Intelligent data acquisition system for weaving machine

Abstract

The invention relates to the technical field of textile production, and discloses a loom data intelligent acquisition system, which collects performance, environment and configuration parameters of a loom in real time through a data acquisition module, and ensures effective communication with various loom controllers in a micro-service mode. The data preprocessing module performs cleaning, format unification and standardization processing on the collected data to form a feature data set. Based on this, a performance prediction model building module builds a prediction model to predict the variation trend of future performance parameters of the loom. And the real-time data acquisition module and the prediction data generation module work cooperatively to provide data support for performance monitoring. The system also has the functions of performance monitoring, data comparison and analysis, deviation detection and alarm, and can monitor the performance of the loom in real time, compare prediction and actual data, timely find and mark deviation, trigger an alarm mechanism, and send detailed alarm information to operation and maintenance personnel, thereby realizing intelligent management of the loom.

Description

Technical Field

[0001] The invention relates to the technical field of textile production, and in particular to an intelligent data collection system for a loom. Background Art

[0002] In the textile industry, looms are core equipment in the production process, and their operating efficiency and performance status are of vital importance to ensuring production quality, improving production efficiency, and controlling production costs. However, with the continuous expansion of production scale and the increasing production requirements, the current loom management and monitoring methods can no longer meet the needs of modern textile production. Although advanced technologies such as the Internet of Things, big data, and artificial intelligence have been widely used in various industries, providing new ideas and means for loom monitoring and management, the loom data acquisition system in the existing technology still has many shortcomings. Specifically, these problems are mainly reflected in the following aspects:

[0003] Existing systems often only focus on some key performance parameters of the loom, while ignoring the important impact of environmental parameters (such as temperature, humidity, etc.) and loom configuration parameters (such as machine model, accessory specifications, etc.) on loom performance. This one-sided data collection method makes it impossible for the system to fully and accurately reflect the actual operating status of the loom, thus affecting subsequent data analysis and decision support.

[0004] Most existing systems lack intelligent performance prediction capabilities based on historical and real-time data. This means that the system cannot predict the future performance parameter trends of the loom based on its historical operating data and current status. This defect makes it difficult for companies to detect potential failures or performance degradation in advance, making it impossible to perform preventive maintenance, increasing the risk of production interruptions and repair costs.

[0005] The existing system's monitoring and alarm mechanism often has problems such as slow response speed and insufficiently detailed alarm information. When the loom is abnormal or fails, the system may not be able to issue an alarm in time, or the alarm information provided is too general to help operation and maintenance personnel quickly locate the problem and take effective countermeasures. This delayed and inaccurate alarm mechanism seriously affects production efficiency and the timeliness of equipment maintenance. Summary of the invention

[0006] The purpose of the present invention is to provide a loom data intelligent collection system to solve the problems raised in the above background technology.

[0007] To achieve the above object, the present invention provides the following technical solution: a loom data intelligent collection system, the system comprising:

[0008] The data acquisition module is used to collect various performance parameters, environmental parameters and loom configuration parameters of the loom in real time during operation; the data acquisition module adopts a microservice mode to build multiple front-end adapters for data reception. In the data acquisition module, each type of loom corresponds to a matching front-end adapter; the front-end adapter is responsible for establishing a communication mechanism with various loom controllers, including the formulation of communication protocols, the definition of data formats, disconnection processing and status monitoring;

[0009] The data preprocessing module is connected to the data acquisition module and is used to clean, unify and standardize the collected data, and then extract features to form a feature data set;

[0010] The performance prediction model building module builds a performance prediction model for predicting the future performance parameter change trend of the loom based on the feature data set, combined with environmental parameters and loom configuration parameters;

[0011] A real-time data acquisition module is used to acquire the current operating status data and environmental condition data of the loom in real time;

[0012] A prediction data generation module inputs the data acquired by the real-time data acquisition module into the performance prediction model to generate prediction performance parameters;

