Remote Monitoring and Adaptive Adjustment Method and System for Garment Production Equipment

Abstract

The present invention belongs to the technical field of control and regulation of clothing production systems, and specifically provides a method and system for remote monitoring and adaptive regulation of clothing production equipment. The method mainly includes: obtaining clothing production data monitored by remote devices; generating a time-domain feature matrix and a working condition scalar according to the clothing production data; generating equipment control parameters according to the time-domain feature matrix; correcting the equipment control parameters to obtain corrected parameters; processing the working condition scalar and the corrected parameters by using a hybrid prediction model to generate a production quality prediction value and an adaptive regulation instruction; and adjusting the control parameters of the clothing production equipment according to the adaptive regulation instruction. This application can quickly respond to changes in complex production scenarios, effectively reduce the lag of regulation instructions, ensure the precise matching of equipment parameter adjustment and real-time working conditions, thereby enhancing the stability of global control and comprehensively improving production quality and equipment operation efficiency.

Description

Technical Field

[0001] The present invention belongs to the technical field of control and regulation of clothing production systems, and in particular relates to a remote monitoring and self-adaptive regulation method and system for clothing production equipment. Background Art

[0002] Under complex and changeable working conditions, it is difficult for clothing production equipment to achieve dynamic adaptation of operating parameters due to differences in fabric properties, environmental fluctuations and other factors, resulting in unstable sewing quality, increased defective rate, accelerated wear of key equipment components due to frequent overloads, and a significant increase in maintenance costs.

[0003] The current adaptive adjustment method for clothing production equipment relies on the adjustment of a single parameter threshold and lacks the ability to collaboratively process multi-source heterogeneous data. It is difficult to accurately generate real-time adjustment strategies, resulting in low quality prediction accuracy, delayed adjustment instructions, and difficulty in adapting to complex production scenarios. Summary of the invention

[0004] The present application provides a remote monitoring and adaptive adjustment method and system for clothing production equipment, which effectively solves the problem that the current adaptive adjustment method for clothing production equipment in the prior art relies on a single parameter threshold adjustment, resulting in low quality prediction accuracy, delayed adjustment instructions, and difficulty in adapting to complex production scenarios. The application can quickly respond to changes in complex production scenarios, effectively reduce the lag of adjustment instructions, and ensure the accurate matching of equipment parameter adjustments with real-time working conditions, thereby enhancing the stability of global control and comprehensively improving production quality and equipment operation efficiency.

[0005] In order to achieve the above object, the present invention adopts the following technical solutions:

[0006] In a first aspect, the present application provides a method for remote monitoring and adaptive adjustment of clothing production equipment, comprising:

[0007] Acquire clothing production data monitored by remote equipment, wherein the clothing production data includes: drive motor current, actuator displacement, fabric tension, and ambient temperature and humidity.

[0008] Based on the clothing production data, the time domain feature matrix and working condition scalar are generated.

[0009] Generate device control parameters based on the time domain characteristic matrix.

[0010] The device control parameters are corrected to obtain corrected parameters.

[0011] A hybrid prediction model is used to process the operating condition scalar and correction parameters. The hybrid prediction model processes the time series change characteristics of the correction parameters, obtains the feature vector, and then splices it with the operating condition scalar to generate production quality prediction values ​​and adaptive adjustment instructions.

[0012] When the predicted production quality value is lower than the preset threshold, adjust the control parameters of the clothing production equipment according to the adaptive adjustment instruction.

[0013] Furthermore, generate a time-domain feature matrix, including: obtaining the working cycle of the clothing production equipment; performing data segmentation on the drive motor current and the actuator displacement based on the working cycle; respectively extracting the time-domain integral feature and the speed distribution feature from the drive motor current and the actuator displacement after data segmentation; and combining the time-domain integral feature and the speed distribution feature into a time-domain feature matrix.

[0014] Furthermore, generate a working condition scalar, including: calculating the coefficient of variation of the fabric tension according to a time window; calculating the temperature and humidity correlation index of the ambient temperature and humidity according to a time window; and combining the coefficient of variation and the temperature and humidity correlation index into a working condition scalar.

