Roving frame spinning quality prediction method and system based on machine learning

Abstract

The invention discloses a roving frame spinning quality prediction method and system based on machine learning, and relates to the technical field of spinning quality prediction.The roving frame spinning quality prediction method comprises the following steps that historical yarn quality data are collected, yarn quality is divided into a plurality of quality gears, and an environment index and a machine state index corresponding to each quality gear are obtained; performing correlation analysis on the environment indexes and the machine state indexes to obtain the optimal correlation degree of the environment indexes and the yarn quality and the optimal correlation degree of the machine state indexes and the yarn quality respectively; inputting the optimal relevancy between the environment index and the yarn quality and the optimal relevancy between the machine state index and the yarn quality into a preset machine learning model for training to obtain a spinning quality prediction model of the roving frame; and inputting to-be-predicted yarn data into the roving frame spinning quality prediction model for quality prediction analysis. The problem that the accuracy of roving frame spinning quality prediction is reduced in the modern textile production process is solved.

Description

Technical Field

[0001] The present invention relates to the technical field of spinning quality prediction, and more specifically, to a method and system for predicting spinning quality of a roving frame based on machine learning. Background Art

[0002] The background of roving frame spinning quality prediction mainly stems from the textile industry's high requirements for product quality and the challenges of quality control in the production process. The quality of textiles directly affects the performance and market competitiveness of the final product, and the quality of yarn, as the basic material in textile production, has an important impact on the durability, feel, gloss, strength, etc. of the fabric. The roving frame is the key equipment responsible for twisting fibers into yarns in spinning production. The quality of roving directly affects the subsequent textile process and the quality of the finished product.

[0003] However, due to the complexity and variability of modern textile production processes, a single quality assessment method is difficult to fully reflect the changing trend of yarn quality, resulting in low accuracy in roving frame spinning quality prediction.

[0004] In view of the above problems, the present invention proposes a solution. Summary of the invention

[0005] In order to overcome the above-mentioned defects of the prior art, an embodiment of the present invention provides a roving frame spinning quality prediction method and system based on machine learning, which solves the problem of reduced accuracy of roving frame spinning quality prediction in modern textile production processes.

[0006] To achieve the above object, the present invention provides the following technical solutions:

[0007] A roving frame spinning quality prediction method based on machine learning includes the following steps: collecting historical yarn quality data, dividing the yarn quality into several quality gears, and obtaining environmental indicators and machine status indicators corresponding to each quality gear; performing correlation analysis on the environmental indicators and machine status indicators to obtain the best correlation between the environmental indicators and the yarn quality and the best correlation between the machine status indicators and the yarn quality respectively; inputting the best correlation between the environmental indicators and the yarn quality and the best correlation between the machine status indicators and the yarn quality into a preset machine learning model for training to obtain a roving frame spinning quality prediction model; and inputting the yarn data to be predicted into the roving frame spinning quality prediction model for quality prediction analysis.

[0008] In a preferred embodiment, the historical quality data of the yarn is collected, the yarn quality is divided into several quality gears, and the environmental indicators and machine status indicators corresponding to each quality gear are obtained, specifically: the maximum yarn quality and the minimum yarn quality in the historical quality data of the yarn are obtained; the maximum yarn quality and the minimum yarn quality are input into a preset quality gear division formula to obtain several quality gears; the environmental indicators and machine status indicators corresponding to each quality gear are obtained, the environmental indicators include but are not limited to temperature, humidity and air flow speed; the machine status indicators include but are not limited to machine speed, voltage and voltage.

[0009] In a preferred embodiment, correlation analysis is performed on the environmental indicators and machine status indicators to obtain the best correlation between the environmental indicators and the yarn quality and the best correlation between the machine status indicators and the yarn quality, respectively. Specifically, the comprehensive influence of the environmental factors at time point t is calculated based on the comprehensive influence calculation formula of the environmental factors using a sliding window method; the comprehensive influence of the environmental factors at time point t is input into the linear correlation coefficient calculation formula to obtain the correlation between the environmental indicators and the yarn quality; the comprehensive influence of the machine status at time point t is calculated based on the comprehensive influence formula of the machine status using a sliding window method; the comprehensive influence of the machine status at time point t is input into the linear correlation coefficient calculation formula to obtain the correlation between the machine indicators and the yarn quality.