[0013] The performance monitoring module is used to monitor the performance of the loom in real time, obtain the actual performance parameter data, and perform anomaly detection on the original monitoring data, and remove or correct abnormal values;

[0014] The data comparison and analysis module plots the real-time performance parameter data obtained by the performance monitoring module and the predicted performance parameter data output by the predicted data generation module into time-varying curves, aligns the timestamps, and places the two curves in the same coordinate system for comparison;

[0015] The deviation detection module sets the performance deviation threshold, determines whether each corresponding point on the two curves in the data comparison and analysis module has a deviation exceeding the threshold, and marks the deviation point;

[0016] The alarm module is connected to the deviation detection module. When a deviation point is detected, the alarm mechanism is triggered and an alarm message is sent to the operation and maintenance personnel. The alarm message includes the specific value of the deviation, the corresponding performance indicator, the time of occurrence and the recommended response measures.

[0017] Preferably, the performance parameters include loom speed, fabric tension, yarn consumption rate, motor load, loom temperature, and failure rate; the environmental parameters include workshop temperature, humidity, and air cleanliness; and the loom configuration parameters include loom model, yarn type, fabric specifications, and degree of wear of loom components.

[0018] Preferably, the data acquisition module also includes a service registration unit, an automation management unit and a log management unit; the service registration unit provides registration and status monitoring functions for microservice instances; the automation management unit is responsible for the deployment, start, stop and restart operations of microservices; the log management unit collects the operation data and log information of the service for data monitoring and management.

[0019] Preferably, the deviation detection module further comprises:

[0020] A threshold setting unit is used to set reasonable deviation thresholds for various performance indicators based on historical performance monitoring data, loom requirements, and operation and maintenance experience;

[0021] The deviation detection unit is used to traverse the aligned data, calculate the deviation value between the real-time performance monitoring data and the predicted data, and mark the deviation points that exceed the threshold.

[0022] Preferably, the performance prediction model construction module adopts a long short-term memory network LSTM algorithm to construct a performance prediction model, and the network structure of the model includes an input layer, at least one LSTM hidden layer and an output layer, wherein the input layer receives a feature data set, the LSTM hidden layer is used to capture long-term dependencies in time series data, and the output layer outputs predicted performance parameter values.

[0023] Preferably, the step of training the performance prediction model includes:

[0024] Step 1: Randomly divide the preprocessed feature data set into training set, validation set and test set. The training set is used for model training, the validation set is used for model verification and parameter adjustment, and the test set is used to evaluate the final model performance.

[0025] Step 2: Build an LSTM neural network model, set the model's input layer, LSTM layer, fully connected layer, and output layer, and initialize the model parameters;

[0026] Step 3: Use the training set to iteratively train the LSTM model, calculate the gradient of the loss function through the back propagation algorithm, and use the gradient descent method to update the model parameters. The parameter update formula is:

[0027]

[0028] Among them, θ represents the model parameters, including the weights and biases of the convolutional layers, η is the learning rate, which is used to control the step size of parameter updates, is the gradient of the loss function with respect to the parameter θ;

[0029] The loss function adopts the cross entropy loss function, and its calculation formula is:

[0030]

[0031] Where N is the number of samples, y i is the true label of the i-th sample, p i is the predicted probability of the i-th sample;

[0032] Step 4: After each round of training, use the validation set to evaluate the performance of the model. Adjust the model's hyperparameters, including the learning rate, the number of units in the LSTM layer, and the number of training rounds, based on the performance on the validation set to minimize the loss on the validation set.

[0033] Step 5: After model training and parameter adjustment are completed, use the test set to evaluate the performance of the final model and quantify the performance of the model by calculating the prediction accuracy and mean square error indicators on the test set.

[0034] Preferably, in step 4, an early stopping method is used to prevent model overfitting. When the loss on the validation set no longer decreases within a certain number of rounds, the training process is stopped, and the model parameters with the best performance on the validation set are saved as the final model parameters.