[0015] Furthermore, according to the time-domain feature matrix, generate equipment control parameters, including: performing feature fusion processing on the time-domain feature matrix to generate a joint feature tensor; using an improved GRU parameter optimization model to process the joint feature tensor and output equipment control parameters including a drive frequency compensation value and a travel reference quantity of the conveying mechanism.

[0016] Wherein, the improved GRU parameter optimization model adds a feature cross layer and a sliding window attention mechanism to the standard gated recurrent unit structure. The feature cross layer performs Hadamard product operation on the joint feature tensor to generate interaction features, and the window attention mechanism sets the window length to be inversely proportional to the equipment operation speed.

[0017] Furthermore, perform feature fusion processing on the time-domain feature matrix to generate a joint feature tensor, including: calculating the multi-scale entropy of the time-domain integral feature in the time-domain feature matrix; performing dynamic time warping alignment on the time-domain integral feature and the speed distribution feature in the time-domain feature matrix to obtain an alignment result; and fusing the multi-scale entropy and the alignment result to generate a joint feature tensor.

[0018] Furthermore, the speed distribution feature includes: speed mean, speed standard deviation, and speed peak-valley difference.

[0019] Furthermore, correct the equipment control parameters to obtain corrected parameters, including: obtaining historical equipment control parameters, performing moving average filtering on the generated equipment control parameters according to the historical equipment control parameters; and performing dynamic range constraint on the equipment control parameters after moving average filtering to generate corrected parameters.

[0020] Furthermore, the adaptive adjustment instruction includes a drive motor pulse width modulation compensation coefficient and a pneumatic valve opening adjustment step of the conveying mechanism.

[0021] Further, adjust the control parameters of the clothing production equipment according to the adaptive adjustment instruction, including: calculating a first compensation gain based on the pulse width modulation compensation coefficient of the drive motor and the tension variation coefficient, and calculating a second compensation gain based on the adjustment step of the pneumatic valve opening of the conveying mechanism and the ambient temperature and humidity correlation index; using the first compensation gain and the second compensation gain as coupled control quantities to synchronously adjust the pulse width modulation parameters of the drive motor and the adjustment step of the pneumatic valve opening of the conveying mechanism.

[0022] In a second aspect, the present application provides a remote monitoring and adaptive adjustment system for clothing production equipment, including:

[0023] A data acquisition module: acquiring clothing production data monitored by a remote device, where the clothing production data includes: drive motor current, actuator displacement, fabric tension, and ambient temperature and humidity.

[0024] A feature extraction module: generating a time-domain feature matrix and a working condition scalar according to the clothing production data.

[0025] A parameter generation module: generating equipment control parameters according to the time-domain feature matrix.

[0026] A parameter correction module: correcting the equipment control parameters to obtain corrected parameters.

[0027] A prediction and instruction generation module: processing the working condition scalar and the corrected parameters using a hybrid prediction model. The hybrid prediction model processes the time-series change characteristics of the corrected parameters, obtains a feature vector, and then splices it with the working condition scalar to generate a production quality prediction value and an adaptive adjustment instruction.

[0028] A judgment and adjustment module: when the production quality prediction value is lower than a preset threshold, adjusting the control parameters of the clothing production equipment according to the adaptive adjustment instruction.

[0029] In a third aspect, the present application provides a remote monitoring and adaptive adjustment device for clothing production equipment, which includes a memory and a processor; the memory is used to store a computer program; the processor is used to implement the steps of the remote monitoring and adaptive adjustment method for clothing production equipment as described in the first aspect when executing the computer program.

[0030] In a fourth aspect, the present application provides a storage medium, in which computer program instructions are stored. When the computer program instructions are read and run by a processor, the steps of the remote monitoring and adaptive adjustment method for clothing production equipment as described in the first aspect are executed.

[0031] Advantages of the present invention:

[0032] Through multi-dimensional data integration and dynamic feature correlation analysis, this application realizes the collaborative optimization of equipment operation parameters and environmental conditions. Based on the multi-modal feature fusion prediction model, it deeply couples the temporal parameter changes with the working conditions, and finally generates production quality prediction values and adaptive adjustment instructions, effectively solving the problems in the prior art that the current adaptive adjustment method of clothing production equipment relies on the adjustment of a single parameter threshold, resulting in low quality prediction accuracy, lagging adjustment instructions, and difficulty in adapting to complex production scenarios. It can quickly respond to the changes in complex production scenarios, effectively reduce the lag of adjustment instructions, ensure the precise matching of equipment parameter adjustment and real-time working conditions, thereby enhancing the stability of global control and comprehensively improving production quality and equipment operation efficiency.

