CN118311914B - A production line data acquisition control method and system for intelligent workshop - Google Patents

Abstract

The invention belongs to the technical field of intelligent control and processing of workshop production lines, and particularly relates to a production line data acquisition control method and system of an intelligent workshop. The method comprises the following steps: collecting equipment state multisource data; selecting the equipment state multisource characteristic data from the equipment state multisource data to generate the equipment state multisource characteristic data; carrying out time sequence data integration on the equipment state multi-source characteristic data to generate equipment state time sequence multi-source characteristic data; acquiring equipment control parameters of a workshop production line; establishing a device dynamic driving data model based on device control parameters and device state time sequence multi-source characteristic data of a workshop production line; and analyzing and optimizing the self-adaptive control strategy of the equipment based on the dynamic driving data model of the equipment, and generating an optimized self-adaptive control strategy of the equipment. The invention optimizes the control strategy of the production line by accurately and intelligently collecting the production line data of the workshop.

Description

A production line data acquisition control method and system for intelligent workshop

Technical Field

The present invention relates to the field of intelligent control and processing technology for workshop production lines, and in particular to a production line data acquisition control method and system for an intelligent workshop.

Background Art

At present, the manufacturing industry is developing rapidly, and the data acquisition and control technology of intelligent workshop production lines has been widely used in the field of industrial automation. By introducing advanced sensor technology, data acquisition and processing technology, and intelligent control technology, comprehensive perception and intelligent control of the production process can be achieved, thereby greatly improving production efficiency, reducing production costs and improving product quality, effectively managing various variables and risks in the production process, improving production efficiency and product quality, and helping workshop production lines to achieve reasonable resource allocation and effective cost control. However, the traditional workshop production line data acquisition and control method mainly relies on manual operation and a single data acquisition method, and there are problems such as information islands, incomplete data collection, poor real-time performance, and difficulty in effective integration and analysis. The existing workshop intelligent technology attempts to apply multi-source heterogeneous data integration processing and analysis to the field of intelligent manufacturing, but there are still problems such as low data processing efficiency, insufficient analysis accuracy, and inflexible control strategies.

Summary of the invention

Based on this, the present invention provides a production line data acquisition control method and system for an intelligent workshop to solve at least one of the above technical problems.

To achieve the above purpose, a production line data acquisition control method for an intelligent workshop comprises the following steps:

Step S1: collecting multi-source heterogeneous data of equipment status of a workshop production line according to a multimodal sensor to generate multi-source heterogeneous data of equipment status; performing multi-source heterogeneous data integration processing on the multi-source heterogeneous data of equipment status to generate multi-source data of equipment status;

Step S2: performing multi-source data classification processing on the multi-source data of the device status to generate single-source classified data of the device status; performing single-source subset distribution feature node analysis of the device status based on the single-source classified data of the device status to generate single-source subset distribution feature node data; performing multi-source feature data selection on the multi-source data of the device status according to the single-source subset distribution feature node data to generate multi-source feature data of the device status;

Step S3: integrating the multi-source characteristic data of the device status in time series to generate the multi-source characteristic data of the device status in time series; identifying the time series trend of abnormal device status based on the multi-source characteristic data of the device status in time series to generate the time series trend data of abnormal device status;

Step S4: Acquire equipment control parameters of the workshop production line; perform dynamic drive mapping relationship analysis of equipment control parameters and equipment status based on the equipment control parameters of the workshop production line and the equipment status time series multi-source feature data to generate an equipment dynamic drive data model; perform equipment adaptive control strategy analysis based on abnormal equipment status trend data and the equipment dynamic drive data model to generate an equipment adaptive control strategy;

Step S5: Perform scheduling optimization processing on the control parameters of the equipment adaptive control strategy to generate an optimized equipment adaptive control strategy, and feed back the optimized equipment adaptive control strategy to the terminal device to execute the intelligent control operation of the production line equipment.