[0010] In a preferred embodiment, the optimal correlation between the environmental indicators and yarn quality and the optimal correlation between the machine state indicators and yarn quality are input into a preset machine learning model for training to obtain a roving frame spinning quality prediction model, specifically: the optimal correlation between the environmental indicators and yarn quality and the optimal correlation between the machine state indicators and yarn quality, environmental indicators, machine state indicators and historical yarn quality data are combined into a data set; the data set is divided into a training set and a validation set according to a preset ratio; the training data set is input into a preset machine learning model for training to obtain a trained model and an evaluation result; the trained model is verified using a validation set, and the parameters of the trained model are optimized according to the evaluation result to obtain a roving frame spinning quality prediction model.

[0011] In a preferred embodiment, the sliding window method is used to calculate the comprehensive influence of environmental factors at time point t based on the environmental factor comprehensive influence calculation formula, specifically: randomly select a time period W to define the size of a sliding window, and the size of the sliding window can be adjusted according to actual conditions; calculate the comprehensive influence of environmental factors at time point t based on the environmental factor comprehensive influence calculation formula; whenever the time point t increases, the window moves forward and recalculates the comprehensive influence of environmental factors at the new time point t.

[0012] In a preferred embodiment, the yarn data to be predicted is input into the roving frame spinning quality prediction model for quality prediction analysis, specifically: the yarn data to be predicted is input into the roving frame spinning quality prediction model to obtain a quality evaluation value; if the quality evaluation value is greater than a preset quality threshold, the yarn quality is unqualified; if the quality evaluation value is less than or equal to the preset quality threshold, the yarn quality is qualified

[0013] Technical effects and advantages of the roving frame spinning quality prediction method and system based on machine learning of the present invention:

[0014] 1. The present invention can more accurately analyze the differences between different quality levels by collecting yarn historical quality data and dividing it into several quality gears, and provide a clear direction for the optimization of the spinning process. For each quality gear, the corresponding environmental indicators (such as temperature, humidity, etc.) and machine status indicators (such as speed, current, etc.) are collected, which can provide comprehensive data support for subsequent analysis. This method can reveal the intrinsic connection between yarn quality changes and external factors and machine status through correlation analysis between environmental indicators and yarn quality, and between machine status indicators and yarn quality, so as to find the best influencing factors. By calculating the correlation between environment and machine status on yarn quality, the specific influence of various factors on quality can be quantified, providing a scientific basis for optimizing the spinning process and improving production efficiency. In addition, correlation analysis can effectively screen out the factors that have the greatest influence on yarn quality, help focus on optimization strategies, and improve stability and controllability in the production process. The advantage of this method lies in its systematicity and data-driven nature, which can provide strong support for the construction of a spinning quality prediction model, thereby achieving accurate prediction and real-time regulation, and improving the intelligent level of textile production.

[0015] 2. The present invention can achieve efficient prediction of yarn quality by inputting the best correlation between environmental indicators and yarn quality and the best correlation between machine state indicators and yarn quality into the machine learning model for training. In this way, the model can automatically learn the influence of environment and machine state on yarn quality based on historical data, thereby improving the accuracy and reliability of prediction. Compared with the traditional artificial experience method, machine learning can process a large amount of complex data and discover the nonlinear relationship hidden behind the data, so as to better cope with the changes in environment and machine state during the production process. In addition, the quality prediction model based on machine learning has the ability of self-optimization. With the accumulation of more data and continuous training, the prediction ability of the model will gradually improve, and it can adapt to the quality changes under different production conditions. Inputting the yarn data to be predicted into the model for quality prediction analysis can provide accurate quality prediction in real time, provide strong support for quality control and process adjustment in the production process, help enterprises reduce quality fluctuations, reduce production costs, and improve product consistency and stability. This intelligent prediction method not only improves production efficiency, but also promotes the transformation of the textile industry to digitalization and intelligence. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 It is a schematic flow chart of the method for predicting spinning quality of roving frame based on machine learning of the present invention.