[0033] Other features and advantages of the present invention will be described in the following specification, and some of them will become obvious from the specification or can be understood by implementing the present invention. The objectives and other advantages of the present invention can be achieved and obtained through the structures pointed out in the specification and the drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the following drawings are some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0035] Figure 1 Shows a schematic flow chart of a method for remote monitoring and adaptive adjustment of a clothing production equipment according to the present invention;

[0036] Figure 2 Shows a schematic module diagram of a system for remote monitoring and adaptive adjustment of a clothing production equipment according to the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0037] To solve the problems raised in the background art, this application realizes the collaborative optimization of equipment operation parameters and environmental conditions through multi-dimensional data integration and dynamic feature correlation analysis. Based on the multi-modal feature fusion prediction model, it deeply couples the temporal parameter changes with the working conditions, and finally generates production quality prediction values and adaptive adjustment instructions, which can quickly respond to the changes in complex production scenarios, effectively reduce the lag of adjustment instructions, ensure the precise matching of equipment parameter adjustment and real-time working conditions, thereby enhancing the stability of global control and comprehensively improving production quality and equipment operation efficiency.

[0038] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0039] In some embodiments, as Figure 1 shown, the present application provides a method for remote monitoring and adaptive adjustment of a garment production device, including:

[0040] S100. Obtain the garment production data monitored by the remote device, where the garment production data includes: drive motor current, actuator displacement, fabric tension, and environmental temperature and humidity.

[0041] S200. Generate a time-domain feature matrix and a working condition scalar according to the garment production data.

[0042] S300. Generate device control parameters according to the time-domain feature matrix.

[0043] S400. Correct the device control parameters to obtain corrected parameters.

[0044] S500. Process the working condition scalar and the corrected parameters using a hybrid prediction model. The hybrid prediction model processes the temporal variation characteristics of the corrected parameters, obtains a feature vector, and then splices it with the working condition scalar to generate a production quality prediction value and an adaptive adjustment instruction.

[0045] S600. When the production quality prediction value is lower than a preset threshold, adjust the control parameters of the garment production device according to the adaptive adjustment instruction.

[0046] In some embodiments, in S100, the drive motor current of the production device is directly measured by a Hall current sensor to reflect the load state of the drive motor.

[0047] The production device can be a garment production device for sewing, cutting, finishing, and decoration, such as an automatic button sewing machine, an automatic cutting machine, a fabric pre-shrinking machine, an automatic embroidery machine, etc.

[0048] For example, taking sewing as an example, when sewing thick fabric, the drive motor current will increase due to increased resistance. The drive motor current can be normalized.

[0049] Use a laser displacement sensor to monitor the vertical movement trajectory of the actuator of the production device, and record the temporal data of the actuator displacement in each working cycle, including at least a timestamp and a displacement amount.

[0050] The execution device can be the presser foot mechanism of an automatic button sewing machine, the cutting knife lifting mechanism of an automatic cutting machine, the roller spacing adjustment mechanism of a fabric pre-shrinking machine, the embroidery frame clamping mechanism of an automatic embroidery machine, etc.

[0051] Measure the real-time tension value during the fabric transmission process through a sensor.

[0052] Use a digital temperature and humidity sensor to collect the temperature and humidity data of the sewing workshop environment.

[0053] In some embodiments, in S200, generate a time-domain feature matrix, including:

[0054] Sa210. Obtain the working cycle of the garment production equipment.

[0055] The working cycle is the time required for the production equipment to complete a complete production action (such as forming a single stitch). The working cycle can be determined by analyzing the periodic fluctuations of the drive motor current.

[0056] Sa220. Perform data segmentation on the drive motor current and the displacement of the actuator based on the working cycle.