[0035] Preferably, the data preprocessing module further comprises:

[0036] Data cleaning unit, used to process missing values ​​and outliers in collected data;

[0037] Data format unification unit, used to convert data from different sources into a unified format;

[0038] The data standardization unit is used to standardize the data and eliminate the dimensional differences in the distribution range of different data.

[0039] Preferably, the method for the data cleaning unit to process missing values ​​includes:

[0040] For numerical features, the median of the non-missing values ​​of the feature is used to fill the missing values;

[0041] For categorical features, the missing values ​​are filled with the mode of the non-missing values ​​of the feature;

[0042] For missing values ​​in time series data, the average value of the adjacent values ​​is used to fill them;

[0043] For features with obvious business meaning and few missing values, they are manually filled according to business logic;

[0044] The method for processing abnormal values ​​by the data cleaning unit includes:

[0045] For numerical features, the 3σ principle is used to identify and process outliers; specifically, the mean and standard deviation of the feature are calculated, and values ​​exceeding the mean ± 3 times the standard deviation are considered outliers.

[0046] Preferably, the method for the data standardization unit to perform standardization processing on the data is: using the Z-score method to subtract the mean of each numerical data and divide it by its standard deviation, so that the processed data conforms to the standard normal distribution, and the formula for standardization processing is: Z = (X-μ) / σ, wherein X is the original data, μ is the mean of the original data, σ is the standard deviation of the original data, and Z is the standardized data.

[0047] Compared with the prior art, the present invention has the following beneficial effects:

[0048] Through the data acquisition module, this system can collect various performance parameters, environmental parameters and loom configuration parameters of the loom in real time and comprehensively during operation. Through comprehensive data collection methods, it is ensured that the system can accurately reflect the actual operating status of the loom, providing a solid foundation for subsequent data analysis, performance prediction and decision support. The system integrates the performance prediction model construction module and the prediction data generation module, which can build and update the performance prediction model based on historical data and real-time data, so as to accurately predict the change trend of the loom's future performance parameters. This intelligent performance prediction capability enables enterprises to detect potential faults or performance degradation in advance, perform preventive maintenance in time, and effectively reduce the risk of production interruption and maintenance costs.

[0049] Through the performance monitoring module and data comparison and analysis module, the system can monitor the performance of the loom in real time and accurately, and automatically compare and analyze the real-time performance parameters with the predicted performance parameters. Once a deviation is detected, the deviation detection module will immediately mark the deviation point and trigger the alarm mechanism. Through accurate monitoring and rapid response mechanisms, it ensures that operation and maintenance personnel can quickly locate problems and take effective countermeasures, greatly improving production efficiency and the timeliness of equipment maintenance.

[0050] When the system detects a deviation point, the alarm module will send detailed alarm information to the operation and maintenance personnel, including the specific value of the deviation, the corresponding performance indicator, the time of occurrence, and the recommended response measures. The detailed alarm information not only helps the operation and maintenance personnel to quickly understand the problem, but also provides them with targeted solution suggestions, further improving the efficiency and accuracy of equipment maintenance. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1 This is a working principle diagram of a loom data intelligent collection system according to the present invention;

[0052] Figure 2 Flowchart for building performance prediction model;

[0053] Figure 3 This is the working principle diagram of the deviation detection module. DETAILED DESCRIPTION

[0054] 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.

[0055] See also Figure 1-3 The present invention provides a technical solution: a loom data intelligent acquisition system, the system comprising:

[0056] Data acquisition module: This module is designed in a microservice mode, and a matching front-end adapter is developed for different types of looms. Each front-end adapter is responsible for establishing a stable communication connection with its corresponding loom controller, including but not limited to the customization of the communication protocol, the definition of the data format, the disconnection and reconnection mechanism, and the status monitoring function. In this way, it is ensured that all kinds of performance parameters (such as speed, tension, temperature, etc.), environmental parameters (such as humidity, temperature, dust concentration, etc.) and loom configuration parameters (such as machine model, yarn type, weaving process, etc.) of the loom during operation can be collected in real time and comprehensively.

