A paving machine self-optimization control method and system based on multi-source scoring feedback

Abstract

This invention discloses a self-optimizing control method and system for a fabric spreading machine based on multi-source scoring feedback, belonging to the field of textile machinery automation technology. The method includes: real-time acquisition of multi-source data from the fabric spreading machine; calculation of parameter jitter score Q1, operational continuity score Q2, and fabric smoothness score Q3 respectively, and weighted fusion to obtain a comprehensive quality score; when the comprehensive quality score is below a threshold, root cause analysis is performed to generate parameter tuning suggestions; parameter regression prediction is performed using historical high-scoring samples, optimized parameter combinations are output and executed, and the score is updated based on newly acquired data, forming a self-optimizing closed loop. This invention, through multi-dimensional quantitative evaluation and data-driven closed-loop optimization, solves problems such as reliance on experience in manual parameter tuning, subjective quality assessment, and difficulty in fault attribution, significantly improving the consistency of fabric spreading quality and the level of equipment intelligence.

Description

Technical Field

[0001] This invention relates to the field of textile machinery automation and intelligent control technology, and in particular to a self-optimization control method and system for a fabric spreading machine based on multi-source scoring feedback. Background Technology

[0002] Fabric spreading machines are core equipment in automated production lines for industries such as apparel and home textiles, and the quality of fabric spreading directly determines the precision of subsequent cutting and sewing. In actual production, the fabric spreading effect is affected by a combination of factors, including the physical properties of the fabric, environmental conditions, equipment status, and process parameter settings.

[0003] However, the existing fabric spreading machine control technology has the following main problems: (1) The fabric spreading effect relies heavily on the operator's visual inspection or is inferred indirectly through a few sensor parameters, lacking objective quantitative standards for key indicators such as flatness and alignment accuracy. (2) Although the equipment operating parameters are recorded, they are not effectively correlated with the fabric spreading quality, and the automated closed loop of "evaluation-attribution-optimization-execution" cannot be realized. (3) Parameter adjustment still relies on manual trial and error, which is inefficient and difficult to adapt to dynamic changes in fabric, environment, etc. (4) When fabric spreading quality problems occur, it is difficult to quickly locate whether the problem is caused by tension fluctuations, equipment shutdown, or other factors, and the troubleshooting cycle is long. The system cannot learn autonomously from historical successful or failed cases, and the level of intelligence will not improve with the accumulation of data.

[0004] Therefore, there is an urgent need for a fabric laying machine control method and system that can integrate multi-source data perception, multi-dimensional quality assessment, intelligent root cause analysis and parameter self-optimization, so as to realize dynamic recommendation and continuous iteration of fabric laying parameters and promote the evolution of fabric laying process towards true intelligence and self-adaptation. Summary of the Invention

[0005] In view of this, the present invention provides a self-optimization control method and system for a fabric spreading machine based on multi-source scoring feedback, in order to solve the problems of subjective quality assessment, lag in parameter adjustment, difficulty in attribution, and lack of self-learning ability in the prior art.

[0006] To address the aforementioned problems, in a first aspect, the present invention provides a self-optimizing control method for a fabric spreading machine based on multi-source scoring feedback, comprising the following steps:

[0007] S1. Real-time acquisition of multi-source data from the fabric spreading machine;

[0008] S2. Calculate the parameter jitter score Q1, the operation continuity score Q2, and the fabric flatness score Q3 based on the multi-source data. Then, weight and fuse Q1, Q2, and Q3 according to preset weights to obtain a comprehensive quality score.

[0009] S3. When the overall quality score is lower than a preset threshold, root cause analysis is performed based on the low score data to generate parameter tuning suggestions;

[0010] S4. Combine historical high-scoring samples to perform parameter regression prediction to generate an optimized parameter combination;

[0011] S5. The optimized parameter combination is sent to the fabric spreading machine for execution, and the comprehensive quality score is updated based on the newly collected multi-source data after execution.

[0012] In one possible implementation, the multi-source data in step S1 includes: the multi-source data includes the tension value sequence during the fabric laying process, the number of unplanned shutdowns, and fabric images.

