of the independent hydraulic cylinder is adjusted. This includes: extracting the enqueue timestamp of the data packet at the head of the buffer queue, calculating the time difference between the enqueue timestamp and the current system real-time clock; when it is determined that the time difference reaches the pre-trigger advance time and does not exceed the failure threshold calculated based on the width of the anti-bending roll's rolling contact area, extracting the corresponding original pressure compensation vector from the buffer queue; multiplying each element in the extracted original pressure compensation vector by a set dimensionless control gain coefficient, subtracting the product from the corresponding basic setting value in the current process's basic line pressure vector, generating a target line pressure setting vector as the target pressure setting command, and sending it to the electro-hydraulic servo valve group of the segmented anti-bending roll to perform closed-loop control.

[0014] A second aspect of the present invention provides an adaptive control system for textile printing and dyeing process parameters, comprising: The variable frequency servo guide roller is set at the infeed end of the dye bath and is used to receive torque commands superimposed with a sine wave signal to make the fabric generate warp dynamic tension fluctuations. A near-field photoelectric sensor array is installed below the liquid surface inside the dye bath. It contains multiple independent photoelectric sampling nodes arranged along the weft direction of the fabric to synchronously collect transient absorbance data and generate a one-dimensional transient concentration response signal vector. The segmented anti-bending roller is set at the fabric exit end of the dye bath. The segmented anti-bending roller is equipped with multiple independent hydraulic cylinders distributed along the weft direction, which are used to adjust the local linear pressure distribution state according to the target pressure setting command received. The central processing unit is communicatively connected to the variable frequency servo guide roller, the near-field photoelectric sensor array, and the segmented anti-bending roller, respectively, and is used to execute the adaptive control method for textile printing and dyeing process parameters provided in the first aspect.

[0015] A third aspect of the present invention provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the adaptive control method for textile printing and dyeing process parameters provided in the first aspect.

[0016] This invention provides an adaptive control method, system, and storage medium for textile printing and dyeing process parameters. It offers the following advantages: 1. This invention excites the warp strain of the fabric by outputting a sinusoidal torque command to the variable frequency servo guide roller, and uses a near-field photoelectric sensor array to synchronously collect transient absorbance data of the working fluid boundary layer, converting the microscopic capillary effect into a measurable concentration response signal. The system further performs digital orthogonal coherent demodulation to extract the phase delay value, which can evaluate the microscopic physical differences in transient liquid absorption rate of different regions of the fabric in real time and quantitatively without changing the macroscopic fabric movement state.

[0017] 2. In this invention, after the system extracts the transient spatial phase matrix, it extracts the phase trend deviation by constructing a first-order inertial low-pass filter, thus filtering out high-frequency random noise caused by fluid turbulence. Subsequently, using a preset control stiffness transformation matrix, the high-dimensional sensing signal matrix is ​​multiplied and mapped into a low-dimensional compensation vector that conforms to the number of hydraulic cylinders. This mapping mechanism transforms the stress coupling of the mechanical continuous elastic body into linear decoupling in the mathematical domain, ensuring that the compensation actions of the independent hydraulic cylinders in each section do not interfere with each other.

[0018] 3. In this invention, the system calculates the transmission delay time by using the equivalent fabric travel distance between the sensor array and the execution roller, and subtracts it from the mechanical inherent time constant of the hydraulic system to obtain the pre-trigger advance time. By establishing an additional timestamp-stamped first-in-first-out buffer queue, the system can proactively issue the target pressure setting command in advance at the calculated pre-trigger node. The delay window of the physical fabric travel compensates for the time required for the hydraulic system to build up pressure, so that the equipment action and the spatial position of the running fabric achieve dynamic matching in the time domain and spatial domain. Attached Figure Description

[0019] Figure 1 This is a schematic diagram of the architecture of an adaptive control system for textile printing and dyeing process parameters according to an embodiment of the present invention. Figure 2 This is a flowchart illustrating an adaptive control method for textile printing and dyeing process parameters according to an embodiment of the present invention. Figure 3 This is a simulation comparison of the data smoothing effect of a first-order inertial low-pass filter according to an embodiment of the present invention; Figure 4 This is a simulation diagram comparing the dynamic response of spatiotemporal feedforward control and traditional feedback control under pressure according to an embodiment of the present invention. Detailed Implementation

[0020] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0021] See attached document Figure 1 , Figure 1 This is a schematic diagram of the architecture of an adaptive control system for textile printing and dyeing process parameters according to an embodiment of the present invention. The present invention provides an adaptive control system for textile printing and dyeing process parameters, which may include: a frequency conversion servo guide roller, a dye liquor trough, a near-field photoelectric sensor array, a segmented anti-bending roller, and a central processing unit.

[0022] The variable frequency servo guide roller is positioned at the infeed end of the dye bath trough to pull the fabric and control the warp tension of the fabric before it enters the dye bath trough.

[0023] The dye bath is used to hold the working solution required for the dyeing and printing process and to provide space for the fabric to be impregnated.

[0024] The near-field photoelectric sensor array is installed below the liquid surface inside the dye bath. The array is positioned close to the fabric's movement trajectory within the bath and is arranged in a straight line along the fabric's weft direction. The near-field photoelectric sensor array contains multiple independent photoelectric sampling nodes.

[0025] The segmented anti-bending roller is installed at the fabric exit end of the dye bath and is used to squeeze the fabric after it has been impregnated with the working solution. The segmented anti-bending roller is equipped with multiple independent hydraulic cylinders distributed along the weft direction to adjust the local linear pressure of different sections of the roller.

[0026] There is a physical distance between the installation location of the near-field photoelectric sensor array and the extrusion line of the segmented anti-bending roll.

[0027] The central processing unit (CPU) is connected via an industrial field communication bus to the driver of the variable frequency servo guide roller, the data interface of the near-field photoelectric sensor array, and the electro-hydraulic servo valve group of the segmented anti-bending roller. The CPU internally includes a perturbation injection module, a synchronous sensing and acquisition module, an orthogonal coherent demodulation module, a frequency reduction smoothing and mapping decoupling module, and a spatiotemporal prediction and feedforward execution module.

[0028] See attached document Figure 2 , Figure 2 This is a flowchart illustrating an adaptive control method for textile printing and dyeing process parameters according to an embodiment of the present invention. The adaptive control method is executed by the aforementioned central processing unit and specifically includes the following steps: S100: The central processing unit calls the perturbation injection module to obtain the basic linear velocity parameters and steady-state basic tension parameters of the current fabric operation. The perturbation injection module outputs a torque command superimposed with a sine wave signal to the frequency conversion servo guide roller, causing the fabric entering the dye bath to generate warp dynamic tension fluctuations.

[0029] S200: The fabric yarn network responds to the warp dynamic tension fluctuations by undergoing volume deformation, causing a transient concentration change in the working fluid boundary layer near the fabric surface. The central processing unit calls the synchronous sensing and acquisition module to control multiple photoelectric sampling nodes of the near-field photoelectric sensor array to synchronously acquire transient absorbance data and generate a one-dimensional transient concentration response signal vector.

[0030] S300: The central processing unit calls the quadrature coherent demodulation module to receive the one-dimensional transient concentration response signal vector. The quadrature coherent demodulation module uses the reference frequency of the aforementioned sinusoidal waveform signal as a reference, performs digital quadrature coherent demodulation operation within a set time window, calculates the phase delay value of the concentration fluctuation of each photoelectric sampling node relative to the perturbation command, and summarizes them to generate a transient spatial phase matrix.

[0031] The S400 central processing unit calls the frequency reduction smoothing and mapping decoupling module to perform a first-order inertial low-pass filter operation on the transient spatial phase matrix to extract the phase trend deviation parameter. The frequency reduction smoothing and mapping decoupling module reads the control stiffness transformation matrix pre-stored in the system and converts the phase trend deviation parameter into the original pressure compensation vector of each independent hydraulic cylinder in the corresponding segmented anti-bending roller through matrix multiplication.

[0032] S500, the central processing unit calls the spatiotemporal prediction and feedforward execution module to calculate the transmission delay time parameter required for the fabric to travel the aforementioned physical distance. Based on this transmission delay time parameter and the mechanical inherent time constant of the segmented anti-bending roller hydraulic system, the spatiotemporal prediction and feedforward execution module calculates the pre-trigger lead time.