[0017] Figure 2 It is a structural schematic diagram of the roving frame spinning quality prediction system based on machine learning of the present invention. DETAILED DESCRIPTION

[0018] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0019] Embodiment 1, Figure 1 The present invention provides a method for predicting spinning quality of a roving frame based on machine learning, which includes the following steps:

[0020] S1, collecting historical yarn quality data, dividing the yarn quality into several quality levels, and obtaining the environmental index and machine status index corresponding to each quality level;

[0021] In this example, historical yarn quality data is collected, the yarn quality is divided into several quality levels, and the environmental indicators and machine status indicators corresponding to each quality level are obtained, specifically:

[0022] Obtaining the maximum yarn quality and the minimum yarn quality in the yarn historical quality data;

[0023] Inputting the maximum yarn mass and the minimum yarn mass into a preset quality level division formula to obtain a plurality of quality levels;

[0024] Obtain environmental indicators and machine status indicators corresponding to each quality gear, the environmental indicators include but are not limited to temperature, humidity and air flow speed; the machine status indicators include but are not limited to machine speed, voltage and voltage.

[0025] It should be noted that by collecting yarn historical quality data and dividing it into several quality gears, the subtle differences between different quality grades can be accurately captured, helping the production process to control quality more scientifically. First, by obtaining the maximum and minimum quality in the yarn historical data and inputting the preset quality gear division formula, the yarn quality can be systematically divided into multiple grades, thereby providing a clear classification basis for subsequent analysis. This division method not only simplifies the quality assessment process, but also equips each quality gear with corresponding environmental and machine status indicators. These environmental indicators (such as temperature, humidity, air flow speed, etc.) and machine status indicators (such as speed, voltage, etc.) are key factors affecting quality in textile production. By associating the specific conditions of different quality gears, it helps to reveal the comprehensive impact of environmental and machine status on yarn quality. The advantage of this method is that it can quantify and refine the causes of quality fluctuations, ensure that each quality gear has corresponding production environment and machine status indicators as a basis, and provide data support for optimizing production processes, achieving precise control, and improving production stability. Through this scientific quality gear division method, textile production will be more intelligent and refined, and improve overall production efficiency and product consistency.

[0026] In this example, the quality level classification formula is as follows:

[0027]

[0028] S mid =S min +(S max -S min )×β

[0029] Shigh=Smax

[0030] Among them, S low is the upper bound of the low quality gear, S min is the minimum yarn mass, S max is the maximum yarn quality, S is the coefficient for dividing low-quality and medium-quality gears. midis the upper limit of the medium quality gear, β is the coefficient for dividing the medium quality and high quality gears, S high It is the upper limit of high quality level.

[0031] S2, performing correlation analysis on the environmental index and the machine status index to obtain the best correlation between the environmental index and the yarn quality and the best correlation between the machine status index and the yarn quality respectively;

[0032] In this example, correlation analysis is performed on the environmental index and the machine status index to obtain the best correlation between the environmental index and the yarn quality and the best correlation between the machine status index and the yarn quality, which are specifically:

[0033] The sliding window method is used to calculate the comprehensive impact of environmental factors at time point t based on the calculation formula of comprehensive impact of environmental factors;

[0034] Inputting the comprehensive influence of environmental factors at the time point t into the linear correlation coefficient calculation formula to obtain the correlation between environmental indicators and yarn quality;

[0035] The sliding window method is used to calculate the comprehensive influence of the machine state at time point t based on the comprehensive influence formula of the machine state;

[0036] The comprehensive influence of the machine state at the time point t is input into the linear correlation coefficient calculation formula to obtain the correlation between the machine index and the yarn quality.

[0037] In this example, the calculation formula for the comprehensive impact of environmental factors is as follows:

[0038]

[0039] Among them, L env (t) is the comprehensive influence calculated at time t, wi is the weight of time point i, Ti, Hi and Vi are the temperature, humidity and air flow rate at time i respectively, W is the sliding window size, and t is the time point.

