Material consumption analysis method based on textile technology

Abstract

This invention discloses a method for analyzing material consumption based on textile processes, relating to the field of textile technology. The method includes: reading predetermined analysis dimensions and collecting features of the textile process to obtain textile process feature parameters; comparing the target process feature parameters with first historical process feature parameters to obtain a first comparison result; when the first process similarity obtained from analyzing the first comparison result reaches a similarity limit, extracting first historical material consumption records; matching the first material quantity of a first material category among multiple material quantities and generating a first material quantity time series; analyzing the first material quantity time series to obtain a first predicted material quantity; and generating a predicted material consumption result for the textile process based on a first correspondence between the first predicted material quantity and the first material category. This method solves the technical problem of accurately predicting material consumption in textile processes in existing technologies, achieving the technical effect of improving the accuracy of material consumption prediction in textile processes.

Description

Technical Field

[0001] This invention relates to the field of textile technology, and more specifically to a method for analyzing material consumption based on textile processes. Background Technology

[0002] In the textile industry, material consumption analysis has always been a key factor affecting production efficiency, cost control, and product quality. With increasing market competition and diversified consumer demands, textile companies face the challenge of improving production efficiency, reducing production costs, and ensuring product quality. However, due to the complexity and variability of textile processes, the types, quantities, and proportions of raw materials used in the production of different textiles vary greatly, posing significant challenges to material consumption analysis. Traditional methods of material consumption analysis often rely on empirical estimations and simple data statistics, making it difficult to accurately reflect the material consumption of textile processes, let alone effectively predict future material consumption. This leads to problems in inventory management, such as inventory backlogs or shortages, which in turn affect the execution of production plans and the guarantee of delivery dates. Summary of the Invention

[0003] This application provides a method for analyzing material consumption based on textile processes, which solves the technical problem in the prior art that it is difficult to accurately predict material consumption in textile processes.

[0004] In view of the above problems, embodiments of this application provide a method for analyzing material consumption based on textile processes.

[0005] This application provides a method for analyzing material consumption based on textile processes, the method comprising:

[0006] Read predetermined analysis dimensions and collect features of the textile process based on the predetermined analysis dimensions to obtain textile process feature parameters; compare the target process feature parameters in the textile process feature parameters with the first historical process feature parameters in the first historical textile record to obtain a first comparison result, where the first historical textile record represents any historical textile record of the same type of textile of the textile process; when the first process similarity obtained from analyzing the first comparison result reaches the similarity limit, extract the first historical material consumption record in the first historical textile record, where the first historical material consumption record includes multiple material consumption amounts of multiple types of materials; match the first material consumption amount of the first material category among the multiple material consumption amounts, and generate a first material consumption time series in combination with a predetermined unit period; call the consumption integrated prediction model to analyze the first material consumption time series to obtain a first predicted material consumption amount, where the first predicted material consumption amount refers to the predicted value of the first material category consumed by the textile process at the first time; based on the first correspondence between the first predicted material consumption amount and the first material category, generate the material consumption prediction result of the textile process.

[0007] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0008] First, predetermined analysis dimensions are read, and features of the textile process are collected based on these dimensions to obtain textile process feature parameters. Next, the target process feature parameters are compared with the first historical process feature parameters in the first historical textile record to obtain a first comparison result. The first historical textile record represents any historical textile record of the same type of textile process. When the first process similarity obtained from the first comparison result reaches a similarity limit, the first historical material consumption record is extracted from the first historical textile record. This record includes multiple material consumption quantities for multiple material categories. Then, the first material quantity of the first material category is matched among the multiple material quantities, and a first material quantity time series is generated based on a predetermined unit period. The first material quantity time series is analyzed by calling a consumption integrated prediction model to obtain the first predicted material quantity, which refers to the predicted value of the textile process consuming the first material category at the first time. Finally, based on the first correspondence between the first predicted material quantity and the first material category, a material consumption prediction result for the textile process is generated. This invention solves the technical problem of accurately predicting material consumption in textile processes, thereby improving the accuracy of material consumption prediction in textile processes. Attached Figure Description

[0009] 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.