[0033] The spatiotemporal prediction and feedforward execution module stores the aforementioned original pressure compensation vector into a data buffer queue. When the trigger node with the pre-trigger advance time setting is reached, the spatiotemporal prediction and feedforward execution module sends a target pressure setting command, incorporating the original pressure compensation vector, to the segmented anti-bending roller. The electro-hydraulic servo valve group of the segmented anti-bending roller adjusts the actual pressure output values ​​of multiple independent hydraulic cylinders according to this command, thereby changing the lateral pressure distribution state.

[0034] As a specific implementation of the aforementioned step S100, the perturbation injection module achieves dynamic modulation of the warp tension of the fabric through underlying communication commands.

[0035] The variable frequency servo guide roller specifically includes a traction guide roller, an AC servo motor coaxially connected to the traction guide roller, and a variable frequency drive for controlling the AC servo motor. The perturbation injection module establishes a data communication connection with the variable frequency drive through an industrial field communication bus, thereby enabling real-time issuance of torque control commands.

[0036] To achieve frequency perturbation injection and dynamic tension output, as a specific implementation method, the central processing unit runs the perturbation injection module to execute the following control steps: S101. In the initial stage of dynamic modulation, the perturbation injection module acquires the basic operating parameters of the fabric under the current process. The perturbation injection module reads the preset reference operating linear velocity of the system. and the steady-state basic tension that maintains the smooth operation of the fabric. Obtaining these fundamental parameters aims to establish a macroscopically stable fabric travel reference state, ensuring that subsequent superimposed dynamic perturbation signals do not disrupt the overall continuous dyeing process cycle. For the basic control process of the system obtaining actual tension feedback through a tension sensor and performing closed-loop adjustment, those skilled in the art can employ conventional proportional-integral-differential algorithms. The process of maintaining constant tension is well-known in the field and will not be elaborated upon here.

[0037] S102, the perturbation injection module under steady-state tension Based on this, a dynamic tension setting command is calculated and generated. Specifically, the perturbation injection module superimposes a low-amplitude, high-frequency sinusoidal perturbation control signal into the torque control loop of the frequency conversion servo guide roller. From the general principles of physical processes, this sinusoidal perturbation control signal causes the fabric to undergo periodic slight stretching and relaxation.

[0038] This minute mechanical disturbance, without altering the macroscopic fabric morphology, induces periodic micropore contraction and expansion within the porous medium of the yarn, providing the necessary physical excitation source for subsequent detection of the dynamic exchange process of dye liquor within the microcapillaries. After receiving this control command, the transient tension applied to the warp of the fabric specifically manifests as a steady-state baseline tension. Using the reference center and the set perturbation tension amplitude The amplitude is determined according to the preset perturbation angular frequency. A dynamic tensile force that fluctuates sinusoidally over time. The physical model of this dynamic tensile force is specifically expressed as follows: ; In the formula, This indicates the transient warp tension experienced by the fabric; Indicates steady-state fundamental tension; Indicates the amplitude of the perturbation tension; Indicates the perturbation angular frequency; Represents a time variable.

[0039] S103, the perturbation injection module performs amplitude boundary constraint calculations on the output dynamic tension setting command. To avoid irreversible plastic tensile deformation of the fabric caused by high-frequency mechanical tension fluctuations and to reduce the potential impact on the dimensional stability of the finished printed and dyed products, the perturbation injection module controls the perturbation tension amplitude. It is limited to the elastic deformation threshold range of the tested fabric.

[0040] As a preferred method, the perturbation injection module calculates the maximum allowable perturbation based on the current weight parameters of the processed fabric and the Young's modulus corresponding to the fiber material, and outputs the actual perturbation tension amplitude. Constraints are steady-state fundamental tensions Within a preset ratio range. In conventional continuous dyeing processes for cotton or blended fabrics, this preset ratio range is typically set to 1% to 3%, while the micro-perturbation angular frequency... The value is typically set between 5 rad / s and 20 rad / s. Using this range not only ensures that the mechanical actuator of the variable frequency servo guide roller has sufficient dynamic following response capability, but also provides a measurement excitation signal of sufficient intensity for the fluid boundary layer.

[0041] After receiving the constrained dynamic tension setting command, the variable frequency drive inside the variable frequency servo guide roller adjusts the speed according to the perturbation angular frequency. Adjust the output torque of the AC servo motor. Based on the above mechanism, the fabric is subjected to periodic mechanical excitation in the warp direction that varies sinusoidally with time before entering the dye bath.

[0042] Regarding the near-field photoelectric sensor array in the aforementioned step S200, the near-field photoelectric sensor array is arranged below the liquid surface of the dye bath to capture the transient optical characteristics of the microfluidic boundary layer of the fabric during the impregnation process.

[0043] To address the high-temperature working fluid environment containing suspended particles in continuous dyeing processes, a fluid-guiding protective shell is installed around each photoelectric sampling node to ensure the near-field photoelectric sensor array can acquire effective signals over a long period. As a preferred embodiment, the optical detection end face of this fluid-guiding protective shell is fitted with a corrosion-resistant sapphire transparent window. Multiple photoelectric sampling nodes of the near-field photoelectric sensor array are arranged equidistantly along the weft direction of the fabric at a predetermined spacing, forming a one-dimensional linear array. This predetermined spacing is typically determined by matching the segment width of the independent hydraulic cylinder of the subsequent actuator, i.e., the segmented anti-bending roller, to establish a one-to-one or many-to-one spatial mapping relationship.

[0044] After assembly, a preset millimeter-level mechanical gap is maintained between the optical detection end face of each photoelectric sampling node and the fabric's running trajectory. In industrial applications, this mechanical gap is typically controlled between 1mm and 3mm. Selecting this range ensures that the optical detection field of view effectively covers the working fluid boundary layer area closely adhering to the fabric surface, while preventing direct friction damage to the sensor window due to mechanical vibration of the high-speed moving fabric.

[0045] To achieve synchronous sensing of boundary layer concentration under spatial distribution conditions, the central processing unit calls the synchronous sensing and acquisition module to perform the following collaborative operation steps: S201. The synchronous sensing and acquisition module monitors the evolution of the physical state of the fabric boundary layer region. Accompanying the dynamic tension excitation experienced by the fabric in the warp direction during the aforementioned steps, the yarn network within the fabric undergoes periodic volume contraction and expansion. From the perspective of microscopic porous media fluid dynamics, when the fabric is stretched due to increased tension, the yarn gaps narrow, and some of the dye solution originally retained in the pores is squeezed out to the boundary layer; conversely, when the tension decreases and the fabric relaxes, the pores expand, and fresh dye solution in the boundary layer is drawn into the interior. This mechanical deformation forces the fluid inside the porous media of the fabric to undergo alternating material exchange with the external boundary layer fluid, thereby causing the local dye concentration in the working fluid boundary layer to exhibit transient fluctuation characteristics in sync with the warp tension.

[0046] S202, the synchronous sensing and acquisition module sends a global clock synchronization command to the near-field photoelectric sensor array. Within the near-field photoelectric sensor array... Based on this synchronization command, each independent photoelectric sampling node performs photometric measurements on the boundary layer fluid in its respective latitudinal region at the same discrete sampling time. To fully reproduce the concentration fluctuation curve caused by high-frequency tension perturbations and avoid algorithmic dead zones due to signal aliasing during data sampling, the digital sampling frequency set by the synchronous sensing and acquisition module must satisfy the Nyquist sampling theorem. In specific configurations, this sampling frequency is typically set to at least 10 times the highest frequency component of the aforementioned sinusoidal perturbation control signal. Regarding the specific circuit structure for driving the light-emitting elements within the photoelectric sampling node, receiving transmitted or reflected light signals, and converting analog electrical signals into digital signals, those skilled in the art can employ conventional spectral measurement chip-level solutions. The underlying photoelectric conversion and sample-and-hold mechanisms are well-known technologies in the field and will not be elaborated upon here.