[0013] In one possible implementation, step S2 further includes: calculating a jitter score Q1 based on the statistical variance of the tension value sequence during the fabric laying process; calculating an operational continuity score Q2 based on the number of unplanned downtimes; and calculating a fabric smoothness score Q3 based on the fabric wrinkles after the fabric laying is completed by performing image processing on the fabric image.

[0014] In one possible implementation, the step of performing image processing on the fabric image to identify fabric wrinkles after the fabric is laid further includes:

[0015] (a) Convert an RGB image to a grayscale image;

[0016] (b) Blur denoising was performed using a 5×5 Gaussian kernel;

[0017] (c) Adopt Gaussian weighted adaptive thresholding and binarization;

[0018] (d) Use a 3×3 rectangular kernel for opening operations to reduce noise;

[0019] (e) Extract the outer contour;

[0020] (f) Filter out valid wrinkles and calculate the total area.

[0021] In one possible implementation, the formula for calculating the comprehensive quality score is Q. 综 =ω1×Q1+ω2×Q2+ω3×Q3, where ω1 is the parameter jitter score weight, ω2 is the running continuity score weight, ω3 is the fabric flatness score weight, and ω1+ω2+ω3=1.

[0022] In one possible implementation, the root cause analysis in step S3 includes:

[0023] When Q1 is lower than the first threshold and Q3 is lower than the second threshold, it is determined that the tension parameters are not compatible.

[0024] When Q2 is below the third threshold and Q1 is normal, it is determined to be a device malfunction or sensor abnormality.

[0025] When Q3 is below the fourth threshold alone, it is determined that the speed parameter or environmental conditions are abnormal.

[0026] When Q1, Q2, and Q3 are all below their corresponding thresholds, the overall operating condition is determined to be abnormal, triggering a manual intervention prompt.

[0027] In one possible implementation, step S4, which involves combining historical high-scoring samples to perform parameter regression prediction to generate an optimized parameter combination, further includes: using the current deployment task parameters and the parameter tuning suggestions as query conditions, searching for similar historical high-scoring parameter combinations in a pre-built historical high-scoring database; and based on the search results, outputting the final optimized parameter combination using weighted averaging, interpolation fitting, or artificial intelligence model prediction.

[0028] In one possible implementation, the parameter regression prediction in step S4 uses a multiple linear regression model, a random forest model, or a Bayesian optimization model; wherein the input features include fabric type, weight, elastic modulus, ambient temperature and humidity, and target number of layers, and the output features are fabric length, tension value, fabric laying speed, and target number of layers.

[0029] In one possible implementation, step S5 further includes: after the fabric spreading machine executes the optimized parameter combination, the system automatically re-collects data and calculates the comprehensive quality score. If the comprehensive quality score is still lower than the preset threshold, the root cause analysis and parameter regression prediction steps are repeated until the standard is met or the preset number of iterations is reached.

[0030] Secondly, the present invention also provides a self-optimizing control system for a fabric spreading machine based on multi-source scoring feedback, comprising:

[0031] The data acquisition module is used to collect multi-source data from the fabric spreading machine in real time;

[0032] The multi-dimensional scoring module is used to calculate the parameter jitter score Q1, the running continuity score Q2, and the fabric flatness score Q3 based on the multi-source data, and to perform weighted fusion of Q1, Q2, and Q3 according to preset weights to obtain a comprehensive quality score.

[0033] The root cause analysis module is used to perform root cause analysis based on low-score data to generate parameter tuning suggestions when the overall quality score is lower than a preset threshold.

[0034] The parameter regression prediction module is used to perform parameter regression prediction by combining historical high-scoring samples to generate optimized parameter combinations.

[0035] The execution and feedback module is used to send the optimized parameter combination to the fabric spreading machine for execution, and update the comprehensive quality score based on the newly collected multi-source data after execution.

[0036] As described above, the self-optimization control method and system for a fabric spreading machine based on multi-source scoring feedback provided by the present invention has the following beneficial effects:

[0037] (1) By collecting equipment operating parameters, shutdown events and fabric image data in real time, a three-layer evaluation system including tension stability, operation continuity and fabric flatness is constructed, and a weighted fusion model is used to output the comprehensive quality score.