[0057] For example, when performing data segmentation on the drive motor current, the current peak point can be used as the starting point of segmentation, and a data segment with a duration of 10% is intercepted. Each data segment contains 100 sampling points; where T represents the working cycle. 10% duration of the data segment, each data segment contains 100 sampling points; where, T represents the working cycle.

[0058] When performing data segmentation on the displacement of the actuator, according to the periodic rise minus the fall waveform of the actuator displacement, align the current segmentation points through the dynamic time warping algorithm to ensure that the displacement data segment is synchronized with the current data segment, which can effectively avoid segmentation deviation caused by sewing speed fluctuations.

[0059] For example, the actuator displacement data segment contains a sequence of displacement values x .

[0060] Sa230. Extract the time-domain integral feature and the speed distribution feature from the drive motor current and the actuator displacement after data segmentation respectively.

[0061] Calculate the time-domain integral value for each drive motor current data segment , representing the current energy distribution: ; where, represents the current value at the t-th sampling point, represents the sampling point interval, such as , N represents the number of sampling points, which can be 100.

[0062] Sa240. Combine the time-domain integral feature and the speed distribution feature into a time-domain feature matrix.

[0063] The time-domain integral features corresponding to each working cycle and the velocity distribution features are combined by column to form a cycle feature vector, and all cycle feature vectors are arranged in chronological order to form a time-domain feature matrix.

[0064] The time-domain integral features reflect the motor load energy, and the velocity distribution features describe the motion stability of the actuator. The combination of the two can capture the implicit correlation of the sewing quality.

[0065] In some embodiments, the velocity distribution features in Sa230 include: the mean velocity, the standard deviation of velocity, and the peak-valley difference of velocity.

[0066] By performing a first-order difference calculation on the displacement of the actuator, the instantaneous velocity of the actuator can be obtained : (t = 1, 2, , N - 1); where represents the displacement value at the i-th sampling point.

[0067] Based on the sequence of the instantaneous velocity of the actuator, calculate the velocity distribution features:

[0068] The mean velocity is: .

[0069] The standard deviation of velocity is: .

[0070] The peak-valley difference of velocity is: ; where and represent the maximum and minimum values in the sequence of the instantaneous velocity of the actuator, respectively.

[0071] In some embodiments, generating the working condition scalar in S200 includes:

[0072] Sb210. Calculate the coefficient of variation for the fabric tension in a time window; calculate the temperature-humidity correlation index for the ambient temperature and humidity in a time window.

[0073] The time series data of the fabric tension can be segmented with a fixed time duration (e.g., 1 second) as the time window. For example, a 1-second window contains 10 tension data points .

[0074] Calculate the mean tension and the standard deviation within the window, and then calculate the coefficient of variation , is: , reflecting the fluctuation degree of the fabric tension.

[0075] When calculating the temperature-humidity correlation index, align it with the fabric tension window, collect the average values of the ambient temperature T and humidity H, and then calculate the temperature-humidity correlation index , ; where and represent the weight coefficients of temperature T and humidity H, which can be determined through experimental verification. For example, in the sewing of cotton fabrics, temperature has a greater impact on stitch shrinkage, so = 0.6, = 0.4.

[0076] Sb220. Combine the coefficient of variation and the temperature-humidity correlation index into a working condition scalar.

[0077] Combine the coefficient of variation and the temperature-humidity correlation index by column into a two-dimensional vector as the working condition scalar.

[0078] The coefficient of variation can quantify the tension stability, and the temperature-humidity correlation index can predict the stitch shrinkage risk by integrating environmental factors, so as to reduce the sewing speed in advance.

[0079] In some embodiments, in S300, device control parameters are generated according to the time-domain feature matrix, including:

[0080] S310. Perform feature fusion processing on the time-domain feature matrix to generate a joint feature tensor.

[0081] S320. Process the joint feature tensor using an improved GRU parameter optimization model, and output device control parameters including a drive frequency compensation value and a travel reference quantity of the transfer mechanism.

[0082] Among them, the improved GRU parameter optimization model adds a feature cross layer and a sliding window attention mechanism to the standard gated recurrent unit structure. The feature cross layer performs Hadamard product operation on the joint feature tensor to generate interaction features, and the window attention mechanism sets the window length to be inversely proportional to the device running speed.