[0047] S203, the synchronous sensing and acquisition module collects digital sampling values ​​fed back from all photoelectric sampling nodes, and structurally integrates multiple absorbance data acquired within the same sampling period to construct a one-dimensional transient concentration response signal vector representing the concentration distribution state of the fabric across its entire width at the current physical moment. Specifically, this one-dimensional transient concentration response signal vector is composed of... The data elements are arranged sequentially according to the spatial physical arrangement of the sensors, among which This represents the total number of photoelectric sampling nodes in the near-field photoelectric sensor array. Each data element in this signal vector corresponds one-to-one with the first to the second... The instantaneous absorbance values ​​actually measured at the corresponding sampling time by each photoelectric sampling node. After the vector data is constructed, it is temporarily stored in the cache of the central processing unit to serve as the data input source for subsequent orthogonal signal demodulation algorithms.

[0048] In this embodiment, the central processing unit runs the orthogonal coherent demodulation module to preprocess and align the previously acquired data, providing a computational basis for the extraction of subsequent feature signals.

[0049] In the actual industrial environment of continuous dyeing, the dye bath is filled with broadband background noise introduced by fluid turbulence, mechanical vibration, and temperature drift. The directly acquired one-dimensional transient concentration response signal vector is easily submerged in this strong noise. To effectively filter out background noise and accurately extract the phase characteristics of microcapillary concentration fluctuations, the orthogonal coherent demodulation module constructs a synchronous demodulation reference according to the following steps: S301, the quadrature coherent demodulation module performs cross-node data timing alignment. Since the variable frequency servo guide roller and the near-field photoelectric sensor array are located on different physical network nodes, their data transmission typically suffers from nonlinear network delays and communication jitter. The quadrature coherent demodulation module retrieves the distributed clock information of the underlying hardware via the industrial field communication bus. For the timestamp alignment process of the underlying data link, those skilled in the art can use the IEEE 1588 precise time protocol or industrial Ethernet distributed clock synchronization technology. Its microsecond-level clock synchronization mechanism is well-known in the field and will not be elaborated upon here. Based on this synchronization mechanism, the quadrature coherent demodulation module binds the received one-dimensional transient concentration response signal vector with the global timestamp of the excitation command issued by the perturbation injection module, thereby effectively suppressing the basis phase calculation error introduced by asynchronous sampling and network jitter.

[0050] As a preferred method for handling data anomalies, during the execution of time-series binding, if the system detects data gaps in some sampling points due to network packet loss, the orthogonal coherent demodulation module will call a linear interpolation algorithm to reconstruct and complete the missing values ​​by combining adjacent valid data frames. This compensation mechanism aims to maintain the continuity of the sampled data sequence on the time axis and avoid subsequent integration operations from falling into logical dead zones due to abrupt changes in data dimensions.

[0051] S302. After completing data alignment and verification, the quadrature coherent demodulation module acquires and generates the digital reference signal required for quadrature demodulation. From the general physical principles of signal processing, the core premise of coherent demodulation technology in extracting weak signals is to use orthogonal carriers to shift the target characteristic signal from the noise band to the DC component. This decoupling process highly depends on the absolute frequency consistency between the reference signal and the measured physical excitation signal. Based on this principle, the quadrature coherent demodulation module directly retrieves the previously set perturbation angular frequency from the memory mapping area of ​​the perturbation injection module. The parameters are used as a digital reference base frequency.

[0052] Based on the digital reference base frequency, the S303 quadrature coherent demodulation module generates two mutually orthogonal digital reference sequences within the digital signal processing unit of the central processing unit. Specifically, these two sequences include an in-phase reference sequence and a quadrature reference sequence, whose amplitudes change over time as standard sine and cosine function waveforms with the same frequency as the digital reference base frequency, respectively. Employing an internal method of source extraction and generation avoids the introduction of additional physical measurement sensors and minimizes reference frequency offset issues caused by external electrical interference or temperature drift from the underlying hardware logic. After the two orthogonal digital reference sequences are established, they are stored in the processing buffer for subsequent synchronous multiplication and addition operations.

[0053] In this embodiment, the orthogonal coherent demodulation module performs low-level mathematical operations for feature extraction within the central processing unit based on the synchronous demodulation benchmark and reference sequence generated in the aforementioned steps.

[0054] To effectively remove target feature signals from the complex fluid environment within the dye bath, the central processing unit (CPU) invokes the orthogonal coherent demodulation module to perform the following specific computational steps, in response to the complex fluid environment within the dye bath: S304. The quadrature coherent demodulation module sets the integration time window parameters according to the current operating conditions. To suppress interference from non-co-frequency noise and prevent spectral leakage from causing deviations in the calculation results, the quadrature coherent demodulation module needs to be configured with strict integration time boundaries. As a preferred parameter configuration rule, the duration of this integration time window is usually set to an integer multiple of the period of the aforementioned sinusoidal perturbation control signal. By implementing this period-aligned setting mechanism, the system can ensure that the expected value of the mathematical integral of all background noise orthogonal to the digital reference fundamental frequency approaches zero within this duration.

[0055] S305. After determining the operational boundaries, the quadrature coherent demodulation module performs parallel orthogonal integration operations on all data channels in the one-dimensional transient concentration response signal vector. For any photoelectric sampling node in the near-field photoelectric sensor array, the quadrature coherent demodulation module multiplies the acquired concentration signal with both the in-phase reference sequence and the quadrature reference sequence. Since the data object actually processed by the central processing unit is a discrete digital sampling sequence dependent on a fixed analog-to-digital conversion period, the quadrature coherent demodulation module performs discrete accumulation and summation logic on the results of the above multiplication operations in the underlying computing architecture.

[0056] The specific operation is as follows: Within the set single integration time window, the concentration observation value of each discrete sampling point is multiplied by the standard sinusoidal reference value corresponding to the current time. The arithmetic mean of the product results of all discrete sampling points is then calculated to obtain the in-phase component corresponding to the photoelectric sampling node. Similarly, the concentration observation value is multiplied by the corresponding standard cosine reference value, and the same cumulative averaging process is performed to calculate the orthogonal component corresponding to the photoelectric sampling node. This discrete transformation mechanism enables coherent algorithms to be deployed directly in industrial-grade digital signal processing chips without relying on continuous calculus models.

[0057] S306. From the perspective of the underlying signal processing physical mechanism, the aforementioned operational logic, which includes multiplication and integration accumulation steps, essentially constructs a narrow-bandpass digital filter. The working fluid in the dye bath is often filled with high-frequency random pulsations caused by fluid turbulence, and low-frequency baseline drift caused by the slow consumption of chemical concentration or gradual temperature changes. These noise signals, located in non-target frequency bands, are attenuated and canceled out after being multiplied by standard sine and cosine function carriers of the same frequency and averaged over a complete periodic sequence. This operational mechanism effectively strips away the minute concentration periodic fluctuations originally hidden in strong background noise and transforms them into a DC spatial component that does not change rapidly with time, thus laying the numerical logical foundation for the subsequent extraction of phase parameters representing microscopic fluid resistance.

[0058] In this embodiment, the central processing unit runs the orthogonal coherent demodulation module to perform mathematical calculations on the extracted DC spatial components, and then maps them into quantitative features of the fabric's microscopic properties.

[0059] To construct a data model characterizing the microscopic state of the fabric across its entire width, the central processing unit calls the orthogonal coherent demodulation module to continuously execute the following computational steps: S307. After extracting the in-phase and quadrature components, the quadrature coherent demodulation module calculates the transient phase delay corresponding to each photoelectric sampling node based on the channel data obtained in the preceding steps. For any photoelectric sampling node in the near-field photoelectric sensor array, the quadrature coherent demodulation module extracts the values ​​of the quadrature and in-phase components corresponding to that node and performs arctangent mathematical division. Specifically, the module uses the quadrature component as the dividend and the in-phase component as the divisor to calculate the transient phase delay angle of the concentration fluctuation of that node relative to the mechanical perturbation excitation using the arctangent function. The mathematical model for this calculation is expressed as follows: ; In the formula, Indicates the first The transient phase delay angle calculated from each photoelectric sampling node; Indicates the first The orthogonal components corresponding to each photoelectric sampling node; Indicates the first Each photoelectric sampling node corresponds to an in-phase component. To avoid division-by-zero overflow errors caused by the in-phase component approaching zero in the underlying digital calculations, and to ensure that the calculated phase angle accurately falls into the true mathematical quadrant, as a specific implementation method, the quadrature coherent demodulation module often calls the four-quadrant arctangent function instruction containing quadrant judgment logic in the underlying control algorithm to execute the calculation process.

[0060] This computational mechanism can effectively avoid the logical dead zone that is easily caused by conventional trigonometric function operations in industrial applications.