[0040] In this example, the comprehensive influence formula of the machine status is as follows:

[0041]

[0042] Among them, L mac (t) is the comprehensive influence of the machine state at time point t, K is the number of weight functions, ΔNt, ΔVt and ΔIt are the change rates of speed, voltage and current at time point t, respectively, N , σV and σI are the nonlinear weight function values ​​of the speed, voltage and current at the time point t respectively.

[0043] In this example, the sliding window method is used to calculate the comprehensive impact of environmental factors at time point t based on the calculation formula of the comprehensive impact of environmental factors, which is specifically:

[0044] Randomly select a time period W to define the size of a sliding window, and the size of the sliding window can be adjusted according to actual conditions;

[0045] The comprehensive impact of environmental factors at time point t is calculated based on the calculation formula of comprehensive impact of environmental factors;

[0046] Whenever the time point t increases, the window moves forward and the comprehensive impact of environmental factors at the new time point t is recalculated.

[0047] It should be noted that the sliding window method can dynamically capture and evaluate the real-time impact of environmental factors on yarn quality based on the calculation formula of the comprehensive impact of environmental factors. First, by randomly selecting a time period W, the size of the sliding window is defined, and the window size can be adjusted according to the actual production situation. This flexibility enables the method to adapt to different production environments and data fluctuations, and helps to capture the impact of environmental changes on yarn quality at different time scales. For example, if the environment changes relatively slowly during the production process, a larger window can be selected for a smooth transition; if the environment changes rapidly, a smaller window can be selected to respond to fluctuations in a timely manner.

[0048] In each window, based on the calculation formula of the comprehensive impact of environmental factors, the comprehensive impact of multiple environmental factors (such as temperature, humidity, air flow speed, etc.) on yarn quality can be evaluated. By calculating the comprehensive impact of environmental factors at each time point t, a dynamic and constantly updated impact value can be obtained, which reflects the effect of environmental factors on yarn quality in real time. Whenever the time point ttt increases, the window moves forward and recalculates the impact of the new time point, thereby continuously updating and correcting the prediction of yarn quality.

[0049] The advantage of this method is that it is highly real-time and adaptable. The sliding window can capture the fluctuations of environmental factors in a timely manner as the production process changes, avoiding information lags that may be caused by static data analysis. In addition, the flexible adjustment of the window size provides more control space, allowing the method to optimize the calculation according to actual conditions and ensure a more accurate assessment of the impact on yarn quality. This dynamic and continuous evaluation method helps to improve the quality prediction accuracy of the spinning process and supports more intelligent and refined control strategies in the production process.

[0050] Furthermore, the comprehensive influence of environmental factors and machine status is calculated by the sliding window method, which can dynamically capture the real-time impact of the environment and machine status on yarn quality at different time points. This process first calculates the comprehensive influence of environmental factors at each time point t by the sliding window method to consider the fluctuating effects of environmental changes such as temperature, humidity, and air flow rate on yarn quality. The sliding window technology can effectively smooth noise and avoid the interference of accidental fluctuations in data at a single moment, thereby obtaining a more accurate environmental impact. Then, the calculated comprehensive influence of environmental factors is input into the linear correlation coefficient calculation formula to obtain the correlation between environmental factors and yarn quality, which can quantify the specific impact of the environment on yarn quality, thereby providing a scientific basis for adjusting the production environment.

[0051] Similarly, the comprehensive influence of machine status is also calculated through the sliding window method, and the influence of machine status on yarn quality is evaluated according to the changes in machine parameters such as speed and voltage. The comprehensive influence of machine status is input into the linear correlation coefficient calculation formula to obtain the correlation between machine status and yarn quality. This method can reflect the impact of machine operation status on yarn quality in real time, helping the production team to adjust machine parameters in time to ensure stable quality.

[0052] This method based on sliding window and linear correlation analysis is highly dynamic and real-time. Through continuous data analysis, the time-varying effects of different environments and machine states on quality can be accurately captured, avoiding the shortcomings of static analysis. At the same time, the calculation of the linear correlation coefficient provides a clear quantitative indicator, which allows the contribution of each factor to quality to be quantified and compared, providing strong data support for optimizing production processes and achieving refined management. The advantages of this method lie in its flexibility, real-time and accuracy, which can effectively promote the development of textile production towards intelligence and precision, and improve production efficiency and product consistency.