Preferably, step S1 comprises the following steps:

Step S11: configuring the multimodal sensor as a sensor node based on a preset sensor configuration requirement to obtain sensor node configuration data, and setting a sensor network topology structure through the sensor configuration data;

Step S12: Real-time monitoring of equipment status signals of the workshop production line is performed according to the multimodal sensor to generate real-time monitoring equipment status signals;

Step S13: Perform multi-source heterogeneous data analysis on the equipment status of the workshop production line according to the real-time monitoring equipment status signal to generate multi-source heterogeneous data on the equipment status;

Step S14: performing time sequence synchronization correction on the device state multi-source heterogeneous data to generate synchronized device state multi-source heterogeneous data;

Step S15: analyzing the device status integration queue according to the sensor network topology structure, and generating the device status integration queue;

Step S16: performing multi-source heterogeneous data integration processing on the synchronous device status multi-source heterogeneous data according to the device status integration queue to generate device status multi-source data.

Preferably, step S13 comprises the following steps:

Perform sensor self-noise interference signal analysis on multimodal sensors to generate sensor self-noise interference signals;

Perform effective signal analysis on the real-time monitoring device status signal based on the sensor self-noise interference signal to generate an effective monitoring device status signal;

According to the sensor network topology, the signal source of the effective monitoring device status signal is identified to generate an identification measuring device status signal;

The device status signal of the identification measurement device is converted into multi-source heterogeneous data of the equipment status of the workshop production line to generate multi-source heterogeneous data of the equipment status.

Preferably, step S2 comprises the following steps:

Step S21: performing multi-source data classification processing on the multi-source data of the device status to generate single-source classified data of the device status;

Step S22: selecting the single-source classified data of device status from the multi-source data of device status one by one as a single analysis variable, performing a single-source subset impact similarity evaluation of the device status on the single-source classified data of device status corresponding to the single analysis variable, and generating single-source subset impact similarity data;

Step S23: performing single-source subset distribution feature node analysis based on the single-source subset impact similarity data to generate single-source subset distribution feature node data;

Step S24: selecting device status multi-source feature data from the device status multi-source data according to the single-source subset distribution feature node data to generate the device status multi-source feature data.

Preferably, step S23 includes the following steps:

Perform single-source subset distribution impact probability evaluation based on single-source subset impact similarity data to generate single-source subset distribution impact probability data;

Perform KL divergence calculation on the single-source subset distribution impact probability data to generate single-source subset distribution impact KL divergence data;

According to the KL divergence data of single-source subset distribution influence, single-source subset distribution feature node analysis is performed to generate single-source subset distribution feature node data.

Preferably, step S3 comprises the following steps:

Step S31: integrating the time series data of the multi-source characteristic data of the device status to generate the time series multi-source characteristic data of the device status;

Step S32: performing device state time series trend analysis on the device state time series multi-source feature data based on a preset long short-term memory neural network algorithm to generate device state time series trend data;

Step S33: selecting the equipment status single-source classification data corresponding to the equipment status time series trend data to perform equipment status time series trend evaluation binary tree design, so as to establish an equipment status time series trend evaluation binary tree;

Step S34: identifying abnormal device state timing trend on the device state timing trend data according to the device state timing trend evaluation binary tree, and generating abnormal device state timing trend data.

Preferably, step S34 includes the following steps:

Transmitting the equipment status timing trend data to the equipment status timing trend evaluation binary tree for binary tree recursive evaluation to generate binary tree recursive evaluation data;

Perform cluster analysis on the binary tree recursive evaluation data to generate clustered binary tree recursive evaluation data; perform outlier identification based on the clustered binary tree recursive evaluation data to generate binary tree recursive evaluation outlier data;

The abnormal nodes of the device status time series trend data are identified based on the outlier data recursively evaluated by the binary tree to generate abnormal device status time series trend data.

Preferably, step S4 comprises the following steps:

Step S41: Obtaining equipment control parameters of the workshop production line;

Step S42: performing device driver association analysis based on the equipment control parameters of the workshop production line and the multi-source characteristic data of the equipment status time series to generate device driver association data;

Step S43: analyzing the dynamic drive mapping relationship of the device control parameters and the device status according to the device drive association data to generate a device dynamic drive data model;

Step S44: performing a preferred device driving characteristic data analysis on the device dynamic driving data model based on the abnormal device status trend data to generate preferred device driving characteristic data;

Step S45: Perform device adaptive control strategy analysis based on the preferred device driving characteristic data to generate a device adaptive control strategy.