[0061] S308. Based on the single-point phase calculation results, the system further transforms the discrete data into a spatial feature model. After completing the phase calculation operation for all independent measurement nodes within the current discrete sampling period, the orthogonal coherent demodulation module arrays these discrete phase delay angles according to the physical arrangement order of the sensor nodes in the weft direction of the fabric. Specifically, the system extracts the first to the second... The phase delay calculation results of each node in this sampling period are arranged sequentially according to the topological order of their spatial lateral positions, thereby constructing a [database / structure]. A single-dimensional transverse data sequence of data elements, i.e., the transient spatial phase matrix. By establishing this sequence structure, the system transforms discrete single-point physical signals into a spatial sequence that can completely map the transverse state distribution of the fabric cross-section at the current measuring point, and this sequence is refreshed cyclically with the discrete sampling period of the system.

[0062] S309. Analyzing the engineering implications of this transient spatial phase matrix from the perspective of continuous dyeing process mechanism, this matrix essentially constitutes a quantitative characterization of the microscopic physicochemical properties of the fabric in the current region. When the capillary effect decreases due to uneven pretreatment processes in the micropores of the fabric's porous medium, the dynamic mass transfer resistance of dye molecules in the working fluid diffusing into the fiber interior across the boundary layer increases accordingly. This fluid viscous resistance, at the macroscopic physical level, causes the response time of the working fluid concentration fluctuation to lag behind the application time of the mechanical tension excitation. It should be noted that due to the physical inertia of the mechanical transmission process of the variable frequency guide roller and the fluid medium in the dyeing trough itself, even for an ideal fabric in a standard state, its boundary layer concentration response will have an inherent systemic basis phase delay compared to the perturbation command. Therefore, the phase delay value extracted by the orthogonal coherent demodulation module objectively superimposes this basic physical inertia. Based on this, when the fabric's microscopic fluid resistance changes, the fluctuation of this phase delay relative to the substrate is usually positively correlated with the microscopic permeation resistance of the local area. The transient spatial phase matrix, in the form of a numerical matrix, provides an objective algorithmic criterion for the subsequent implementation of latitudinal differential pressure distribution compensation in the system.

[0063] In this embodiment, the central processing unit runs a frequency reduction smoothing and mapping decoupling module to perform digital filtering on the high-frequency sampled spatial matrix to adapt to the physical response limits of the rear actuator.

[0064] After obtaining the aforementioned transient spatial phase matrix, since the refresh frequency of the electrical signal is much higher than the response capability of the mechanical system, the hydraulic servo system of the segmented anti-bending roller has intrinsic mechanical inertia and fluid damping, exhibiting typical low-pass response characteristics in the physical dimension. If the unprocessed high-frequency transient phase matrix is ​​directly sent to the hydraulic system as a control deviation, it can easily lead to frequent opening and closing actions of the electro-hydraulic servo valve, thereby causing oscillation and instability in the control system. To establish a matching relationship between the electrical high-frequency computation domain and the mechanical low-frequency execution domain, the central processing unit calls the frequency reduction smoothing and mapping decoupling module to perform the following calculation steps: The S401 frequency reduction smoothing and mapping decoupling module constructs a first-order inertial low-pass filter in the digital computing domain. For each data channel in the transient spatial phase matrix, the frequency reduction smoothing and mapping decoupling module uses a discrete recursive algorithm to smooth the high-frequency phase data. The specific operation logic is as follows: within any discrete sampling period, the module extracts the old phase trend feature data cached in the previous period and the newly acquired high-frequency transient phase data in the current period. Subsequently, using a set dimensionless digital filter coefficient as a weighting factor, a weighted summation operation is performed on the old and new data. The weight of the newly acquired data is the digital filter coefficient, while the weight of the old cached data is one minus the digital filter coefficient. The result of the weighted summation is the updated phase trend feature matrix for the current period. The discrete mathematical model of this first-order inertial digital filtering process is expressed as: ; In the formula, Indicates the current number The phase trend feature matrix obtained by calculating a discrete sampling period; This represents the phase trend feature matrix obtained from the calculation in the previous discrete sampling period; This represents the high-frequency transient spatial phase matrix of the input during the current discrete sampling period; This represents the dimensionless digital filter coefficients. To ensure the logical integrity of the algorithm during system power-on startup, the frequency reduction smoothing and mapping decoupling module initializes the cache size of the previous cycle to a zero matrix or assigns it to the pre-calibrated reference state parameters of the system during the first operation to avoid null pointer or undefined variable errors in the program.

[0065] S402, the frequency reduction smoothing and mapping decoupling module calibrates the aforementioned digital filter coefficients based on the mechanical bandwidth constraint. In control system engineering practice, the value of the digital filter coefficients determines the filter's cutoff frequency and response hysteresis. As a specific implementation method, the frequency reduction smoothing and mapping decoupling module calculates these coefficients by combining the discrete sampling period of the underlying control with the set filter time constant. The specific calculation rule is as follows: divide the value of the discrete sampling period by the sum of the sampling period and the filter time constant; the quotient obtained is the digital filter coefficient. The specific calculation rule model is expressed as follows: ; In the formula, Represents the dimensionless digital filter coefficients; Indicates the discrete sampling period; This represents the filter time constant. To ensure that the smoothed command can be effectively followed by the actuator, the lower limit of this filter time constant is subject to the inherent mechanical time constant of the segmented anti-bending roller hydraulic system. Technicians often obtain the inherent physical time required for the hydraulic cylinder to build up pressure to 63.2% of the steady-state value through step response testing, and use this as a boundary reference for the mechanical bandwidth, thereby determining whether the filter time constant is greater than or equal to this inherent mechanical time constant.

[0066] S403, the frequency reduction smoothing and mapping decoupling module, cyclically executes the above recursive filtering operation, continuously outputting the smoothed phase trend feature matrix. Through this filtering mechanism, high-frequency phase abrupt changes caused by random turbulence of the fluid in the dyeing tank or minute mechanical vibrations of the fabric are effectively attenuated, while the gradual trend quantities that truly characterize the long-range macroscopic evolution of the fabric's microscopic physicochemical properties are fully preserved. The extracted phase trend feature matrix eliminates high-frequency noise interference, providing a stable feedforward data source for the system's subsequent execution of spatial dimension stiffness mapping and differential pressure calculation.

[0067] In this embodiment, the central processing unit runs a frequency reduction smoothing and mapping decoupling module to perform a benchmark comparison on the smoothed phase data and executes spatial dimensionality reduction and decoupling mapping for the physical characteristics of the mechanical hardware.

[0068] After obtaining the aforementioned stable phase trend feature matrix, the system needs to convert this purely mathematical observation feature into the corresponding control bias of the physical actuator. The central processing unit calls the frequency reduction smoothing and mapping decoupling module to perform the following calculation steps: S404, the frequency reduction smoothing and mapping decoupling module extracts the spatial phase trend deviation parameter. The system internally stores a standard fabric phase reference matrix. This reference matrix is ​​generated by extracting a standard pre-treated fabric sample that meets the process requirements from the current batch of fabric and obtaining a steady-state phase data sequence under calibrated operating conditions. The frequency reduction smoothing and mapping decoupling module uses the phase trend feature matrix calculated in the current cycle. Phase reference matrix of standard fabric Perform matrix subtraction to obtain the spatial phase trend deviation matrix. Each element in this deviation matrix quantifies the degree of physical bias in which the current fabric deviates from the ideal dyeability state in different transverse regions due to microscopic differences in capillary effect.

[0069] The S405 frequency reduction smoothing and mapping decoupling module performs spatial dimension analysis of the hardware topology architecture. In actual industrial printing and dyeing equipment topologies, the number of photoelectric sampling nodes contained in the near-field photoelectric sensor array... The number of independent hydraulic cylinders configured inside the segmented anti-bending roller They are usually not equal; in engineering, this often manifests as an asymmetric structure of high-density sensing and low-density execution. The outer surface of a segmented anti-bending roller is typically covered with a continuous elastic rubber layer. When the output pressure of a local hydraulic cylinder is adjusted, based on the physical law of stress diffusion in a continuous elastic body, the linear pressure on the roller surface of adjacent sections will also undergo deformation response. This mechanical structural feature constitutes a typical multivariable strongly coupled system.