[0057] Data preprocessing module: This module receives the raw data from the data acquisition module and performs a series of data cleaning tasks, including removing invalid data, filling missing values, correcting erroneous data, etc. Subsequently, the cleaned data is formatted and standardized to ensure data consistency and comparability. Finally, key features are extracted from the processed data to form a feature data set, which provides a basis for subsequent analysis.

[0058] Performance prediction model building module: Based on the feature data set, combined with environmental parameters and loom configuration parameters, a performance prediction model is built using machine learning or deep learning algorithms. The model can learn the complex relationship between loom performance parameters and the environment and configuration, thereby predicting the change trend of loom performance parameters in the future.

[0059] Real-time data acquisition module: This module is responsible for acquiring the current operating status data and environmental condition data of the loom in real time to ensure the timeliness and accuracy of the data.

[0060] Prediction data generation module: inputs the data obtained by the real-time data acquisition module into the performance prediction model to generate predicted performance parameters and provide a benchmark for performance monitoring and comparative analysis.

[0061] Performance monitoring module: Real-time performance monitoring of the loom, obtaining actual performance parameter data, and using statistical methods or machine learning algorithms to detect anomalies in the original monitoring data, timely discovering and eliminating or correcting outliers to ensure the accuracy of the monitoring data.

[0062] Data comparison and analysis module: The real-time performance parameter data obtained by the performance monitoring module and the predicted performance parameter data output by the prediction data generation module are plotted into time-varying curves, and the timestamps are aligned. The two curves are placed in the same coordinate system for intuitive comparison, so that operation and maintenance personnel can quickly identify performance differences.

[0063] Deviation detection module: Set a reasonable performance deviation threshold, use an algorithm to determine whether each corresponding point on the two curves in the data comparison and analysis module has a deviation exceeding the threshold, and mark the deviation point to provide a basis for subsequent alarm and fault handling.

[0064] Alarm module: When the deviation detection module detects a deviation point, the alarm mechanism is triggered. The alarm module is responsible for sending detailed alarm information to the operation and maintenance personnel, including the specific value of the deviation, the corresponding performance indicator, the time of occurrence, and the recommended response measures based on historical data and expert experience. The alarm information can be conveyed through SMS, email, APP push and other methods to ensure that the operation and maintenance personnel can respond and handle abnormal situations in a timely manner.

[0065] The present invention will be further described below in conjunction with Examples 1 to 4:

[0066] Embodiment 1:

[0067] The following will describe in detail the service registration unit, automation management unit and log management unit in the data acquisition module, the specific implementation method and the problems they solve, combined with a specific example:

[0068] The service registration unit uses a distributed service registration and discovery mechanism, such as using open source frameworks such as ZooKeeper, Consul, or Eureka. Each front-end adapter, as a microservice instance, registers with the service registration unit at startup, providing its service address, status information, and metadata (such as supported loom types, communication protocols, etc.). The service registration unit maintains a dynamic service registry, updates and monitors the status of each microservice instance in real time.

[0069] Through the service registration unit, the system can dynamically sense the increase and decrease of front-end adapters and state changes, and realize automatic discovery and load balancing of services. This solves the problem of poor flexibility and insufficient scalability caused by hard-coded service addresses in traditional data collection methods, allowing the system to easily cope with the increase or change of loom types, and improving the maintainability and scalability of the system.

[0070] The automated management unit is built on containerization technology (such as Docker) and container orchestration tools (such as Kubernetes). It is responsible for the entire life cycle management of microservices, including deployment, start, stop, restart, and version update of microservices. By defining service templates and configuration files, the automated management unit can automatically create and destroy containers according to demand, realizing elastic scaling of microservices.