Preferably, step S51 includes the following steps:

Step S51: performing simulation equipment scheduling driving efficiency analysis according to the equipment adaptive control strategy and the equipment dynamic driving data model, and generating simulation equipment scheduling driving efficiency data;

Step S52: performing scheduling resource priority analysis according to the simulation device scheduling drive efficiency data to generate scheduling resource priority data;

Step S53: Perform scheduling optimization processing on the control parameters of the equipment adaptive control strategy according to the scheduling resource priority data, generate an optimized equipment adaptive control strategy, and feed back the optimized equipment adaptive control strategy to the terminal device to execute the intelligent control operation of the production line equipment.

This specification provides a production line data acquisition control system for an intelligent workshop, which is used to execute the production line data acquisition control method for the intelligent workshop as described above. The production line data acquisition control system for the intelligent workshop includes:

The equipment status multi-source data acquisition module is used to collect multi-source heterogeneous data of the equipment status of the workshop production line based on multi-modal sensors to generate multi-source heterogeneous data of the equipment status; perform multi-source heterogeneous data integration processing on the multi-source heterogeneous data of the equipment status to generate multi-source data of the equipment status;

The device status multi-source feature analysis module is used to perform multi-source data classification processing on the device status multi-source data to generate the device status single-source classification data; perform single-source subset distribution feature node analysis on the device status based on the device status single-source classification data to generate single-source subset distribution feature node data; perform device status multi-source feature data selection on the device status multi-source data based on the single-source subset distribution feature node data to generate the device status multi-source feature data;

The equipment status time series multi-source feature analysis module is used to integrate the time series data of the equipment status multi-source feature data to generate the equipment status time series multi-source feature data; based on the equipment status time series multi-source feature data, the abnormal equipment status time series trend is identified to generate the abnormal equipment status time series trend data;

The equipment adaptive control strategy analysis module is used to obtain the equipment control parameters of the workshop production line; based on the equipment control parameters of the workshop production line and the multi-source characteristic data of the equipment status time series, the dynamic drive mapping relationship analysis of the equipment control parameters and the equipment status is performed to generate the equipment dynamic drive data model; based on the abnormal equipment status trend data and the equipment dynamic drive data model, the equipment adaptive control strategy analysis is performed to generate the equipment adaptive control strategy;

The equipment adaptive control strategy optimization module is used to schedule and optimize the control parameters of the equipment adaptive control strategy, generate an optimized equipment adaptive control strategy, and feed back the optimized equipment adaptive control strategy to the terminal device to execute the intelligent control operation of the production line equipment.

The present invention can comprehensively monitor the equipment status of the workshop production line by deploying multimodal sensors, realize real-time collection of multiple parameters such as temperature, pressure, vibration, sound, etc., provide comprehensive data support, integrate data of different types and sources through integrated processing of multi-source heterogeneous data, generate unified multi-source data of equipment status, solve the problems of information islands and data inconsistency, and the real-time data collection and processing of multimodal sensors can quickly respond to the status changes of production line equipment, improve the real-time and accuracy of data. The multi-source data of equipment status is classified and processed to generate single-source classified data of equipment status, so that data analysis is more detailed and accurate, which is helpful to identify small changes in equipment status. Through single-source subset distribution feature node analysis and feature data selection, the key feature data of equipment status is extracted. The feature data selection process can effectively reduce the redundant information of data, reduce the complexity and calculation burden of data processing, and improve the operation efficiency of the system. The multi-source characteristic data of the equipment status is integrated to generate the multi-source characteristic data, so that the data has a time dimension and can reflect the dynamic changes of the equipment status. The abnormal equipment status time series trend identification based on the multi-source characteristic data can detect the abnormal trend and potential failure of the equipment in advance, improve the safety and reliability of equipment operation, and provide data support for the predictive maintenance of the equipment, reduce the production downtime caused by sudden failures, and reduce maintenance costs. By analyzing the dynamic driving mapping relationship between the equipment control parameters and the multi-source characteristic data of the equipment status time series, the equipment dynamic driving data model is generated, which can accurately describe the relationship between the equipment status and the control parameters. Based on the abnormal equipment status trend data and the equipment dynamic driving data model, the equipment adaptive control strategy is analyzed and generated to achieve adaptive control of the equipment, improve the automation level and operation efficiency of the production line, and the adaptive control strategy can be flexibly adjusted according to the real-time changes of the equipment status, improve the stability of equipment operation and the flexibility of the production line, and prevent abnormal conditions from occurring during the operation of the production line equipment. The control parameters of the equipment adaptive control strategy are optimized and dispatched, and the corresponding optimal equipment control parameters are analyzed based on the equipment status, such as allocating equipment control scheduling resources based on the optimal operating range of the equipment, generating an optimized adaptive control strategy, ensuring the efficiency and accuracy of the control strategy in execution, and feeding back the optimized adaptive control strategy to the terminal device, executing intelligent control operations, realizing the automation and intelligent control of production line equipment, reducing manual intervention, and improving production efficiency. Intelligent control operations can effectively improve the operating efficiency and product quality of the production line. Therefore, the production line data acquisition and control method of the intelligent workshop of the present invention, by introducing technologies such as multi-modal sensors, multi-source heterogeneous data integration processing, multi-source data classification processing, time series data integration, and adaptive control strategies, aims to solve the problems of incomplete data acquisition, low data processing efficiency, and inflexible intelligent control in the prior art, realize comprehensive monitoring and intelligent control of the equipment status of the production line, and improve the intelligence level and production efficiency of the production line.