[0070] S406. To address the aforementioned issues of spatial dimension asymmetry and mechanical stress coupling, the frequency reduction smoothing and mapping decoupling module performs a stiffness matrix mapping transformation operation. This module retrieves the system's pre-stored control stiffness transformation matrix and uses matrix multiplication to reduce the dimension of the deviation matrix, which contains multiple sensing dimensions, into a compensation vector that conforms to the dimension of the number of hydraulic cylinders. The core mathematical model of this mapping transformation is expressed as follows: ; In the formula, Represents the original pressure compensation vector; This represents the control stiffness transformation matrix; This represents the spatial phase trend deviation matrix. Specifically, the number of rows in the control stiffness transformation matrix corresponds to the total number of independent hydraulic cylinders, and the number of columns corresponds to the total number of photoelectric sampling nodes. The system extracts the previously obtained spatial phase trend deviation matrix as a column vector and performs a standard linear algebra matrix multiplication operation on it with the control stiffness transformation matrix. The resulting column vector after the multiplication is the original pressure compensation vector, where each data element indicates, one-to-one, the required pressure adjustment bias for each corresponding independent hydraulic cylinder in the segmented anti-bending roller.

[0071] S407. The acquisition of the internal elements of the control stiffness transformation matrix relies on prior offline physical calibration. As a specific engineering implementation, technicians established a finite element analysis model by combining the mechanical elastic modulus of the segmented anti-bending roller, the Shore hardness of the roller surface rubber coating, and the fluid constitutive relationship of the current dyeing and printing working fluid. By applying a step pressure excitation to each independent hydraulic cylinder and recording the deformation response curve generated by the system in the transverse direction, technicians calculated the cross-linking influence coefficient between the hydraulic control quantity and the local linear pressure distribution, and then constructed the matrix structure. The control polarity configuration logic of this matrix operation is explained by combining the general principles of pad-dyeing control: when the phase delay of a certain transverse region increases relative to the reference state, it indicates that the microscopic capillary effect of the fabric at that location deteriorates and the rate of dye liquor diffusion into the fiber interior decreases. Under this operating condition, the matrix mapping operation tends to output a positive bias parameter as a compensation command. This positive bias parameter, after being input into the subtraction operation model of the subsequent feedforward execution module, will achieve the goal of reducing the actual pressure of the local hydraulic cylinder in that area. This increases the microscopic liquid carrying capacity of the fabric at that location by reducing the mechanical extrusion pressure, and vice versa. This matrix multiplication mapping operation based on physical stiffness calibration pre-processes the stress interference effect of adjacent areas caused by the continuous elastic body of the roll within the mathematical calculation domain, so that the calculated original pressure compensation vector can drive the hydraulic cylinders in each section to complete the corresponding pressure compensation action relatively independently.

[0072] In this embodiment, the central processing unit's runtime space prediction and feedforward execution module aims to solve the problem of large control lag caused by the inertia of heavy machinery by converting the physical walking distance into a time buffer for the control system.

[0073] In the layout of continuous dyeing equipment, there is an objectively fixed physical fabric transport length between the measurement profile where the near-field photoelectric sensor array is located and the execution line where the segmented anti-bending roller is located. To reasonably assess the time point at which the fabric being measured travels from the sensing point to the execution point, and to allow for lead time for the lagging hydraulic system, the central processing unit calls the spatiotemporal prediction and feedforward execution module to perform the following parameter calibration and calculation steps: S501, the spatiotemporal prediction and feedforward execution module calibrates and calculates the dynamic fabric transmission delay time. This module retrieves the physical arrangement parameters stored internally by the system to obtain the equivalent physical fabric transmission distance between the near-field photoelectric sensor array and the segmented anti-bending roller line. Simultaneously, the module reads the actual linear velocity of the fabric under the current processing technology in real time. Based on fundamental kinematic principles, the module divides this equivalent physical transmission distance by the actual linear velocity to calculate the transmission delay time required for the tested fabric segment, which exhibits microscopic dyeability fluctuations, to reach the execution roller line. This time parameter represents the buffer time the system has from sensing local fabric characteristics until that area enters the physical compression zone.

[0074] The S502 spatiotemporal prediction and feedforward execution module obtains the mechanical inherent time constant of the actuator. Due to limitations imposed by the fluid's bulk modulus, the servo valve dead zone, and the inherent mechanical rotational inertia of the anti-bending roll, the hydraulic system of the segmented anti-bending roll cannot respond to electronic control commands in real time. The spatiotemporal prediction and feedforward execution module retrieves the mechanical inherent time constant obtained from the system's factory calibration or engineering measurements. This constant characterizes the physical time elapsed from the moment the system issues a step command to the electro-hydraulic servo valve assembly until the actual linear pressure output of the independent hydraulic cylinder reaches the set target value. As a specific engineering boundary reference, for common heavy-duty uniform rolling mills, this mechanical inherent time constant typically ranges from 0.5 seconds to 2.0 seconds.

[0075] S503, the spatiotemporal prediction and feedforward execution module performs a comparative analysis of inverse kinematic time constraints. The system compares the numerical relationship between the transmission delay time and the inherent mechanical time constant. To ensure that the heavy-duty actuator has sufficient time to build up pressure in advance, enabling the compensation action to match the spatial position of the fabric, the system's process settings often require that the transmission delay time of the physical fabric movement be greater than the inherent physical time consumption of the actuator. If the system detects that the transmission delay time is less than or equal to the inherent mechanical time constant due to an excessively high fabric running speed, the spatiotemporal prediction and feedforward execution module will issue a timing insufficiency warning to the main control system, prompting the operator to adjust the process speed or accept a degraded dynamic tracking accuracy. Under conventional process configurations, this time constraint usually meets the physical requirements for the system to implement spatiotemporal decoupling, thereby generating a pre-trigger period on the timeline for the system to act in advance.

[0076] In this embodiment, the central processing unit runs the space-time prediction and feedforward execution module to perform time-series scheduling on the previously calculated pressure compensation data in order to achieve dynamic convergence of fabric movement and hydraulic action.

[0077] After confirming that the system meets the physical boundary conditions for spatiotemporal decoupling, the system needs to queue and trigger a series of instructions generated by discrete sampling periods in an orderly manner. The central processing unit calls the spatiotemporal prediction and feedforward execution module to perform the following queue scheduling and pre-trigger calculation steps: S504. The spatiotemporal prediction and feedforward execution module calculates the pre-trigger lead time. Based on the time parameters obtained in the preceding steps, the module performs a reverse deduction in the time domain. The specific calculation rule is as follows: subtract the mechanical inherent time constant of the actuator from the transmission delay time of the dynamic fabric movement; the difference is defined as the pre-trigger lead time. Specifically, this pre-trigger lead time, in a physical sense, represents the lag period that the system must maintain while waiting for the fabric segment containing micro-fluctuations to move backward after it has been detected by the sensor array.

[0078] The S505 spatiotemporal prediction and feedforward execution module allocates and maintains a first-in-first-out (FIFO) data buffer queue in the system memory. For a series of raw pressure compensation vectors generated during continuous production, the system sequentially pushes them into this buffer queue as independent discrete data packets. To establish a mapping relationship between spatial data and the time axis, as a preferred method, the spatiotemporal prediction and feedforward execution module appends an enqueue timestamp based on the current system real-time clock to each data packet pushed into the queue. This buffer queue serves as a medium for data temporary storage and timing delay scheduling, achieving a smooth transition between high refresh rates in electrical computation and mechanically delayed execution.

[0079] S506, the spatiotemporal prediction and feedforward execution module executes queue shifting and trigger inspection logic based on the system clock. Within each control scan cycle, the module extracts the enqueue timestamp of the data packet at the head of the buffer queue and calculates the time difference between this timestamp and the current system real-time clock. When the spatiotemporal prediction and feedforward execution module determines that this time difference reaches the pre-trigger lead time calculated in the aforementioned steps, it triggers the instruction extraction operation. Simultaneously, to avoid the algorithm getting stuck in a logical dead zone and to ensure the effectiveness of the compensation action, the module sets an effective time window boundary in the extraction logic.