[0083] Exemplarily, if the feature at a certain moment is , perform Hadamard product operation to generate .

[0084] The window length L is: , where v represents the current device running speed, k represents an empirical constant, such as K can be 6000, represents rounding down.

[0085] The calculation range of the attention weight is strictly limited within the window length L.

[0086] Improved GRU output device control parameters. The drive frequency compensation value is used to adjust the pulse width modulation frequency of the spindle drive motor, and the range can be 0.8 - 1.2. The travel reference amount of the transfer mechanism is used to control the reference position for adjusting the opening step of the pneumatic valve of the transfer mechanism, and the range can be 2.0 - 5.0 mm.

[0087] In some embodiments, in S310, feature fusion processing is performed on the time-domain feature matrix to generate a joint feature tensor, including:

[0088] S311. Calculate the multi-scale entropy of the time-domain integral features in the time-domain feature matrix.

[0089] The multi-scale entropy is defined to measure the complexity of the time-domain integral feature sequence and can be calculated by the sample entropy at different time scales:

[0090] Coarse-grain the time-domain integral feature sequence to generate subsequences of different scales (such as the scale factor ).

[0091] Calculate the sample entropy at each scale , and the formula is: ; where represents the logarithm of the number of templates with a matching length of +1, represents the logarithm of the number of templates with a matching length of , and r represents the similarity tolerance, usually taking 0.2 times the standard deviation.

[0092] S312. Perform dynamic time warping alignment on the time-domain integral features and the velocity distribution features in the time-domain feature matrix to obtain an alignment result.

[0093] Exemplarily, the length of the original time-domain integral feature sequence is 100, and the length of the velocity distribution feature sequence is 95. After dynamic time warping alignment, both are extended to 105 time steps, which can eliminate the timing offset caused by fluctuations between the time-domain integral features and the velocity distribution features.

[0094] S313. Fuse the multi-scale entropy and the alignment result to generate a joint feature tensor.

[0095] Concatenate the multi-scale entropy and the aligned feature matrix along the channel dimension to generate a joint feature tensor.

[0096] In some embodiments, in S400, the device control parameters are corrected to obtain corrected parameters, including:

[0097] S410. Obtain the historical device control parameters and perform a moving average filter on the generated device control parameters according to the historical device control parameters.

[0098] The historical device control parameters refer to those in the past A time window (for example = 10 cycles) to generate the drive frequency compensation value and the travel reference quantity of the transfer mechanism.

[0099] The time window length Is dynamically adjusted according to the equipment operation speed: , where Represents the average equipment operation speed over a past period of time, Represents Rounding down, the time window length m is positively correlated with the sewing speed, and can balance the response speed and the anti-noise ability.

[0100] Perform weighted averaging on the currently generated equipment control parameters, the travel reference quantity of the transfer mechanism, and the historical parameters.

[0101] The drive frequency compensation value after filtering And the travel reference quantity of the transfer mechanism Are respectively:

[0102]

[0103] Where Represents the currently generated drive frequency compensation value, Represents the currently generated travel reference quantity of the transfer mechanism, Represents The weight of, Represents the weight of the historical equipment control parameters, Represents the k-th historical drive frequency compensation value, Represents the k-th historical travel reference quantity of the transfer mechanism, Is given a higher priority, such as , Is given equal weights, such as .

[0104] S420. Perform dynamic range constraint on the equipment control parameters after moving average filtering to generate corrected parameters.

[0105] Adopt The truncation function to perform upper and lower limit truncation on the filtered drive frequency compensation value And the travel reference quantity of the transfer mechanism : ; .

[0106] Fine-tune the constraint range according to the coefficient of variation of the fabric tension : If , the upper limit of the drive frequency compensation value is relaxed to 1.25 to cope with sudden loads; if , the lower limit of the stroke reference amount of the conveying mechanism is reduced to 1.8 mm to improve the cloth feeding accuracy. Combining with the coefficient of variation of cloth tension dynamically adjusts the parameter range to adapt to the requirements of different working conditions.

[0107] In some embodiments, the hybrid prediction model in S500 at least includes: a time series feature extraction layer, a working condition scalar fusion layer, and a fully connected prediction layer.