The beneficial effect of the present application is that the production line data acquisition and control method of the intelligent workshop of the present invention significantly improves the intelligent control level and operation efficiency of the production line by comprehensively using multimodal sensor data acquisition, classification processing and feature analysis of multi-source data, integration of time series data and abnormal trend identification, and dynamic mapping and optimization of adaptive control strategies. Through efficient data acquisition and integrated processing, the equipment status can be fully monitored and accurately analyzed to ensure real-time update and high consistency of data. The use of long short-term memory networks and binary tree structures for trend analysis and abnormal state identification greatly improves the accuracy of fault prediction and early warning capabilities, thereby reducing the risk of equipment failure and downtime. Through simulation and optimization analysis, the optimal allocation of resources and fine adjustment of control strategies are achieved, the response speed and adaptability of the production line are enhanced, and the efficiency and stability of the production process are ensured.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG1 is a schematic diagram of the steps of a production line data acquisition control method for an intelligent workshop according to the present invention;

FIG2 is a schematic diagram of a detailed implementation process of step S2 in FIG1 ;

FIG3 is a schematic diagram of a detailed implementation process of step S3 in FIG1 ;

The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings.

DETAILED DESCRIPTION

The technical method of the present invention is described clearly and completely below in conjunction with the accompanying drawings. Obviously, the described embodiments are 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 technicians in this field without creative work are within the scope of protection of the present invention.

In addition, the accompanying drawings are only schematic illustrations of the present invention and are not necessarily drawn to scale. The same reference numerals in the figures represent the same or similar parts, and their repeated description will be omitted. Some of the block diagrams shown in the accompanying drawings are functional entities and do not necessarily correspond to physically or logically independent entities. The functional entities can be implemented in software form, or implemented in one or more hardware modules or integrated circuits, or implemented in different networks and/or processor methods and/or microcontroller methods.

It should be understood that, although the terms "first", "second", etc. may be used herein to describe various units, these units should not be limited by these terms. These terms are used only to distinguish one unit from another unit. For example, without departing from the scope of the exemplary embodiments, the first unit may be referred to as the second unit, and similarly the second unit may be referred to as the first unit. The term "and/or" used herein includes any and all combinations of one or more of the listed associated items.