[0080] If the system detects that the time difference not only reaches the pre-trigger advance time but also exceeds the failure threshold calculated by the system based on the width of the anti-bending roller contact area, it indicates that the fabric segment corresponding to the data packet has crossed the physical execution line. Faced with this abnormal condition, the module will directly determine that the data packet is invalid and discard it to prevent the subsequent compensation instruction sequence from accumulating misaligned execution. For data packets within the valid time window, the module extracts them from the buffer queue as valid control instructions activated at the current physical moment and releases the memory address they occupy, causing the remaining data packets in the queue to shift forward. By implementing this time judgment mechanism based on timestamp comparison, the system can reduce the risk of timing misalignment caused by transient fabric slippage or speed disturbances, ensuring that the issuance time of compensation instructions more accurately corresponds to the set pre-execution node.

[0081] In this embodiment, the central processing unit's runtime spatial prediction and feedforward execution module converts the aforementioned timing scheduling data into physical control instructions for the underlying execution mechanism, thereby completing the spatial differential adjustment of the dyeing process parameters.

[0082] After retrieving the pressure compensation data corresponding to the current trigger node from the buffer queue, the system enters the final stage of converting front-end perceived data into back-end physical execution data. The central processing unit calls the spatiotemporal prediction and feedforward execution module to execute the following control steps: S507, the spatiotemporal prediction and feedforward execution module calculates and generates the target line pressure setting vector for the underlying drive. The module reads the base line pressure vector used to maintain the fabric's macroscopic roll-off rate in the current dyeing process. Combining the dynamic bias extracted from the buffer queue, the spatiotemporal prediction and feedforward execution module introduces a preset dimensionless control gain coefficient for numerical conversion. The specific calculation logic is as follows: the module multiplies each element of the original pressure compensation vector periodically popped from the buffer queue by this dimensionless control gain coefficient, and then subtracts the product from the corresponding base setting value in the base line pressure vector to calculate the target line pressure setting vector at the current trigger moment. The mathematical model of this calculation logic is expressed as follows: ; In the formula, Indicates the target line pressure setting vector; Represents the baseline pressure vector; This represents the dimensionless control gain coefficient. This represents the original pressure compensation vector. The target linear pressure setting vector contains the control values ​​for each independent hydraulic cylinder. As a specific implementation method, the control gain coefficient is pre-calibrated based on current process requirements and the mechanical gain characteristics of the hydraulic cylinders, used to adjust the control system's sensitivity to changes in micro-resistance. Based on this calculation model, the system can superimpose local compensation adjustments while maintaining overall fabric stress stability.

[0083] The S508, spatiotemporal prediction, and feedforward execution module executes control commands via the fieldbus for hydraulic drive. After the target line pressure setting vector is generated, the central processing unit synchronously sends this multi-dimensional vector command to the electro-hydraulic servo valve group of the segmented anti-bending roller via the industrial field communication bus. Upon receiving the command, the electro-hydraulic servo valve group allocates the corresponding hydraulic oil flow and pressure according to the target values ​​set in each data channel, driving multiple independent hydraulic cylinders inside the segmented anti-bending roller to adjust their movements. For the underlying closed-loop control process of the electro-hydraulic servo valve group responding to digital commands and driving the hydraulic cylinders to establish pressure, those skilled in the art can use a conventional proportional-integral-derivative pressure feedback adjustment loop combined with an internal pressure sensor. Its inner-loop servo drive mechanism is a well-known technology in the field and will not be described in detail here.

[0084] S509. The system achieves dynamic convergence of physical states based on a feedforward command issuance mechanism. Combining the aforementioned spatiotemporal decoupling logic with process analysis, the physical moment when the spatiotemporal prediction and feedforward execution modules issue commands to the electro-hydraulic servo valve group is ahead of the moment the defective fabric reaches the execution rolling line on the time axis. After receiving the command, the segmented anti-bending roller's mechanical system begins the physical pressure-building process. Because the amount of time the command is issued ahead of time is set according to the system's inherent mechanical time constant, the moment when the hydraulic cylinder overcomes fluid damping and mechanical inertia to reach the target line pressure setting is essentially aligned with the moment when the measured fabric segment, containing micro-fluidities, crosses the physical fabric travel distance to reach the extrusion area. Through this control architecture mechanism, the system utilizes the fabric travel time buffer to offset the execution delay of heavy machinery, allowing the differentiated pressure distribution of the segmented anti-bending roller in the transverse weft direction to maintain the set state within the effective tolerance time range of the fabric defect contacting the roller, thereby achieving a feedforward adaptive closed loop for process compensation.

[0085] In this embodiment, the aforementioned central processing unit and related functional modules rely on specific industrial-grade computer equipment hardware entities to perform low-level calculations and network communications.

[0086] As a specific system hardware implementation, this invention provides a computer device for carrying out and executing the aforementioned adaptive control method for textile printing and dyeing process parameters. This computer device may include a processor component, a memory component, and an industrial communication interface unit.

[0087] This processor component, as the core processing unit of the system, is used to parse and execute various software instructions stored in the memory component. In actual industrial control deployments, this processor component can be composed of one or more integrated circuit chips. Those skilled in the art can implement this computing function using, but not limited to, general-purpose central processing units (CPUs), digital signal processors (DSPs), field-programmable gate arrays (FPGAs), or application-specific integrated circuits (ASICs). Considering that this invention relates to a large number of parallel digital orthogonal coherent demodulation operations and high-frequency matrix decoupling mappings, as a preferred approach, this processor component often adopts an architecture of an ARM-based industrial multi-core processor combined with a low-level digital signal processing coprocessor to ensure that the underlying control algorithm has a microsecond-level response capability.

[0088] The memory component establishes an internal data bus connection with the processor component. The memory component stores the underlying operating system program, device driver code, and application program instructions carrying the control algorithm required to execute the aforementioned control method. Simultaneously, the memory component is also divided into a data buffer with predefined contiguous addresses for temporarily storing the acquired one-dimensional transient concentration response signal vector, the extracted transient spatial phase matrix, and pressure compensation data packets in a first-in-first-out buffer queue. In terms of hardware form, this memory component can encompass a combination of volatile and non-volatile storage media, such as static random access memory (SRAM), dynamic random access memory (DRAM), and various read-only flash memory chips or industrial-grade solid-state drives.

[0089] The industrial communication interface unit is connected to the processor component, serving as the physical hardware port for data interaction between the computer device and external physical actuators. To ensure reliable data transmission in the harsh electromagnetic environment of industrial sites, the industrial communication interface unit integrates hardware such as an industrial Ethernet controller or a fieldbus protocol conversion chip. To support the system's requirement for high-precision timing alignment of cross-node data in the preceding algorithm, the industrial communication interface unit is equipped with a distributed clock hardware synchronization circuit supporting precise time protocols (such as the IEEE 1588 protocol). Through this communication interface unit, the processor component can send torque perturbation commands to the variable frequency servo guide roller, receive high-frequency digital signals from the near-field photoelectric sensor array, and issue target line pressure setting vectors to the electro-hydraulic servo valve group of the segmented anti-bending roller. For the network handshake protocol and data packet verification process of the underlying communication interface, those skilled in the art can use well-known industrial communication standards such as PROFINET, EtherCAT, or EtherNet / IP, which will not be elaborated upon here.

[0090] When the computer device is powered on, the processor component retrieves relevant instructions and basic parameters from the memory unit, and then executes and implements the aforementioned perturbation injection module, synchronous sensing and acquisition module, orthogonal coherent demodulation module, frequency reduction smoothing and mapping decoupling module, and spatiotemporal prediction and feedforward execution module in its internal logic unit. These functional modules, after execution, collaboratively perform a series of control steps in the textile printing and dyeing process, such as spatial perception, feature decoupling, and spatiotemporal prediction, to achieve stable and continuous operation of the adaptive control system.

[0091] In this embodiment, to facilitate the replication, deployment and system upgrade of the technical solution, the aforementioned spatiotemporal decoupling-based dyeing control algorithm is encapsulated and solidified within a storage medium of a specific form.

[0092] To facilitate the replication, deployment, and system upgrade of the technical solution, the aforementioned spatiotemporal decoupling-based dyeing control algorithm can be encapsulated and embedded within a specific type of storage medium. As a supplement to the software implementation, this invention provides a computer-readable storage medium. This storage medium stores a computer program or executable instructions. Combined with the conventional operating logic of an industrial automation system, when the computer program or executable instructions are read and executed by the processor component in the aforementioned computer device, the computer device can be prompted to complete the aforementioned... Figure 2 The adaptive control method for textile printing and dyeing process parameters is shown (i.e., steps S100 to S500).