[0053] S3, inputting the optimal correlation between the environmental index and the yarn quality and the optimal correlation between the machine state index and the yarn quality into a preset machine learning model for training to obtain a roving frame spinning quality prediction model;

[0054] In this example, the optimal correlation between the environmental index and the yarn quality and the optimal correlation between the machine state index and the yarn quality are input into a preset machine learning model for training to obtain a roving frame spinning quality prediction model, specifically:

[0055] Combining the best correlation between the environmental index and the yarn quality, the best correlation between the machine status index and the yarn quality, the environmental index, the machine status index and the historical yarn quality data into one data set;

[0056] Dividing the data set into a training set and a validation set according to a preset ratio;

[0057] Input the training data set into a preset machine learning model for training, and obtain a trained model and evaluation results;

[0058] The trained model is verified using a verification set, and parameters of the trained model are tuned according to the evaluation results to obtain a roving frame spinning quality prediction model.

[0059] It should be noted that by combining the best correlation between environmental indicators and yarn quality, the best correlation between machine status indicators and yarn quality, environmental indicators, machine status indicators and historical yarn quality data into a complete data set, comprehensive input information can be provided for roving frame spinning quality prediction. The advantage of this process is that it can comprehensively consider the impact of multiple factors on yarn quality, including not only the direct impact of environmental factors and machine status, but also the integration of historical quality data to ensure that the model has enough information for accurate prediction.

[0060] Next, the data set is divided into a training set and a validation set according to a preset ratio to ensure that the model training process has good generalization ability and avoids overfitting. The training set is used to learn the model, while the validation set is used to evaluate the performance of the model on unseen data to ensure that it can accurately predict the yarn quality in actual production. By inputting data from the training set into the machine learning model for training, the parameters and structure of the model can be optimized, thereby obtaining an efficient and accurate roving frame spinning quality prediction model.

[0061] During the training process, the evaluation results provide the necessary feedback for the model, so that the parameter tuning performed on the validation set can further improve the prediction accuracy of the model. Through continuous iteration and optimization, the trained model can more accurately reflect the changing trend of yarn quality under different environmental conditions and machine conditions. The final roving frame spinning quality prediction model can predict yarn quality in real time in actual production, providing a scientific basis for production process optimization and quality control.

[0062] The advantage of this method is that, through the automatic learning ability of the machine learning model, it can not only process complex multi-factor data, but also continuously optimize and adapt to various changes in the production process, significantly improve the intelligence level of textile production, reduce quality fluctuations, and improve production efficiency and product consistency.

[0063] S4, inputting the yarn data to be predicted into the roving frame spinning quality prediction model for quality prediction analysis.

[0064] In this example, the yarn data to be predicted is input into the roving frame spinning quality prediction model for quality prediction analysis, specifically:

[0065] Inputting the yarn data to be predicted into the roving frame spinning quality prediction model to obtain a quality evaluation value;

[0066] If the quality evaluation value is greater than a preset quality threshold, the yarn quality is unqualified;

[0067] If the quality evaluation value is less than or equal to the preset quality threshold, the yarn quality is qualified.

[0068] It should be noted that by inputting the yarn data to be predicted into the roving frame spinning quality prediction model, an accurate assessment of the yarn quality can be achieved. This process first inputs the real-time yarn data (such as machine status, environmental factors, etc.) collected during the current production process into the trained quality prediction model. The model calculates a quality assessment value based on the rules learned from historical data. This assessment value is a prediction result based on the comprehensive impact of factors such as the environment and machine status on the yarn quality, providing a basis for subsequent quality judgment.

[0069] After obtaining the quality assessment value, the model compares it with the preset quality threshold to determine whether the yarn quality is qualified. If the quality assessment value is greater than the preset quality threshold, it means that the yarn quality is unqualified and needs to be adjusted or optimized to ensure that the products in the production process meet the standards. If the quality assessment value is less than or equal to the quality threshold, it means that the yarn quality is qualified and can continue to enter the subsequent process or delivery. This automated judgment method can perform quality inspection efficiently and accurately, avoiding the errors and lags that may be caused by manual inspection.