[0010] Figure 1 A schematic flowchart of a material consumption analysis method based on textile processes provided in an embodiment of this application;

[0011] Figure 2 This is a schematic diagram of the process for obtaining the first process similarity in the material consumption analysis method based on textile processes provided in the embodiments of this application. Detailed Implementation

[0012] This application provides a method for analyzing material consumption in textile processes, which solves the technical problem in the prior art that it is difficult to accurately predict material consumption in textile processes.

[0013] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0014] It should be noted that the terms "comprising" and "having" are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or server that includes a series of steps or units is not necessarily limited to those steps or units that are explicitly listed, but may include other steps or modules that are not explicitly listed or that are inherent to these processes, methods, products, or devices. Example 1

[0015] like Figure 1 As shown in the embodiments of this application, a method for analyzing material consumption based on textile processes is provided, wherein the method includes:

[0016] Read the predetermined analysis dimensions and collect features of the textile process based on the predetermined analysis dimensions to obtain textile process feature parameters.

[0017] In the analysis of material consumption in textile processes, it is necessary to read predetermined analysis dimensions, including material and process dimensions. The material dimension includes the type of raw materials, fiber composition, and yarn specifications of the textiles. The process dimension includes the textile process steps, machine type, and process parameters (such as tension, speed, and temperature). Based on the predetermined analysis dimensions, feature collection of the textile process is performed. Specifically, based on the material dimension, the types of raw materials used in the textiles, such as cotton, linen, silk, and synthetic fibers, are collected to determine the fiber composition and its proportion, as well as the yarn specifications, such as count and twist. Based on the process dimension, the textile process flow is obtained, including the types of machines used, such as looms and knitting machines, and process parameters, such as the tension and speed of looms, and the needle count and needle pitch of knitting machines, are collected to obtain the characteristic parameters of the textile process.

[0018] The target process feature parameter in the textile process feature parameter is compared with the first historical process feature parameter in the first historical textile record to obtain a first comparison result. The first historical textile record represents any historical textile record of the same type of textile of the textile process.

[0019] From the collected textile process characteristic parameters, target process characteristic parameters closely related to the target analysis are selected, including machine model, process steps, key process parameters (such as tension, speed, temperature, etc.), and material type. The first historical textile record refers to any historical textile record of the same type as the current textile process, i.e., the material and textile type are the same, such as knitting cotton and wool. From the first historical textile record, the historical process characteristic parameters corresponding to the current target process characteristic parameters are extracted; these are the first historical process characteristic parameters. The current target process characteristic parameters are compared with the extracted first historical process characteristic parameters, such as by directly comparing parameter values, calculating parameter difference rates, or performing similarity measurements, to obtain the first comparison result.

[0020] When the first process similarity obtained from the analysis of the first comparison result reaches the similarity limit, the first historical material consumption record in the first historical textile record is extracted. The first historical material consumption record includes multiple material consumption quantities for multiple types of materials.

[0021] After comparative analysis, if the first process similarity (i.e., the similarity between the target textile process and the first historical textile record) reaches the preset similarity limit, it indicates that the two have a high degree of similarity in process. Based on the comparison results, the first historical textile record similar to the current textile process is located. From the first historical textile record, the material consumption data corresponding to the first historical textile record is extracted, namely the first historical material consumption record. The first historical material consumption record contains multiple quantities of various types of materials used in the historical textile process.

[0022] Furthermore, such as Figure 2 As shown, obtaining the first process similarity includes:

[0023] Read the predetermined process labeling scheme; perform labeling processing on the target process feature parameters and the first historical process feature parameters in sequence according to the predetermined process labeling scheme to obtain the target process label and the first historical process label respectively; compare and analyze the target process label and the first historical process label to obtain the first comparison result; obtain the first process similarity of the first comparison result based on the Tanimoto similarity coefficient algorithm principle.