[0071] The application of the automated management unit greatly simplifies the management process of microservices, reduces manual intervention, and improves operation and maintenance efficiency. At the same time, the isolation between microservices is achieved through containerization technology, which enhances the stability and security of the system. In addition, the automated elastic scaling capability enables the system to dynamically adjust resource allocation according to changes in the number of looms and the amount of data collected, thereby optimizing resource utilization.

[0072] The log management unit adopts a centralized log collection and management solution, such as using ELK Stack (Elasticsearch, Logstash, Kibana) or Fluentd and other tools. The log information and operation data generated by each front-end adapter during operation are sent to the log management unit in real time, and are stored in the log database after cleaning and sorting. The log management unit provides functions such as log query, analysis, and alarm, and supports operation and maintenance personnel to monitor the operation status and abnormal information of the service in real time through a visual interface. The establishment of the log management unit solves the problem of scattered and difficult to track logs in traditional data collection systems. Through centralized log management, operation and maintenance personnel can quickly locate and solve faults and improve the maintainability of the system. At the same time, the analysis of log data can also provide strong support for system optimization and performance tuning, helping operation and maintenance personnel to better understand system behavior and predict potential problems.

[0073] Embodiment 2:

[0074] This embodiment is used to describe in detail the implementation of the collection and deviation detection module of performance parameters, environmental parameters and loom configuration parameters in the loom data intelligent collection system.

[0075] Collection of performance parameters, environmental parameters and loom configuration parameters:

[0076] Performance parameters: The system collects key performance indicators of the loom in real time, including but not limited to loom speed (measured in fabric length or revolutions per minute), fabric tension (reflecting the tightness of the fabric during the weaving process), yarn consumption rate (calculating the amount of yarn used per unit time), motor load (monitoring the load of the motor driving the loom), loom temperature (reflecting the temperature condition of the loom when it is working to prevent overheating) and failure rate (statisticing the frequency of loom failures).

[0077] Environmental parameters: The system also monitors environmental conditions that affect loom operation, including workshop temperature, humidity (both of which affect the physical properties of yarn and the operating efficiency of the loom) and air cleanliness (affecting the quality of the fabric and the maintenance cycle of the loom).

[0078] Loom configuration parameters: In order to have a more comprehensive understanding of the operating status of the loom, the system also collects basic configuration information of the loom, such as loom model (reflecting the technical specifications and performance characteristics of the loom), yarn type (affecting the feel and strength of the fabric), fabric specifications (determining the final use of the fabric), and the degree of wear of loom components (assessing the maintenance needs and life of the loom).

[0079] Implementation of the deviation detection module:

[0080] Threshold setting unit: This unit sets reasonable deviation thresholds for each performance indicator based on historical performance monitoring data (providing a baseline for performance indicators), loom manufacturer requirements (ensuring that the loom operates within a safe range), and the experience of maintenance personnel (reflecting best practices in actual operation). These thresholds take into account the normal fluctuation range of the loom and also reserve enough space to deal with abnormal situations.

[0081] Deviation detection unit: This unit is responsible for traversing the aligned real-time performance monitoring data and prediction data in the data comparison and analysis module. For each pair of corresponding data points, it calculates the deviation value between the two and compares the deviation with the threshold set in the threshold setting unit. If the deviation exceeds the threshold, the point is marked as a deviation point, indicating that a certain performance indicator of the loom has an abnormality or deviates from the expected trend.

[0082] Embodiment 3:

[0083] In the loom data intelligent collection system, the construction of the performance prediction model is the key link to predict the future performance status of the loom. This system uses the long short-term memory network (LSTM) algorithm to build a performance prediction model to capture the long-term dependencies in the time series data and accurately predict the performance parameters of the loom. The following is the specific implementation method of model construction and training:

[0084] Performance prediction model construction

[0085] A model structure design:

[0086] Input layer: Receives preprocessed feature data sets, which include historical performance parameters of the loom (such as speed, tension, yarn consumption rate, etc.), environmental parameters (such as workshop temperature, humidity, etc.), and loom configuration parameters (such as loom model, yarn type, etc.).