[0108] The time series feature extraction layer is used to process the time series change characteristics of the correction parameters (drive frequency compensation value, stroke reference amount of the conveying mechanism), such as a single-layer LSTM unit.

[0109] The working condition scalar fusion layer is used to splice the working condition scalars (coefficient of variation of cloth tension , temperature and humidity correlation index ) with the time series feature vector to generate a hybrid feature vector.

[0110] The fully connected prediction layer is used to generate a production quality prediction value and an adaptive adjustment instruction, and at least includes a fully connected layer and an output layer. The activation function of the fully connected layer can be a ReLU function, and the production quality prediction value output by the output layer can adopt Sigmoid activation, mapped to 0-100, and the output of the adaptive adjustment instruction can adopt linear activation.

[0111] In some embodiments, the adaptive adjustment instruction in S500 includes a pulse width modulation compensation coefficient of the drive motor and an opening adjustment step of the pneumatic valve of the conveying mechanism.

[0112] The pulse width modulation compensation coefficient is used to dynamically adjust the duty cycle of the pulse width modulation signal of the main shaft drive motor to compensate for the speed deviation caused by load fluctuations or mechanical losses.

[0113] Exemplarily, the reference value of the pulse width modulation compensation coefficient is 1.0, and the allowable adjustable range is 0.8-1.2. When the compensation coefficient is 1.15, the pulse width modulation frequency is increased by 15%, and the motor output torque increases to cope with the resistance of thick cloth.

[0114] The opening adjustment step of the pneumatic valve represents the linear displacement increment for controlling the pneumatic valve of the conveying mechanism, and is used to adjust the cloth conveying speed and tension.

[0115] Exemplarily, the reference step of the opening adjustment step of the pneumatic valve is 0.1 mm, and the allowable adjustment range is 0.05-0.5 mm. When the step is adjusted to 0.3 mm, the opening of the pneumatic valve increases, and the cloth conveying speed is increased to avoid wrinkles.

[0116] In some embodiments, in S600, the control parameters of the clothing production equipment are adjusted according to the adaptive adjustment instruction, including:

[0117] S610. Calculate the first compensation gain based on the drive motor pulse width modulation compensation coefficient and the tension variation coefficient, and calculate the second compensation gain based on the adjustment step of the pneumatic valve opening of the conveying mechanism and the ambient temperature and humidity correlation index.

[0118] The first compensation gain is: ; where represents the drive motor pulse width modulation compensation coefficient, represents the gain coefficient, which can be determined by verifying historical data. For example, in cotton fabric sewing, the tension fluctuation has a significant impact on the motor load, represents the tension variation coefficient.

[0119] The second compensation gain is: ; where represents the adjustment step of the pneumatic valve opening, represents the gain coefficient, which is determined by verifying historical data. For example, a higher temperature and humidity environment requires more precise fabric feeding control, represents the temperature and humidity correlation index.

[0120] S620. Take the first compensation gain and the second compensation gain as coupled control quantities, and synchronously adjust the pulse width modulation parameters of the drive motor and the adjustment step of the pneumatic valve opening of the conveying mechanism.

[0121] The first compensation gain is used to adjust the pulse width modulation parameters of the drive motor, and the second compensation gain is used to adjust the opening of the pneumatic valve of the conveying mechanism. The first compensation gain and the second compensation gain need to take effect synchronously to avoid system instability caused by single parameter adjustment.

[0122] The pulse width modulation parameters of the drive motor are: ; where represents the reference frequency.

[0123] The pneumatic valve opening is: ; where represents the current pneumatic valve opening.

[0124] Dynamically adjusting the gain coefficient according to the tension variation coefficient and the temperature and humidity correlation index can reduce the action delay of the motor and the pneumatic valve.

[0125] Through the coupled control of the pulse width modulation parameters and the opening degree of the pneumatic valve, the dynamic matching between the core operations (cutting, sewing, pre-shrinking, embroidery, etc.) and the material conveying is ensured, effectively avoiding the systematic imbalance caused by single-parameter adjustment.

[0126] In some embodiments, such as Figure 2 shown, the present application provides a remote monitoring and adaptive adjustment system for a clothing production device, including:

[0127] Data acquisition module: acquiring clothing production data monitored by a remote device, where the clothing production data includes: drive motor current, actuator displacement, fabric tension, and environmental temperature and humidity.