[0093] To ensure the logical consistency of system operation, the computer program encapsulated within the computer-readable storage medium contains multiple interconnected code segments. When the processor component executes this program, its internal operating mechanism specifically maps to the programmed implementation of the aforementioned control steps: During the programmatic perturbation injection phase, the processor component retrieves the underlying interface driver code and clock configuration instructions from the storage medium to initialize the runtime environment of the perturbation injection module. These instructions prompt the processor component to acquire basic operating parameters, generate a dynamic tension setting instruction containing sinusoidal perturbation characteristics, and send this instruction to the physical execution network segment via the underlying network stack, thereby achieving the meridional dynamic tension fluctuation control in step S100.

[0094] During the programmatic sensing acquisition phase, the processor component calls the data synchronization allocation code to instantiate the synchronous sensing acquisition module in memory. This instruction causes the processor component to send a global clock synchronization instruction to the external sensor network and allocate specific contiguous addresses in the memory cache to receive and structure and integrate multiple concurrently transmitted digital sample values, thereby generating a one-dimensional transient concentration response signal vector, thus realizing the data acquisition and vector construction in step S200.

[0095] During the programmed coherent demodulation stage, the processor component calls digital signal processing library functions to activate the quadrature coherent demodulation module. This code instructs the processor component to generate an orthogonal waveform sequence based on the extracted reference frequency, and within a set integration time window, to perform discrete multiplication and addition calculations of the in-phase and quadrature components of the aforementioned response signal vector. Subsequently, the instruction calls the arctangent function from the math library to calculate the phase delay value of each node in space, thereby realizing the summary generation of the transient spatial phase matrix in step S300.

[0096] During the programmed down-frequency mapping stage, the processor component executes digital filtering and matrix operation code, and calls the down-frequency smoothing and mapping decoupling module. This instruction configures the processor component to run a first-order inertial low-pass filtering algorithm based on the mechanical bandwidth parameter to extract the trend deviation, and calls the basic linear algebra subroutine library to perform matrix multiplication operations, converting the spatial phase deviation of the sensing dimension into the original pressure compensation vector of the corresponding hydraulic drive dimension, thereby realizing the mapping conversion function in step S400.

[0097] During the programmed timing scheduling phase, the processor component executes queue maintenance and control feedback code, invoking the spatiotemporal prediction and feedforward execution module. This instruction prompts the processor component to calculate the transmission delay time and pre-trigger lead time based on kinematic laws, and to build and maintain a first-in-first-out (FIFO) data buffer queue in memory. With the system clock polling, this code segment guides the processor component to extract and issue timing trigger instructions based on timestamp comparison logic, thereby directing the external electro-hydraulic servo valve assembly to complete differential pressure regulation, thus realizing the feedforward closed-loop control in step S500.

[0098] Regarding the specific physical form of the computer-readable storage medium in this embodiment, those skilled in the art can employ, but are not limited to, Universal Serial Bus (USB) flash drives, portable hard drives, optical discs (such as CD-ROMs or DVD-ROMs), Secure Digital Storage Cards (SD cards), Solid State Drives (SSDs), and remote code repositories deployed in the cloud or within a local area network server. The compilation and linking process of the code and the physical read / write mechanism of the storage medium are well-known technologies in the art and will not be elaborated upon here.

[0099] Specific application examples: To further aid in understanding the technical solution and control logic of this invention, the following description is provided in conjunction with the appendix. Figure 3 With appendix Figure 4 This paper presents a specific industrial application example and a comparative analysis of its effects.

[0100] Application scenario parameter settings: This embodiment is set on a continuous pad-dyeing production line that processes all-cotton woven fabrics. The relevant physical and process parameters are calibrated based on actual industrial equipment as follows: Operating parameters: actual linear speed of the fabric Steady-state basic tension .

[0101] Perturbation parameter (S100): Perturbation tension amplitude (2% of the basic tension); micro-perturbation angular frequency .

[0102] Spatial parameter (S501): Equivalent physical fabric travel distance from the near-field photoelectric sensor array to the segmented anti-bending roller. The transmission delay time was calculated. .

[0103] Mechanical parameter (S502): The inherent mechanical time constant of the segmented anti-bending roller hydraulic system .

[0104] Digital sampling and filtering parameters (S401-S402): Discrete sampling period ; Set filter time constant The dimensionless digital filter coefficients are calculated according to the formula. .

[0105] Abnormal operating condition injection and system response (combined with formula flow): Assuming the system has been running for a certain period of time... At that time, the sensor array detected a pre-treatment defect in a certain area of ​​the fabric in the transverse direction (the capillary effect suddenly deteriorated and the microfluid resistance suddenly increased).

[0106] Feature extraction: Orthogonal coherent demodulation module based on and The transient phase delay in this region was calculated to be abrupt. Due to fluid turbulence, the unfiltered... It contains a large amount of high-frequency noise.

[0107] Frequency reduction smoothing (in conjunction with appendix) Figure 3 (Note: See attached document) Figure 3 , Figure 3 The data smoothing effect of a first-order inertial low-pass filter is demonstrated. The thin, light gray solid line in the figure represents the high-frequency transient phase data sequence acquired by the system. This data objectively incorporates high-frequency fluid turbulence noise from the industrial environment; directly using it for control could lead to system oscillations. Therefore, the frequency reduction and smoothing module is substituted into the formula. Perform iterative calculations. For example... Figure 3 The thick black line in the figure represents the phase trend feature matrix data after processing by the formula of this invention. The curve shows that the algorithm effectively filters out high-frequency noise and accurately tracks it. The trend of sudden changes in microscopic physicochemical properties (capillary effect) that occur over time.

[0108] Stiffness mapping: through multiplication of stiffness matrices The system determines that this region requires a positive bias parameter, and generates a compensation amount to reduce the hydraulic cylinder pressure after subtraction. The target linear pressure was calculated. The pressure needs to be reduced from the initial 1000 N / cm to 900 N / cm to increase the amount of liquid carried in the defective section.

[0109] Spatiotemporal prediction triggering (S504-S506): The system calculates the pre-trigger lead time. The compensation command containing the target pressure is pushed into the FIFO buffer queue. 1.2 seconds after the defect is detected (i.e., 11.2 seconds of physical clock), the system actively pops up the command and sends it to the servo valve group.

[0110] Conclusion of effect comparison: See attached document Figure 4 , Figure 4 The study provides a detailed comparison of the stress responses of spatiotemporal feedforward control and traditional feedback control when faced with the aforementioned abnormal operating conditions.

[0111] Ideal state benchmark: such as Figure 4 As shown by the solid black line (ideal target pressure curve), it represents the section of fabric defect that reaches the anti-bending roller line. The ideal transient physical pressure change trajectory required at a given moment (from 1000 N / cm to 900 N / cm instantaneously).

[0112] The drawbacks of traditional closed-loop feedback control: Figure 4 The light gray dotted lines with squares (traditional method) represent the response curve of a traditional feedback control system without spatiotemporal decoupling. Traditional systems only initiate action when fabric defects reach the rollers (i.e., at the 12th second) or after a deviation is detected by a subsequent colorimeter. This is constrained by mechanical inertia (i.e., the inherent time constant of the machine). ), until The actual pressure only reached the target value. During this 0.8s pressure build-up period, a huge shadow area (i.e., the dynamic color difference defect generation area) was formed between the response curve and the ideal target, resulting in severe color difference defects in a 0.8m (1.0m / s×0.8s) section of fabric.

[0113] Advantages of the spatiotemporal prediction feedforward control of this invention: Figure 4 The dark gray dashed line with circles (the method of this invention) represents the actual stress response curve of the spatiotemporal prediction and feedforward execution module of this invention. Because the system in Pre-triggered extraction was performed, and the hydraulic cylinder began to build up pressure in advance. When the defective fabric segment precisely reached the extrusion line at 12.0 seconds, the heavy hydraulic cylinder completed its 0.8-second physical pressure build-up process. This ensured that the actual pressure was... Upon arrival, the fabric precisely and smoothly aligns with the ideal target pressure, completely eliminating hysteresis error. This solution utilizes physical spatial distance to fully absorb mechanical inertial hysteresis, eliminating the range of dynamically defective fabric generation from a kinematic perspective.