[0087] LSTM hidden layer: contains at least one layer of LSTM units to capture long-term dependencies in time series data. LSTM units process and store information through their special gating mechanisms (forget gate, input gate, and output gate), thereby effectively learning the dynamic changes of time series data.

[0088] Output layer: Outputs predicted performance parameter values, such as loom speed, fabric tension, etc. in the future.

[0089] B model training steps:

[0090] Step 1: Data partitioning: The preprocessed feature data set is randomly divided into training set, validation set and test set. The training set is used for model training, the validation set is used for model verification and parameter adjustment, and the test set is used to evaluate the final model performance. Usually, the training set accounts for 70% of the total data, the validation set accounts for 15%, and the test set accounts for 15%.

[0091] Step 2: Model construction and initialization: Build the LSTM neural network model, set the model's input layer, LSTM layer, fully connected layer, and output layer. Initialize the model parameters, including the weights and biases of the LSTM unit, and the parameters of the fully connected layer.

[0092] Step 3: Model iterative training: Use the training set to iteratively train the LSTM model. The gradient of the loss function is calculated through the back propagation algorithm, and the model parameters are updated using the gradient descent method. The parameter update formula is:

[0093]

[0094] Among them, θ represents the model parameters, including the weights and biases of the convolutional layers, η is the learning rate, which is used to control the step size of parameter updates, is the gradient of the loss function with respect to the parameter θ;

[0095] The loss function adopts the cross entropy loss function, and its calculation formula is:

[0096]

[0097] Where N is the number of samples, y i is the true label of the i-th sample, p i is the predicted probability of the i-th sample;

[0098] Step 4: Model verification and parameter adjustment:

[0099] After each round of training, the performance of the model is evaluated using the validation set. The model's hyperparameters, including learning rate, number of units in the LSTM layer, number of training rounds, etc., are adjusted based on the performance on the validation set to minimize the loss on the validation set.

[0100] The early stopping method is used to prevent the model from overfitting. When the loss on the validation set no longer decreases within a certain number of rounds (such as 10 rounds), the training process is stopped and the model parameters with the best performance on the validation set are saved as the final model parameters.

[0101] Step 5: Model performance evaluation:

[0102] After model training and parameter adjustment, the test set is used to evaluate the performance of the final model. The performance of the model is quantified by calculating indicators such as prediction accuracy and mean square error (MSE) on the test set. The prediction accuracy reflects the accuracy of the model's prediction results, and the mean square error reflects the degree of difference between the model's prediction results and the true value.

[0103] Embodiment 4:

[0104] The data preprocessing module of the present invention includes a data cleaning unit, a data format unification unit and a data standardization unit. The implementation methods of these units and the problems they solve are described in detail below.

[0105] The data cleaning unit is responsible for processing missing values ​​and outliers in the collected data to ensure the integrity and accuracy of the data.

[0106] ① Missing value processing:

[0107] Numerical features: For missing values ​​in numerical features, the median of the non-missing values ​​of the feature is used to fill in the missing values. The median can reflect the central trend of the data and is not affected by extreme values, so it is a suitable choice for filling in missing values ​​of numerical types.

[0108] Categorical features: For missing values ​​in categorical features, the mode of the non-missing values ​​of the feature is used to fill in the missing values. The mode represents the value that appears most frequently in the data and can reflect the general situation of the data.

[0109] Time series data: For missing values ​​in time series data, the average value of the adjacent values ​​is used to fill in. This method can maintain the continuity and smoothness of time series data.

[0110] Features with obvious business meaning: For features with obvious business meaning and few missing values, fill them manually based on business logic. For example, if a feature represents the status of a specific component of a loom and the missing value corresponds to a known maintenance event, the missing value can be inferred and filled based on the data before and after the maintenance.