[0128] Feature extraction module: generating a time-domain feature matrix and a working condition scalar according to the clothing production data.

[0129] Parameter generation module: generating device control parameters according to the time-domain feature matrix.

[0130] Parameter correction module: correcting the device control parameters to obtain corrected parameters.

[0131] Prediction and instruction generation module: using a hybrid prediction model to process the working condition scalar and the corrected parameters. The hybrid prediction model processes the time-series change characteristics of the corrected parameters, obtains a feature vector and then splices it with the working condition scalar to generate a production quality prediction value and an adaptive adjustment instruction.

[0132] Judgment and adjustment module: when the production quality prediction value is lower than a preset threshold, adjusting the control parameters of the clothing production device according to the adaptive adjustment instruction.

[0133] This embodiment has all the advantages of a method for remotely monitoring and adaptively adjusting a clothing production device, and can automatically execute the steps of the method for remotely monitoring and adaptively adjusting a clothing production device.

[0134] In some embodiments, the present application provides a remote monitoring and adaptive adjustment device for a clothing production device, which includes a memory and a processor; the memory is used to store a computer program; the processor is used to implement the steps of the method for remotely monitoring and adaptively adjusting a clothing production device when executing the computer program.

[0135] In some embodiments, the present application provides a storage medium, in which computer program instructions are stored. When the computer program instructions are read and run by a processor, the steps of the method for remotely monitoring and adaptively adjusting a clothing production device are executed.

[0136] Among them, any reference to a memory, storage, database, or other medium used in the embodiments provided by the present invention may include non-volatile and / or volatile memories. The non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. The volatile memory may include random access memory (RAM) or external cache memory.

[0137] It should be noted that in this text, relational terms such as "first" and "second" are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, such that a process, method, article or device comprising a series of elements not only includes those elements but also includes other elements not expressly listed, or elements inherent to such process, method, article or device. Without further limitation, an element defined by the statement "comprising one..." does not exclude the presence of additional identical elements in the process, method, article or device comprising the element.

[0138] Although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements for some of the technical features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for remote monitoring and adaptive adjustment of a clothing production device, characterized in that Including: Obtain the clothing production data monitored by the remote device, where the clothing production data includes: main spindle motor current, needle bar displacement, fabric tension, and ambient temperature and humidity; Generate a time-domain feature matrix and a working condition scalar according to the clothing production data; Generate device control parameters according to the time-domain feature matrix; Correct the device control parameters to obtain corrected parameters; Process the working condition scalar and the corrected parameters using a hybrid prediction model. The hybrid prediction model processes the temporal variation characteristics of the corrected parameters, splices the obtained feature vectors with the working condition scalar after obtaining them, and generates a sewing quality prediction value and an adaptive adjustment instruction; When the sewing quality prediction value is lower than the preset threshold, adjust the control parameters of the clothing production equipment according to the adaptive adjustment instruction; Among them, generating the time-domain feature matrix includes: obtaining the working cycle of the sewing machine; performing data segmentation on the main spindle motor current and the needle bar displacement based on the working cycle; respectively extracting the time-domain integral feature and the speed distribution feature from the main spindle motor current and the needle bar displacement after data segmentation; combining the time-domain integral feature and the speed distribution feature into a time-domain feature matrix; Generating the working condition scalar includes: calculating the coefficient of variation of the fabric tension according to a time window; calculating the temperature and humidity correlation index of the ambient temperature and humidity according to a time window; combining the coefficient of variation and the temperature and humidity correlation index into a working condition scalar; Generating device control parameters according to the time-domain feature matrix includes: performing feature fusion processing on the time-domain feature matrix to generate a joint feature tensor; processing the joint feature tensor using an improved GRU parameter optimization model, and outputting device control parameters including the motor PWM frequency compensation value and the feed dog stroke reference amount. The improved GRU parameter optimization model adds a feature cross layer and a sliding window attention mechanism to the standard gated recurrent unit structure. The feature cross layer performs Hadamard product operation on the joint feature tensor to generate interaction features, and the window attention mechanism sets the window length to be inversely proportional to the sewing speed.