[0024] Preferably, a predetermined process labeling scheme is read. This scheme defines how process feature parameters are converted into comparable labels. Based on this scheme, the target process feature parameters (i.e., the feature parameters of the current textile process) are labeled, including data cleaning, standardization, classification, and label assignment, ensuring that each feature parameter is correctly converted into its corresponding label. The processed label is called the target process label. Similarly, based on the predetermined process labeling scheme, the process feature parameters in the first historical textile record are labeled. These historical parameters, after processing, yield the first historical process label. The target process label and the first historical process label are compared and analyzed. Since the labels have been standardized, label values ​​can be directly compared or similarity calculations can be performed to identify the similarities and differences between the current and historical processes. Based on the Tanimoto similarity coefficient algorithm, the similarity between the target process label and the first historical process label is calculated. In the comparison of process feature parameters, labels can be considered as elements in a set. The Tanimoto similarity coefficient, i.e., the first process similarity, is obtained by calculating the intersection and union of the labels.

[0025] Furthermore, it also includes:

[0026] Extract the target material feature parameters from the textile process feature parameters; perform weighted analysis on the target material feature parameters to obtain the target material quality index; use the target material quality index as a similarity adjustment coefficient to perform weighted calculation and adjustment on the first process similarity to obtain the first process similarity adjustment result.

[0027] In comparative analysis of textile processes, the similarity of processes is further adjusted by extracting target material characteristic parameters and calculating a target material quality index. Specifically, target material characteristic parameters related to material properties are selected from textile process characteristic parameters, including fiber type, fiber composition ratio, yarn count, and yarn twist. Different weights are assigned to each target material characteristic parameter based on its importance to textile process quality, combined with expert experience, historical data, or its influence in actual production. Each target material characteristic parameter is quantitatively scored, or normalized using existing standard values. Then, the score of each parameter is multiplied by its corresponding weight to obtain a weighted material quality score. All weighted material quality scores are summed to obtain the target material quality index, which reflects the overall performance of the current textile process in terms of material quality. The target material quality index is used as a similarity adjustment coefficient to adjust the weighted calculation of the first process similarity. For example, the first process similarity is multiplied by the target material quality index to obtain the adjusted first process similarity result, or a linear interpolation is performed between the original similarity and the target material quality index to obtain the adjusted similarity. The first process similarity adjustment result not only considers the similarity of process parameters, but also the influence of material properties on process similarity, thus making it more comprehensive and accurate.

[0028] Furthermore, the target material characteristic parameters include fiber length and fiber percentage for multiple fiber material types.

[0029] The target material's characteristic parameters include fiber length and fiber percentage for various fiber types. Fiber length is a crucial physical characteristic that directly affects fiber processing performance, yarn quality, and fabric style. Different fiber materials have different lengths; for example, cotton fibers are typically 25-35 mm long, while wool fibers can be even longer. Generally, longer fibers result in less textile waste. Fiber percentage refers to the proportion of different fiber types within a textile material. This proportion determines the material's overall performance and applications. Higher fiber percentages, especially for fibers with good weaving properties and longer fibers, correspond to less textile waste.

[0030] Match the first material quantity of the first material category among the multiple material quantities, and generate the first material quantity timing sequence in combination with a predetermined unit period.

[0031] From multiple material usage quantities, the material usage quantity corresponding to the first material usage category is selected, i.e., the first material usage quantity. The first material usage quantity is then organized in chronological order to ensure that each data point corresponds to a clear timestamp. According to the predetermined unit period, the data points of the first material usage quantity are divided into different time periods. For example, if the predetermined unit period is weekly, then the first material usage quantity data needs to be grouped by week to generate the first material usage quantity time series. The first material usage quantity time series reflects the changing trend of material usage quantity over time.

[0032] The consumption integrated prediction model is invoked to analyze the time series of the first material consumption to obtain the first predicted material consumption, which refers to the predicted value of the first material type consumed by the textile process in the first time period.

[0033] By analyzing the time series of the first material usage using the consumption ensemble prediction model, the first predicted material usage is obtained. The first predicted material usage refers to the predicted value of the first material category consumed by the textile process at the first time. Specifically, a machine learning model is trained based on historical material usage data, and the consumption ensemble prediction model is obtained after training. The time series of the first material usage is input into the consumption ensemble prediction model, which predicts the first material usage at a given time (the first time) and outputs the first predicted material usage. The first predicted material usage refers to the quantity of the first material category that the textile process is expected to consume at the first time.