[0038] (2) It does not require manual visual inspection, which significantly improves the objectivity, consistency and traceability of fabric laying quality assessment.

[0039] (3) The fabric flatness score adopts the lightweight Python OpenCV image processing flow, which does not rely on GPU or complex deep learning framework, making it easy to deploy on edge industrial control equipment and reducing implementation costs.

[0040] (4) The overall system can be seamlessly integrated with the existing fabric spreading machine control system, providing reliable data support for parameter optimization, fault warning and intelligent production scheduling. Attached Figure Description

[0041] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0042] Figure 1 This is a flowchart illustrating a self-optimization control method for a fabric spreading machine based on multi-source scoring feedback, provided in an embodiment of the present invention. Detailed Implementation

[0043] The preferred embodiments of the present invention will now be described in detail with reference to the accompanying drawings, which constitute a part of the present invention and are used together with the embodiments of the present invention to illustrate the principles of the present invention, but are not intended to limit the scope of the present invention.

[0044] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of the invention. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a mutually exclusive, independent, or alternative embodiment. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0045] A specific embodiment of the present invention, such as Figure 1 As shown, this invention discloses a method for intelligently recommending tension parameters of a fabric spreading machine based on historical data, comprising the following steps:

[0046] S1. Real-time acquisition of multi-source data from the fabric spreading machine.

[0047] In this embodiment of the invention, a data acquisition device is installed on the fabric laying machine and connected to the PLC via an RS485 bus, using the Modbus RTU protocol to read registers at 1-second cycles. Simultaneously, an industrial camera is installed on the top beam of the fabric laying machine to capture images of the fabric pile surface. The following data is collected after each layer of fabric is laid:

[0048] Tension value sequence: Record the tension value per second during the laying process of this layer, and obtain the sequence T=[t1, t2, ...,t n ];

[0049] Unplanned downtime C: Monitor the PLC alarm register to count the number of automatic downtimes caused by fabric edge misalignment, excessive tension, etc.

[0050] Fabric image: Triggers the camera to capture an image of the current layer's surface and saves it in RGB format.

[0051] S2. Calculate the parameter jitter score Q1, the operation continuity score Q2, and the fabric flatness score Q3 based on the multi-source data. Then, weight and fuse Q1, Q2, and Q3 according to preset weights to obtain a comprehensive quality score.

[0052] In this embodiment of the invention, the specific steps for calculating the jitter score Q1 are as follows:

[0053] Calculate the variance σ of the tension sequence 2 :

[0054] Q1 = max(0, 100 - k·σ) 2 );

[0055] Where k is the sensitivity coefficient.

[0056] In this embodiment of the invention, the specific steps for calculating the continuity score Q2 of the parameter operation are as follows:

[0057] Q2 = max(0, 100 - a·C);

[0058] Where 'a' represents the points deducted for a single shutdown.

[0059] In this embodiment of the invention, the specific steps for calculating the fabric smoothness score Q3 are as follows:

[0060] (a) Convert an RGB image to a grayscale image.

[0061] (b) Use a 5×5 Gaussian kernel for blurring and noise reduction to reduce interference from fabric texture, light reflection, etc., and highlight the real fold edges.

[0062] (c) Gaussian weighted adaptive thresholding is used to adapt to the differences in illumination in different regions; binarization is used to turn the inverted wrinkles into white foreground, which facilitates subsequent contour extraction.

[0063] (d) Use a 3×3 rectangular kernel for opening operation to remove noise: Use the opening operation to remove isolated noise points. If the wrinkles are fine, the opening operation can be performed again until the broken wrinkles are connected into a whole.

[0064] (e) Extract the outer contour: Find the contour of all white connected regions.

[0065] (f) Filter valid wrinkles and calculate the total area: Traverse each contour and calculate the corresponding area of ​​each contour; set a minimum effective wrinkle area threshold (e.g., 50 pixels). 2 ), to filter out minute noise; calculated as the sum of the areas W of all effective folds.

[0066] Q3 = max(0, 100 - b·W);

[0067] Where b is the deduction coefficient per unit area.

[0068] In this embodiment of the invention, the formula for calculating the comprehensive quality score is Q. 综 =ω1×Q1+ω2×Q2+ω3×Q3, where ω1 is the parameter jitter score weight, ω2 is the running continuity score weight, ω3 is the fabric flatness score weight, and ω1+ω2+ω3=1.