2. The remote monitoring and adaptive adjustment method for clothing production equipment according to claim 1, characterized in that Performing feature fusion processing on the time-domain feature matrix to generate a joint feature tensor includes: Calculating the multi-scale entropy of the time-domain integral feature in the time-domain feature matrix; Performing dynamic time warping alignment on the time-domain integral feature and the speed distribution feature in the time-domain feature matrix to obtain an alignment result; Fusing the multi-scale entropy and the alignment result to generate a joint feature tensor.

3. The remote monitoring and adaptive adjustment method for clothing production equipment according to claim 1, wherein The speed distribution feature includes: speed mean, speed standard deviation, and speed peak-to-valley difference.

4. The remote monitoring and adaptive adjustment method for clothing production equipment according to claim 1, characterized in that Correcting the device control parameters to obtain corrected parameters includes: Obtain historical device control parameters, and perform moving average filtering on the generated device control parameters according to the historical device control parameters; Perform dynamic range constraint on the device control parameters after moving average filtering to generate corrected parameters.

5. The remote monitoring and adaptive adjustment method for clothing production equipment according to claim 4, characterized in that The adaptive adjustment instruction includes the motor driver pulse width modulation compensation coefficient and the adjustment step of the pneumatic valve opening of the feed mechanism.

6. The remote monitoring and adaptive adjustment method for clothing production equipment according to claim 5, characterized in that, Adjusting the control parameters of the clothing production equipment according to the adaptive adjustment instruction includes: Calculating a first compensation gain based on the motor driver pulse width modulation compensation coefficient and the tension coefficient of variation, and calculating a second compensation gain based on the adjustment step of the pneumatic valve opening of the feed mechanism and the temperature and humidity correlation index; Taking the first compensation gain and the second compensation gain as coupling control quantities, synchronously adjust the pulse width modulation parameters of the motor driver and the opening degree of the pneumatic valve of the fabric feeding mechanism.

7. A remote monitoring and adaptive adjustment system for clothing production equipment, characterized in that, It includes: Data acquisition module: Acquire the garment production data monitored by the remote device, and the garment production data includes: spindle motor current, needle bar displacement, fabric tension, and environmental temperature and humidity; Feature extraction module: Generate a time-domain feature matrix and a working condition scalar according to the garment production data; Parameter generation module: Generate device control parameters according to the time-domain feature matrix; Parameter correction module: Correct the device control parameters to obtain corrected parameters; Prediction and instruction generation module: Process the working condition scalar and the corrected parameters using a hybrid prediction model. The hybrid prediction model processes the time-series change characteristics of the corrected parameters, splices the obtained feature vectors with the working condition scalar after obtaining the feature vectors, and generates a sewing quality prediction value and an adaptive adjustment instruction; Judgment and adjustment module: When the sewing quality prediction value is lower than the preset threshold, adjust the control parameters of the garment production equipment according to the adaptive adjustment instruction; Among them, generating the time-domain feature matrix includes: obtaining the working cycle of the sewing machine; performing data segmentation on the spindle motor current and the needle bar displacement based on the working cycle; respectively extracting the time-domain integral feature and the speed distribution feature from the spindle motor current and the needle bar displacement after data segmentation; merging the time-domain integral feature and the speed distribution feature into a time-domain feature matrix; Generating the working condition scalar includes: calculating the coefficient of variation of the fabric tension according to a time window; calculating the temperature and humidity correlation index of the environmental temperature and humidity according to a time window; merging the coefficient of variation and the temperature and humidity correlation index into a working condition scalar; Generating device control parameters according to the time-domain feature matrix includes: performing feature fusion processing on the time-domain feature matrix to generate a joint feature tensor; processing the joint feature tensor using an improved GRU parameter optimization model, and outputting device control parameters including the motor PWM frequency compensation value and the feed dog stroke reference amount. The improved GRU parameter optimization model adds a feature cross layer and a sliding window attention mechanism to the standard gated recurrent unit structure. The feature cross layer performs Hadamard product operation on the joint feature tensor to generate interaction features, and the window attention mechanism sets the window length to be inversely proportional to the sewing speed.

Images