[0034] Furthermore, this includes:

[0035] Obtain the first prediction layer in the consumption integrated prediction model, and store the first fitting polynomial of the first material consumption time series in the first prediction layer; input the first time into the first prediction layer to obtain the first predicted value; obtain the second prediction layer in the consumption integrated prediction model; input the first time series feature set of the first material consumption time series collected based on predetermined multi-domain features into the second prediction layer to obtain the second predicted value; record the average of the first predicted value and the second predicted value as the first predicted material consumption.

[0036] The first prediction layer is extracted from the consumption integrated prediction model. This first prediction layer is a polynomial fitting model based on historical data, used to capture the basic trend in the first material consumption time series. The first fitting polynomial (i.e., polynomial coefficients) of the first material consumption time series is stored in the first prediction layer. The first fitting polynomial can be obtained by fitting using time series analysis methods (such as least squares). The first time (i.e., the time point to be predicted) is input into the first prediction layer. Using the stored first fitting polynomial, polynomial calculations are performed on the first time to obtain the first predicted value. The second prediction layer is obtained from the consumption integrated prediction model. This second prediction layer is built based on machine learning or deep learning algorithms and is used to capture complex patterns and influencing factors in the first material consumption time series besides the basic trend. Based on predetermined multi-domain features, feature data related to the first material consumption time series is collected and constructed into a first time series feature set. This first time series feature set can comprehensively reflect various factors affecting the first material consumption time series. The first time series feature set is input into the second prediction layer. The second prediction layer uses these feature data, combined with its internal machine learning or deep learning algorithms, to perform complex pattern recognition and prediction to obtain the second predicted value. The average of the first and second predicted values ​​is used as the final first predicted material amount. This average combines the first predicted value based on historical data trends and the second predicted value based on multi-domain features, which improves the accuracy and robustness of the prediction results.

[0037] Furthermore, this includes:

[0038] Draw a scatter plot of the first material usage over time; obtain the first material usage spline curve of the scatter plot of the first material usage based on the principle of random sampling consistency; perform polynomial fitting regression analysis on the first material usage spline curve to obtain the first fitting polynomial.

[0039] Preferably, the first material usage time series includes a timestamp and the corresponding material usage value. A scatter plot of the first material usage time series is plotted using data visualization tools (such as Excel, Python's matplotlib library, etc.). Random sampling consistency is an iterative method for estimating the parameters of a mathematical model from a dataset containing outliers. Based on the principle of random sampling consistency, a spline curve of the first material usage is extracted from the scatter plot of the first material usage. Specifically, a portion of data points are randomly selected from the scatter plot of the first material usage as support points of the initial spline curve. A spline curve is fitted using these support points. The distance from all data points to the fitted curve is calculated, and the number of points with a distance less than the error threshold is counted. If the termination condition is met (such as reaching the maximum number of iterations or finding enough inliers), the iteration stops. The spline curve with the most inliers or the smallest average error is selected from all iterations as the final result. Polynomial fitting regression analysis is performed on the spline curve of the first material consumption. The appropriate polynomial order is selected according to the complexity of the data and the requirements. The least squares method is used to fit the data points on the spline curve to a polynomial function. The sum of squared residuals between the fitted polynomial and the data points is calculated to evaluate the fitting effect. The final polynomial function is the first fitted polynomial, which describes the trend and change law of the first material consumption time series.

[0040] Furthermore, the predetermined multi-domain features include predetermined time-domain features, predetermined frequency-domain features, and predetermined component features, wherein the predetermined component features refer to the features of the optimal material quantity component.