[0070] The advantages of this method are its efficiency and accuracy. With the support of machine learning models, the evaluation of yarn quality no longer relies on manual experience and a single parameter, but is based on intelligent analysis of multi-dimensional data, which can more comprehensively capture the quality change trend. In addition, real-time quality prediction can detect problems in a timely manner, prevent unqualified products from entering the production chain, reduce production losses, and improve overall product consistency and stability. Through the application of this intelligent prediction model, the spinning production process can be more refined and automated, effectively improving production efficiency, reducing costs, and enhancing the market competitiveness of products.

[0071] Embodiment 2, Figure 2 The present invention provides a roving frame spinning quality prediction system based on machine learning, including a data acquisition module, a relevance module, a model building module, and a quality prediction module:

[0072] A data acquisition module is used to collect historical yarn quality data, divide the yarn quality into several quality levels, and obtain the environmental indicators and machine status indicators corresponding to each quality level;

[0073] A correlation module, used for performing correlation analysis on the environmental index and the machine status index to obtain the best correlation between the environmental index and the yarn quality and the best correlation between the machine status index and the yarn quality respectively;

[0074] A model building module, used for inputting the optimal correlation between the environmental index and the yarn quality and the optimal correlation between the machine state index and the yarn quality into a preset machine learning model for training, so as to obtain a roving frame spinning quality prediction model;

[0075] The quality prediction module is used to input the yarn data to be predicted into the roving frame spinning quality prediction model for quality prediction analysis.

[0076] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions.

[0077] The above embodiments may be implemented in whole or in part by software, hardware, firmware or any other combination. When implemented by software, the above embodiments may be implemented in whole or in part in the form of a computer program product.

[0078] Those of ordinary skill in the art will appreciate that the modules and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0079] In addition, each functional module in each embodiment of the present application may be integrated into one processing module, or each module may exist physically separately, or two or more modules may be integrated into one module.

[0080] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

[0081] Finally: The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for predicting spinning quality of roving frames based on machine learning, characterized in that: The following steps are involved: Collect historical yarn quality data, divide the yarn quality into several quality levels, and obtain the environmental indicators and machine status indicators corresponding to each quality level; Performing correlation analysis on the environmental index and the machine status index to obtain the best correlation between the environmental index and the yarn quality and the best correlation between the machine status index and the yarn quality respectively; Inputting the optimal correlation between the environmental index and the yarn quality and the optimal correlation between the machine state index and the yarn quality into a preset machine learning model for training to obtain a roving frame spinning quality prediction model; The yarn data to be predicted is input into the roving frame spinning quality prediction model for quality prediction analysis.

2. The method for predicting spinning quality of roving frame based on machine learning according to claim 1, characterized in that: The yarn historical quality data is collected, the yarn quality is divided into several quality gears, and the environmental index and machine status index corresponding to each quality gear are obtained, specifically: Obtaining the maximum yarn quality and the minimum yarn quality in the yarn historical quality data; Inputting the maximum yarn mass and the minimum yarn mass into a preset quality level division formula to obtain a plurality of quality levels; Obtain environmental indicators and machine status indicators corresponding to each quality gear, the environmental indicators include but are not limited to temperature, humidity and air flow speed; the machine status indicators include but are not limited to machine speed, voltage and voltage.

3. The method for predicting spinning quality of roving frame based on machine learning according to claim 2, characterized in that: The best correlation between the environmental index and the machine status index and the best correlation between the environmental index and the yarn quality and the best correlation between the machine status index and the yarn quality are obtained by performing correlation analysis on the environmental index and the machine status index, respectively, which are: The sliding window method is used to calculate the comprehensive impact of environmental factors at time point t based on the calculation formula of comprehensive impact of environmental factors; Inputting the comprehensive influence of environmental factors at the time point t into the linear correlation coefficient calculation formula to obtain the correlation between environmental indicators and yarn quality; The sliding window method is used to calculate the comprehensive influence of the machine state at time point t based on the comprehensive influence formula of the machine state; The comprehensive influence of the machine state at the time point t is input into the linear correlation coefficient calculation formula to obtain the correlation between the machine index and the yarn quality.