[0111] ②Outlier processing:

[0112] Numerical features: Use the 3σ principle to identify and handle outliers. First, calculate the mean and standard deviation of the feature, and then consider values ​​that exceed ±3 times the standard deviation of the mean as outliers. For the identified outliers, you can choose to delete, replace (such as replacing with the mean, median, or specific value), or retain but perform special processing (such as marking as anomaly) according to the specific situation.

[0113] The data format unification unit is responsible for converting data from different sources into a unified format for subsequent data processing and model training. This includes unifying data in different time formats into a standard timestamp format, converting numerical data in different units to the same unit, and converting text data into numerical encoding (such as one-hot encoding or embedded encoding). By unifying the data format, inconsistencies between data can be eliminated and the efficiency and accuracy of data processing can be improved.

[0114] The data standardization unit is responsible for standardizing the data to eliminate the dimensional differences in the distribution range of different data. Standardization makes data with different characteristics comparable, which helps to improve the training effect and prediction accuracy of the model.

[0115] Standardization method: The Z-score method is used for data standardization. For each numerical data, its mean is subtracted and divided by its standard deviation so that the processed data conforms to the standard normal distribution. The formula for standardization is: Z = (X-μ) / σ, where X is the original data, μ is the mean of the original data, σ is the standard deviation of the original data, and Z is the standardized data.

[0116] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0117] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A loom data intelligent collection system, characterized in that: The system comprises: The data acquisition module is used to collect various performance parameters, environmental parameters and loom configuration parameters of the loom in real time during operation; the data acquisition module adopts a microservice mode to build multiple front-end adapters for data reception. In the data acquisition module, each type of loom corresponds to a matching front-end adapter; the front-end adapter is responsible for establishing a communication mechanism with various loom controllers, including the formulation of communication protocols, the definition of data formats, disconnection processing and status monitoring; The data preprocessing module is connected to the data acquisition module and is used to clean, unify and standardize the collected data, and then extract features to form a feature data set; The performance prediction model building module builds a performance prediction model for predicting the future performance parameter change trend of the loom based on the feature data set, combined with environmental parameters and loom configuration parameters; A real-time data acquisition module is used to acquire the current operating status data and environmental condition data of the loom in real time; A prediction data generation module inputs the data acquired by the real-time data acquisition module into the performance prediction model to generate prediction performance parameters; The performance monitoring module is used to monitor the performance of the loom in real time, obtain the actual performance parameter data, and perform anomaly detection on the original monitoring data, and remove or correct abnormal values; The data comparison and analysis module plots the real-time performance parameter data obtained by the performance monitoring module and the predicted performance parameter data output by the predicted data generation module into time-varying curves, aligns the timestamps, and places the two curves in the same coordinate system for comparison; The deviation detection module sets the performance deviation threshold, determines whether each corresponding point on the two curves in the data comparison and analysis module has a deviation exceeding the threshold, and marks the deviation point; The alarm module is connected to the deviation detection module. When a deviation point is detected, the alarm mechanism is triggered and an alarm message is sent to the operation and maintenance personnel. The alarm message includes the specific value of the deviation, the corresponding performance indicator, the time of occurrence and the recommended response measures.

2. The loom data intelligent collection system according to claim 1, characterized in that: The performance parameters include loom speed, fabric tension, yarn consumption rate, motor load, loom temperature, and failure rate; the environmental parameters include workshop temperature, humidity, and air cleanliness; the loom configuration parameters include loom model, yarn type, fabric specifications, and degree of wear of loom components.

3. The loom data intelligent collection system according to claim 1, characterized in that: The data acquisition module also includes a service registration unit, an automation management unit and a log management unit; the service registration unit provides registration and status monitoring functions for microservice instances; the automation management unit is responsible for the deployment, start, stop and restart operations of microservices; The log management unit collects the service's operating data and log information for data monitoring and management.