Claims

1. An adaptive control method for textile printing and dyeing process parameters, characterized in that, include: The torque command with a superimposed sine wave signal is output to the variable frequency servo guide roller, causing the fabric to produce dynamic tension fluctuations in the warp direction. Multiple photoelectric sampling nodes of the near-field photoelectric sensor array are controlled to synchronously acquire transient absorbance data and generate a one-dimensional transient concentration response signal vector; Based on the one-dimensional transient concentration response signal vector and the reference frequency of the sinusoidal waveform signal, a digital quadrature coherent demodulation operation is performed to calculate the phase delay value of the transient absorbance data relative to the torque command, and to generate a transient spatial phase matrix. A first-order inertial low-pass filter operation is performed on the transient spatial phase matrix to extract the phase trend deviation parameter, and the phase trend deviation parameter is converted into the original pressure compensation vector using the control stiffness transformation matrix; The pre-trigger advance time is calculated by combining the transmission delay time parameter of the fabric operation with the mechanical inherent time constant of the segmented anti-bending roller hydraulic system. When the trigger node is reached, a target pressure setting command incorporating the original pressure compensation vector is sent to the segmented anti-bending roller to adjust the actual pressure output value of the independent hydraulic cylinder.

2. The adaptive control method for textile printing and dyeing process parameters according to claim 1, characterized in that, The torque command superimposed with a sine wave signal is output to the frequency conversion servo guide roller, including: Obtain the reference running linear velocity and steady-state basic tension of the fabric; A sinusoidal perturbation control signal is superimposed on the torque control loop of the variable frequency servo guide roller. The transient tension applied to the warp of the fabric takes the steady-state basic tension as the reference center, the perturbation tension amplitude as the amplitude, and fluctuates sinusoidally with time according to the preset perturbation angular frequency. By combining the weight parameters of the processed fabric and the Young's modulus corresponding to the fiber material, the maximum allowable perturbation is calculated, and the amplitude of the micro-perturbation tension is limited within the elastic deformation threshold range of the tested fabric.

3. The adaptive control method for textile printing and dyeing process parameters according to claim 1, characterized in that, The control of the near-field photoelectric sensor array involves multiple photoelectric sampling nodes synchronously acquiring transient absorbance data to generate a one-dimensional transient concentration response signal vector, including: Based on the segment width of the independent hydraulic cylinder of the segmented anti-bending roller, the spacing parameters of multiple photoelectric sampling nodes are determined to be equidistantly arranged along the weft direction of the fabric, and the preset mechanical gap between the optical detection end face of each photoelectric sampling node and the fabric running trajectory is maintained. Based on the highest frequency component of the sinusoidal waveform signal, a digital sampling frequency that satisfies the Nyquist sampling theorem is set. The transient absorbance data of all photoelectric sampling nodes are collected at the same discrete sampling time, and the one-dimensional transient concentration response signal vector containing multiple data elements is generated according to the spatial physical arrangement order.

4. The adaptive control method for textile printing and dyeing process parameters according to claim 1, characterized in that, The step of performing digital quadrature coherent demodulation to calculate the phase delay value of the transient absorbance data relative to the torque command includes: The one-dimensional transient concentration response signal vector is bound and associated with the global timestamp of the torque command through a distributed clock synchronization mechanism; Within an integration time window equal to an integer multiple of the period of the sinusoidal waveform signal, the transient absorbance data of the discrete sampling points are multiplied by the standard sine reference value and the standard cosine reference value of the same frequency, respectively, and the arithmetic mean is obtained after accumulation. The in-phase component and the quadrature component are calculated respectively. Using the quadrature component as the dividend and the in-phase component as the divisor, the transient phase delay angle corresponding to each photoelectric sampling node is calculated by using the arctangent function.

5. The adaptive control method for textile printing and dyeing process parameters according to claim 1, characterized in that, The step of performing a first-order inertial low-pass filter operation on the transient spatial phase matrix to extract the phase trend deviation parameter includes: The dimensionless digital filter coefficient is calculated by dividing the discrete sampling period of the underlying control by the sum of the discrete sampling period and the set filter time constant. The filter time constant is set to be greater than or equal to the mechanical inherent time constant of the segmented anti-bending roller hydraulic system. Within any discrete sampling period, the phase trend feature matrix cached in the previous period is multiplied by the difference between the dimensionless digital filter coefficient and the product of the transient spatial phase matrix input in the current period and the dimensionless digital filter coefficient, and a weighted summation operation is performed to obtain the updated phase trend feature matrix. The updated phase trend feature matrix is ​​subtracted from the pre-stored standard fabric phase reference matrix to obtain the spatial phase trend deviation matrix, which is used as the phase trend deviation parameter.

6. The adaptive control method for textile printing and dyeing process parameters according to claim 5, characterized in that, The process of converting the phase trend deviation parameter into the original pressure compensation vector using the control stiffness transformation matrix includes: Retrieve a pre-stored control stiffness transformation matrix, and set the number of rows of the control stiffness transformation matrix to correspond to the total number of independent hydraulic cylinders, and the number of columns of the control stiffness transformation matrix to correspond to the total number of photoelectric sampling nodes; The spatial phase trend deviation matrix is ​​extracted as a column vector, and a linear algebraic matrix multiplication operation is performed on it with the control stiffness transformation matrix. The resulting column vector is the original pressure compensation vector, thus completing the dimensionality reduction mapping from multiple sensing dimensions to the dimension of the number of hydraulic cylinders.

7. The adaptive control method for textile printing and dyeing process parameters according to claim 1, characterized in that, The calculation of the pre-trigger advance time by combining the transmission delay time parameter of the fabric movement with the mechanical inherent time constant of the segmented anti-bending roller hydraulic system includes: The equivalent physical fabric travel distance between the near-field photoelectric sensor array and the segmented anti-bending roller line is obtained, and the equivalent physical fabric travel distance is divided by the actual running line speed of the fabric to calculate the transmission delay time parameter. The difference between the transmission delay time parameter and the mechanical inherent time constant of the segmented anti-bending roller hydraulic system is defined as the pre-trigger lead time. The continuously generated series of raw pressure compensation vectors are appended with an enqueue timestamp generated based on the current system real-time clock, and then pushed into the first-in-first-out data buffer queue as independent data packets.

8. The adaptive control method for textile printing and dyeing process parameters according to claim 7, characterized in that, Upon reaching the trigger node, sending a target pressure setting command incorporating the original pressure compensation vector to the segmented anti-bending roller, and adjusting the actual pressure output value of the independent hydraulic cylinder, includes: Extract the enqueue timestamp of the data packet at the head of the buffer queue, and calculate the time difference between the enqueue timestamp and the current system real-time clock; When it is determined that the time difference reaches the pre-trigger advance time but does not exceed the failure threshold calculated based on the width of the anti-bending roll contact area, the corresponding original pressure compensation vector is extracted from the buffer queue. Each element in the extracted original pressure compensation vector is multiplied by a set dimensionless control gain coefficient. The product is then subtracted from the corresponding basic set value in the current process's basic line pressure vector to generate a target line pressure setting vector as the target pressure setting command. This vector is then sent to the electro-hydraulic servo valve group of the segmented anti-bending roller to perform closed-loop control.

9. An adaptive control system for textile printing and dyeing process parameters, characterized in that, include: The variable frequency servo guide roller is set at the infeed end of the dye bath and is used to receive torque commands superimposed with a sine wave signal to make the fabric generate warp dynamic tension fluctuations. A near-field photoelectric sensor array is installed below the liquid surface inside the dye bath and includes multiple independent photoelectric sampling nodes arranged along the weft direction of the fabric. It is used to synchronously collect transient absorbance data and generate a one-dimensional transient concentration response signal vector. The segmented anti-bending roller is set at the fabric exit end of the dye bath. The segmented anti-bending roller is equipped with multiple independent hydraulic cylinders distributed along the weft direction, which are used to adjust the local linear pressure distribution state according to the received target pressure setting command. The central processing unit is communicatively connected to the variable frequency servo guide roller, the near-field photoelectric sensor array, and the segmented anti-bending roller, respectively, and is used to execute the adaptive control method for textile printing and dyeing process parameters as described in any one of claims 1-8.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the adaptive control method for textile printing and dyeing process parameters as described in any one of claims 1-8.

Images