4. The method for predicting spinning quality of roving frame based on machine learning according to claim 3, characterized in that: The optimal correlation between the environmental index and the yarn quality and the optimal correlation between the machine status index and the yarn quality are input into a preset machine learning model for training to obtain a roving frame spinning quality prediction model, specifically: Combining the best correlation between the environmental index and the yarn quality, the best correlation between the machine status index and the yarn quality, the environmental index, the machine status index and the historical yarn quality data into one data set; Dividing the data set into a training set and a validation set according to a preset ratio; Input the training data set into a preset machine learning model for training, and obtain a trained model and evaluation results; The trained model is verified using a verification set, and parameters of the trained model are tuned according to the evaluation results to obtain a roving frame spinning quality prediction model.

5. The method for predicting spinning quality of roving frame based on machine learning according to claim 4, characterized in that: The sliding window method is used to calculate the comprehensive impact of environmental factors at time point t based on the comprehensive impact calculation formula of environmental factors, specifically: Randomly select a time period W to define the size of a sliding window, and the size of the sliding window can be adjusted according to actual conditions; The comprehensive impact of environmental factors at time point t is calculated based on the calculation formula of comprehensive impact of environmental factors; Whenever the time point t increases, the window moves forward and the comprehensive impact of environmental factors at the new time point t is recalculated.

6. The method for predicting spinning quality of roving frame based on machine learning according to claim 5, characterized in that: The method of inputting the yarn data to be predicted into the roving frame spinning quality prediction model for quality prediction analysis is specifically as follows: Inputting the yarn data to be predicted into the roving frame spinning quality prediction model to obtain a quality evaluation value; If the quality evaluation value is greater than a preset quality threshold, the yarn quality is unqualified; If the quality evaluation value is less than or equal to the preset quality threshold, the yarn quality is qualified.

7. The method for predicting spinning quality of roving frame based on machine learning according to claim 6, characterized in that: The quality level classification formula is specifically: S mid =S min +(S max -S min )×β S high =S max Among them, S low is the upper bound of the low quality gear, S min is the minimum yarn mass, S max is the maximum yarn quality, S is the coefficient for dividing low-quality and medium-quality gears. mid is the upper limit of the medium quality gear, β is the coefficient for dividing the medium quality and high quality gears, S high It is the upper limit of high quality level.

8. The method for predicting spinning quality of roving frame based on machine learning according to claim 7, characterized in that: The calculation formula for the comprehensive impact of environmental factors is specifically: Among them, L env (t) is the comprehensive influence calculated at time t, w i is the weight of time point i, T i , H i and V i are the temperature, humidity, and air velocity at time i, W is the sliding window size, and t is the time point.

9. The method for predicting spinning quality of roving frame based on machine learning according to claim 8, characterized in that: The comprehensive influence formula of the machine state is specifically: Among them, L mac (t) is the comprehensive influence of the machine state at time point t, K is the number of weight functions, ΔN t , ΔV t and ΔI t are the rates of change of speed, voltage and current at time point t, σ N , σ V and σ I They are the nonlinear weight function values ​​of the speed, voltage and current at time point t respectively.

10. A system using the roving frame spinning quality prediction method based on machine learning as described in any one of claims 1 to 9, characterized in that: Including data acquisition module, relevance module, model building module, quality prediction module: A data acquisition module is used to collect historical yarn quality data, divide the yarn quality into several quality levels, and obtain the environmental indicators and machine status indicators corresponding to each quality level; A correlation module, used for performing correlation analysis on the environmental index and the machine status index to obtain the best correlation between the environmental index and the yarn quality and the best correlation between the machine status index and the yarn quality respectively; A model building module, used for inputting the optimal correlation between the environmental index and the yarn quality and the optimal correlation between the machine state index and the yarn quality into a preset machine learning model for training, so as to obtain a roving frame spinning quality prediction model; The quality prediction module is used to input the yarn data to be predicted into the roving frame spinning quality prediction model for quality prediction analysis.

Images