[0069] In this embodiment of the invention, the weights can be dynamically adjusted according to the actual production priorities. For example, during the trial production stage, the weight ω3 can be increased to prioritize cutting quality, while during large-scale stable production, the weight ω2 can be increased to improve efficiency.

[0070] S3. When the overall quality score is lower than a preset threshold, root cause analysis is performed based on the low score data to generate parameter tuning suggestions.

[0071] In this embodiment of the invention, the root cause analysis in step S3 includes:

[0072] When Q1 is lower than the first threshold and Q3 is lower than the second threshold, it is determined that the tension parameters are not compatible.

[0073] When Q2 is below the third threshold and Q1 is normal, it is determined to be a device malfunction or sensor abnormality.

[0074] When Q3 is below the fourth threshold alone, it is determined that the speed parameter or environmental conditions are abnormal.

[0075] When Q1, Q2, and Q3 are all below their corresponding thresholds, the overall operating condition is determined to be abnormal, triggering a manual intervention prompt.

[0076] S4. Combine historical high-scoring samples to perform parameter regression prediction to generate an optimized parameter combination.

[0077] In this embodiment of the invention, step S4, which involves combining historical high-scoring samples to perform parameter regression prediction to generate an optimized parameter combination, further includes: using the current deployment task parameters and the parameter tuning suggestions as query conditions, searching for similar historical high-scoring parameter combinations in a pre-built historical high-scoring database; and based on the search results, outputting the final optimized parameter combination using weighted average, interpolation fitting, or artificial intelligence model prediction.

[0078] In this embodiment of the invention, the parameter regression prediction in step S4 uses a multiple linear regression model, a random forest model, or a Bayesian optimization model; wherein, the input features include fabric type, weight, elastic modulus, ambient temperature and humidity, and target number of layers, and the output features are fabric laying length, tension value, fabric laying speed, and target number of layers.

[0079] S5. The optimized parameter combination is sent to the fabric spreading machine for execution, and the comprehensive quality score is updated based on the newly collected multi-source data after execution.

[0080] In this embodiment of the invention, step S5 further includes: after the fabric spreading machine executes the optimized parameter combination, the system automatically re-collects data and calculates the comprehensive quality score. If the comprehensive quality score is still lower than the preset threshold, the root cause analysis and parameter regression prediction steps are repeated until the standard is met or the preset number of iterations is reached.

[0081] Another specific embodiment of the present invention discloses a self-optimizing control system for a fabric spreading machine based on multi-source scoring feedback, comprising:

[0082] The data acquisition module is used to collect multi-source data from the fabric spreading machine in real time;

[0083] The multi-dimensional scoring module is used to calculate the parameter jitter score Q1, the running continuity score Q2, and the fabric flatness score Q3 based on the multi-source data, and to perform weighted fusion of Q1, Q2, and Q3 according to preset weights to obtain a comprehensive quality score.

[0084] The root cause analysis module is used to perform root cause analysis based on low-score data to generate parameter tuning suggestions when the overall quality score is lower than a preset threshold.

[0085] The parameter regression prediction module is used to perform parameter regression prediction by combining historical high-scoring samples to generate optimized parameter combinations.

[0086] The execution and feedback module is used to send the optimized parameter combination to the fabric spreading machine for execution, and update the comprehensive quality score based on the newly collected multi-source data after execution.

[0087] In summary, this invention solves the problems of reliance on experience for manual parameter adjustment, subjective quality assessment, and difficulty in attributing faults by using multi-dimensional quantitative evaluation and data-driven closed-loop optimization, and significantly improves the consistency of fabric laying quality and the level of equipment intelligence.

[0088] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention.