4. The loom data intelligent collection system according to claim 1, characterized in that: The deviation detection module further comprises: A threshold setting unit is used to set reasonable deviation thresholds for various performance indicators based on historical performance monitoring data, loom requirements, and operation and maintenance experience; The deviation detection unit is used to traverse the aligned data, calculate the deviation value between the real-time performance monitoring data and the predicted data, and mark the deviation points that exceed the threshold.

5. The loom data intelligent collection system according to claim 1, characterized in that: The performance prediction model construction module adopts the long short-term memory network LSTM algorithm to construct a performance prediction model. The network structure of the model includes an input layer, at least one LSTM hidden layer and an output layer, wherein the input layer receives a feature data set, the LSTM hidden layer is used to capture long-term dependencies in time series data, and the output layer outputs predicted performance parameter values.

6. The loom data intelligent collection system according to claim 5, characterized in that: The steps to train a performance prediction model include: Step 1: Randomly divide the preprocessed feature data set into training set, validation set and test set. The training set is used for model training, the validation set is used for model verification and parameter adjustment, and the test set is used to evaluate the final model performance. Step 2: Build an LSTM neural network model, set the model's input layer, LSTM layer, fully connected layer, and output layer, and initialize the model parameters; Step 3: Use the training set to iteratively train the LSTM model, calculate the gradient of the loss function through the back propagation algorithm, and use the gradient descent method to update the model parameters. The parameter update formula is: Among them, θ represents the model parameters, including the weights and biases of the convolutional layers, η is the learning rate, which is used to control the step size of parameter updates, is the gradient of the loss function with respect to the parameter θ; The loss function adopts the cross entropy loss function, and its calculation formula is: Where N is the number of samples, y i is the true label of the i-th sample, p i is the predicted probability of the i-th sample; Step 4: After each round of training, use the validation set to evaluate the performance of the model. Adjust the model's hyperparameters, including the learning rate, the number of units in the LSTM layer, and the number of training rounds, based on the performance on the validation set to minimize the loss on the validation set. Step 5: After model training and parameter adjustment are completed, use the test set to evaluate the performance of the final model and quantify the performance of the model by calculating the prediction accuracy and mean square error indicators on the test set.

7. The loom data intelligent collection system according to claim 6, characterized in that: In step 4, the early stopping method is used to prevent the model from overfitting. When the loss on the validation set no longer decreases within a certain number of rounds, the training process is stopped, and the model parameters with the best performance on the validation set are saved as the final model parameters.

8. The loom data intelligent collection system according to claim 1, characterized in that: The data preprocessing module further comprises: Data cleaning unit, used to process missing values ​​and outliers in collected data; Data format unification unit, used to convert data from different sources into a unified format; The data standardization unit is used to standardize the data and eliminate the dimensional differences in the distribution range of different data.

9. The loom data intelligent collection system according to claim 8, characterized in that: The method for the data cleaning unit to process missing values ​​includes: For numerical features, the median of the non-missing values ​​of the feature is used to fill the missing values; For categorical features, the missing values ​​are filled with the mode of the non-missing values ​​of the feature; For missing values ​​in time series data, the average value of the adjacent values ​​is used to fill them; For features with obvious business meaning and few missing values, they are manually filled according to business logic; The method for processing abnormal values ​​by the data cleaning unit includes: For numerical features, the 3σ principle is used to identify and process outliers; specifically, the mean and standard deviation of the feature are calculated, and values ​​exceeding the mean ± 3 times the standard deviation are considered outliers.

10. The loom data intelligent collection system according to claim 8, characterized in that: The method for the data standardization unit to perform standardization processing on the data is: using the Z-score method to subtract the mean of each numerical data and divide it by its standard deviation, so that the processed data conforms to the standard normal distribution, and the formula for standardization processing is: Z = (X-μ) / σ, wherein X is the original data, μ is the mean of the original data, σ is the standard deviation of the original data, and Z is the standardized data.

Images