[0041] Predetermined multi-domain features include predetermined time-domain features, predetermined frequency-domain features, and predetermined component features. Predetermined time-domain features describe the patterns of data variation over time. In material consumption forecasting, predetermined time-domain features may include time-series data of historical material consumption, such as daily, weekly, and monthly consumption. By analyzing this time-series data, seasonal, periodic, or trend changes in material consumption can be captured. Predetermined frequency-domain features describe the characteristics of data in the frequency domain. For periodic or periodically varying material consumption data, frequency domain analysis can provide information about different frequency components. Predetermined frequency-domain features may include the results of spectral analysis, such as spectrograms and power spectral density. These features help identify the dominant frequency components in the material consumption data. Predetermined component characteristics refer to the characteristics of the optimal material usage component. In textile processing, the optimal material usage component may refer to the minimum amount of material required while meeting product quality and production efficiency requirements. Predetermined component characteristics may include various parameters related to the optimal material usage, such as the optimal material usage ratio, the optimal material usage distribution pattern, and the optimal material usage adjustment factor. These characteristics can be determined through historical data analysis and expert knowledge, and are used to guide the adjustment and optimization of material usage during the production process. In summary, predetermined time-domain characteristics and predetermined frequency-domain characteristics describe the characteristics of data in the time and frequency domains, respectively, helping to capture the changing trends and periodic changes in material usage. Predetermined component characteristics specifically refer to the characteristics related to the optimal material usage component, and are of great significance for guiding the adjustment and optimization of material usage during the production process.

[0042] Furthermore, this includes:

[0043] The first material consumption time series is subjected to modal decomposition processing by a multi-domain analyzer to obtain multiple material consumption components; the first material consumption component among the multiple material consumption components is subjected to multi-dimensional component feature collection to obtain the first material consumption component feature set; when the deviation between the first material consumption component feature set and the material consumption characteristics of the first material consumption time series reaches the deviation limit, the first material consumption component is recorded as the optimal material consumption component.

[0044] By performing mode decomposition on the first material consumption time series using a multi-domain analyzer, different material consumption variation patterns can be identified, resulting in multiple material consumption components. Specifically, an appropriate mode decomposition method is selected based on the data characteristics, such as Empirical Mode Decomposition (EMD), Variational Mode Decomposition (VMD), or Fourier Decomposition. This selected method is then applied to decompose the first material consumption time series, yielding multiple material consumption components. Multi-dimensional component features are collected from the first material consumption component among the decomposed components, including time-domain features, frequency-domain features, or other statistical features. These features describe different aspects of the first material consumption component. Specifically, appropriate features are selected based on actual needs and data characteristics. For example, time-domain features may include mean, variance, maximum, and minimum values; frequency-domain features may include spectral distribution and dominant frequency. The selected features are extracted from the first material consumption component to form a feature set for the first material consumption component. The deviation between the first material quantity component feature set and the material quantity characteristics of the first material quantity time series is compared to determine whether a deviation limit has been reached. If the deviation reaches or exceeds the limit, the first material quantity component is recorded as the optimal material quantity component. Specifically, the deviation between the first material quantity component feature set and the material quantity characteristics of the first material quantity time series is calculated, and the calculated deviation is compared with a preset deviation limit. The deviation limit can be set according to actual needs and data characteristics. If the deviation reaches or exceeds the limit, the first material quantity component is recorded as the optimal material quantity component.

[0045] Based on the first correspondence between the first predicted material consumption and the first material category, the predicted material consumption result of the textile process is generated.

[0046] Based on the first correspondence between the first predicted material consumption and the first material category, the predicted material consumption of each first material category is calculated, thereby generating the predicted material consumption results for the textile process.

[0047] In summary, the embodiments of this application have at least the following technical effects:

[0048] First, predetermined analysis dimensions are read, and features of the textile process are collected based on these dimensions to obtain textile process feature parameters. Next, the target process feature parameters are compared with the first historical process feature parameters in the first historical textile record to obtain a first comparison result. The first historical textile record represents any historical textile record of the same type of textile process. When the first process similarity obtained from the first comparison result reaches a similarity limit, the first historical material consumption record is extracted from the first historical textile record. This record includes multiple material consumption quantities for multiple material categories. Then, the first material quantity of the first material category is matched among the multiple material quantities, and a first material quantity time series is generated based on a predetermined unit period. The first material quantity time series is analyzed by calling a consumption integrated prediction model to obtain the first predicted material quantity, which refers to the predicted value of the textile process consuming the first material category at the first time. Finally, based on the first correspondence between the first predicted material quantity and the first material category, a material consumption prediction result for the textile process is generated. This invention solves the technical problem of accurately predicting material consumption in textile processes, thereby improving the accuracy of material consumption prediction in textile processes.