Claims

1. A self-optimization control method for a fabric spreading machine based on multi-source scoring feedback, characterized in that, Includes the following steps: S1. Real-time acquisition of multi-source data from the fabric spreading machine; the multi-source data includes: tension value sequence, number of unplanned shutdowns, and fabric surface images during the fabric spreading process; S2. Calculate the parameter jitter score Q1, the operational continuity score Q2, and the fabric smoothness score Q3 based on the multi-source data. Then, weight and fuse Q1, Q2, and Q3 according to preset weights to obtain a comprehensive quality score. Specifically, the parameter jitter score Q1 is calculated based on the statistical variance of the tension value sequence during the fabric laying process; the operational continuity score Q2 is calculated based on the number of unplanned downtimes; and the fabric smoothness score Q3 is calculated based on the fabric surface wrinkles after the fabric laying is completed by image processing of the fabric image. S3. When the overall quality score is lower than a preset threshold, root cause analysis is performed based on the low-score data to generate parameter tuning suggestions; wherein, the root cause analysis includes: When Q1 is lower than the first threshold and Q3 is lower than the second threshold, it is determined that the tension parameters are not compatible. When Q2 is below the third threshold and Q1 is normal, it is determined to be a device malfunction or sensor abnormality. When Q3 is below the fourth threshold alone, it is determined that the speed parameter or environmental conditions are abnormal. When Q1, Q2, and Q3 are all below their corresponding thresholds, it is determined that the overall operating condition is abnormal, triggering a manual intervention prompt. S4. Combine historical high-scoring samples to perform parameter regression prediction to generate an optimized parameter combination; S5. The optimized parameter combination is sent to the fabric spreading machine for execution, and the comprehensive quality score is updated based on the newly collected multi-source data after execution.

2. The method according to claim 1, characterized in that, The step of processing the fabric image to identify fabric wrinkles after laying further includes: (a) Convert an RGB image to a grayscale image; (b) Blur denoising was performed using a 5×5 Gaussian kernel; (c) Adopt Gaussian weighted adaptive thresholding and binarization; (d) Use a 3×3 rectangular kernel for opening operations to reduce noise; (e) Extract the outer contour; (f) Filter out valid wrinkles and calculate the total area.

3. The method according to claim 1, characterized in that, The formula for calculating the overall quality score is Q. 综 =ω1×Q1+ω2×Q2+ω3×Q3, where ω1 is the parameter jitter score weight, ω2 is the running continuity score weight, ω3 is the fabric flatness score weight, and ω1+ω2+ω3=1.

4. The method according to claim 1, characterized in that, Step S4, which involves combining historical high-scoring samples to perform parameter regression prediction to generate an optimized parameter combination, further includes: using the current deployment task parameters and the parameter tuning suggestions as query conditions, searching for similar high-scoring historical parameter combinations in a pre-built high-scoring historical database; and based on the search results, outputting the final optimized parameter combination using weighted averaging, interpolation fitting, or artificial intelligence model prediction.

5. The method according to claim 1, characterized in that, The parameter regression prediction in step S4 uses a multiple linear regression model, a random forest model, or a Bayesian optimization model; the input features include fabric type, weight, elastic modulus, ambient temperature and humidity, and target number of layers, and the output features are fabric length, tension value, fabric laying speed, and target number of layers.

6. The method according to claim 1, characterized in that, Step S5 further includes: after the fabric spreading machine executes the optimized parameter combination, the system automatically re-collects data and calculates the comprehensive quality score. If the comprehensive quality score is still lower than the preset threshold, the root cause analysis and parameter regression prediction steps are repeated until the standard is met or the preset number of iterations is reached.

7. A self-optimizing control system for a fabric spreading machine based on multi-source scoring feedback, wherein the method described in any one of claims 1-6 is characterized in that, include: The data acquisition module is used to collect multi-source data from the fabric spreading machine in real time; The multi-dimensional scoring module is used to calculate the parameter jitter score Q1, the running continuity score Q2, and the fabric flatness score Q3 based on the multi-source data, and to perform weighted fusion of Q1, Q2, and Q3 according to preset weights to obtain a comprehensive quality score. The root cause analysis module is used to perform root cause analysis based on low-score data to generate parameter tuning suggestions when the overall quality score is lower than a preset threshold. The parameter regression prediction module is used to perform parameter regression prediction by combining historical high-scoring samples to generate optimized parameter combinations. The execution and feedback module is used to send the optimized parameter combination to the fabric spreading machine for execution, and update the comprehensive quality score based on the newly collected multi-source data after execution.

Images