[0049] It should be noted that the order of the embodiments described above is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. Furthermore, the above description focuses on specific embodiments of this specification. The processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired results. In some implementations, multitasking and parallel processing are possible or may be advantageous.

[0050] The above description is only a preferred embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

[0051] This specification and accompanying drawings are merely illustrative examples of this application and are intended to cover any and all modifications, variations, combinations, or equivalents within the scope of this application. Clearly, those skilled in the art can make various alterations and modifications to this application without departing from its scope. Therefore, if such modifications and modifications fall within the scope of this application and its equivalents, this application intends to include such modifications and modifications.

Claims

1. A method for analyzing material consumption based on textile processes, characterized in that, The method includes: Read the predetermined analysis dimensions and collect features of the textile process based on the predetermined analysis dimensions to obtain textile process feature parameters; The target process feature parameter in the textile process feature parameter is compared with the first historical process feature parameter in the first historical textile record to obtain a first comparison result. The first historical textile record represents any historical textile record of the same type of textile in the textile process. When the first process similarity obtained from the analysis of the first comparison result reaches the similarity limit, the first historical material consumption record in the first historical textile record is extracted. The first historical material consumption record includes multiple material consumption quantities of multiple types of materials. Match the first material quantity of the first material category among the multiple material quantities, and generate the first material quantity timing sequence in combination with a predetermined unit period; The consumption integrated prediction model is invoked to analyze the time series of the first material consumption to obtain the first predicted material consumption. The first predicted material consumption refers to the predicted value of the first material type consumed by the textile process in the first time period. Based on the first correspondence between the first predicted material consumption and the first material category, the predicted material consumption result of the textile process is generated; Extract the target material characteristic parameters from the textile process characteristic parameters; The target material quality index is obtained by weighted analysis of the characteristic parameters of the target material; The target material quality index is used as a similarity adjustment coefficient to perform a weighted calculation and adjustment of the first process similarity, thereby obtaining the first process similarity adjustment result; The target material characteristic parameters include fiber length and fiber percentage for multiple fiber material types; Obtain the first prediction layer in the consumption integrated prediction model, and store the first fitting polynomial of the first material consumption time series into the first prediction layer; The first time is input into the first prediction layer to obtain the first prediction value; Obtain the second prediction layer in the consumption integrated prediction model; The first time-series feature set, which is collected based on the first material consumption time series based on the predetermined multi-domain features, is input into the second prediction layer to obtain the second prediction value; The average of the first predicted value and the second predicted value is recorded as the first predicted material consumption.

2. The material consumption analysis method based on textile processes according to claim 1, characterized in that, include: Read the pre-defined process label scheme; According to the predetermined process labeling scheme, the target process feature parameters and the first historical process feature parameters are labeled sequentially to obtain the target process label and the first historical process label, respectively. By comparing and analyzing the target process label and the first historical process label, the first comparison result is obtained; The first process similarity of the first comparison result is obtained based on the Tanimoto similarity coefficient algorithm principle.

3. The material consumption analysis method based on textile processes according to claim 1, characterized in that, include: Draw a scatter plot of the first material consumption time sequence; The first material usage spline curve of the scatter plot of the first material usage is obtained based on the principle of random sampling consistency. A polynomial fitting regression analysis was performed on the first material usage spline curve to obtain the first fitting polynomial.

4. The material consumption analysis method based on textile processes according to claim 1, characterized in that, The predetermined multi-domain features include predetermined time-domain features, predetermined frequency-domain features, and predetermined component features, wherein the predetermined component features refer to the features of the optimal material quantity component.

5. The material consumption analysis method based on textile processes according to claim 4, characterized in that, include: The first material consumption time series is subjected to modal decomposition processing by a multi-domain analyzer to obtain multiple material consumption components. Multi-dimensional component features are collected from the first material quantity component among the multiple material quantity components to obtain the first material quantity component feature set; When the deviation between the first material quantity component feature set and the material quantity feature of the first material quantity time series reaches the deviation limit, the first material quantity component is recorded as the optimal material quantity component.

Images