To achieve the above object, referring to FIG. 1 to FIG. 3 , the present invention provides a production line data acquisition control method for an intelligent workshop, comprising the following steps:

Step S1: collecting multi-source heterogeneous data of equipment status of a workshop production line according to a multimodal sensor to generate multi-source heterogeneous data of equipment status; performing multi-source heterogeneous data integration processing on the multi-source heterogeneous data of equipment status to generate multi-source data of equipment status;

Step S2: performing multi-source data classification processing on the multi-source data of the device status to generate single-source classified data of the device status; performing single-source subset distribution feature node analysis of the device status based on the single-source classified data of the device status to generate single-source subset distribution feature node data; performing multi-source feature data selection on the multi-source data of the device status according to the single-source subset distribution feature node data to generate multi-source feature data of the device status;

Step S3: integrating the multi-source characteristic data of the device status in time series to generate the multi-source characteristic data of the device status in time series; identifying the time series trend of abnormal device status based on the multi-source characteristic data of the device status in time series to generate the time series trend data of abnormal device status;

Step S4: Acquire equipment control parameters of the workshop production line; perform dynamic drive mapping relationship analysis of equipment control parameters and equipment status based on the equipment control parameters of the workshop production line and the equipment status time series multi-source feature data to generate an equipment dynamic drive data model; perform equipment adaptive control strategy analysis based on abnormal equipment status trend data and the equipment dynamic drive data model to generate an equipment adaptive control strategy;

Step S5: Perform scheduling optimization processing on the control parameters of the equipment adaptive control strategy to generate an optimized equipment adaptive control strategy, and feed back the optimized equipment adaptive control strategy to the terminal device to execute the intelligent control operation of the production line equipment.

The present invention can comprehensively monitor the equipment status of the workshop production line by deploying multimodal sensors, realize real-time collection of multiple parameters such as temperature, pressure, vibration, sound, etc., provide comprehensive data support, integrate data of different types and sources through integrated processing of multi-source heterogeneous data, generate unified multi-source data of equipment status, solve the problems of information islands and data inconsistency, and the real-time data collection and processing of multimodal sensors can quickly respond to the status changes of production line equipment, improve the real-time and accuracy of data. The multi-source data of equipment status is classified and processed to generate single-source classified data of equipment status, so that data analysis is more detailed and accurate, which is helpful to identify small changes in equipment status. Through single-source subset distribution feature node analysis and feature data selection, the key feature data of equipment status is extracted. The feature data selection process can effectively reduce the redundant information of data, reduce the complexity and calculation burden of data processing, and improve the operation efficiency of the system. The multi-source characteristic data of the equipment status is integrated to generate the multi-source characteristic data, so that the data has a time dimension and can reflect the dynamic changes of the equipment status. The abnormal equipment status time series trend identification based on the multi-source characteristic data can detect the abnormal trend and potential failure of the equipment in advance, improve the safety and reliability of equipment operation, and provide data support for the predictive maintenance of the equipment, reduce the production downtime caused by sudden failures, and reduce maintenance costs. By analyzing the dynamic driving mapping relationship between the equipment control parameters and the multi-source characteristic data of the equipment status time series, the equipment dynamic driving data model is generated, which can accurately describe the relationship between the equipment status and the control parameters. Based on the abnormal equipment status trend data and the equipment dynamic driving data model, the equipment adaptive control strategy is analyzed and generated to achieve adaptive control of the equipment, improve the automation level and operation efficiency of the production line, and the adaptive control strategy can be flexibly adjusted according to the real-time changes of the equipment status, improve the stability of equipment operation and the flexibility of the production line, and prevent abnormal conditions from occurring during the operation of the production line equipment. The control parameters of the equipment adaptive control strategy are optimized and dispatched, and the corresponding optimal equipment control parameters are analyzed based on the equipment status, such as allocating equipment control scheduling resources based on the optimal operating range of the equipment, generating an optimized adaptive control strategy, ensuring the efficiency and accuracy of the control strategy in execution, and feeding back the optimized adaptive control strategy to the terminal device, executing intelligent control operations, realizing the automation and intelligent control of production line equipment, reducing manual intervention, and improving production efficiency. Intelligent control operations can effectively improve the operating efficiency and product quality of the production line. Therefore, the production line data acquisition and control method of the intelligent workshop of the present invention, by introducing technologies such as multi-modal sensors, multi-source heterogeneous data integration processing, multi-source data classification processing, time series data integration, and adaptive control strategies, aims to solve the problems of incomplete data acquisition, low data processing efficiency, and inflexible intelligent control in the prior art, realize comprehensive monitoring and intelligent control of the equipment status of the production line, and improve the intelligence level and production efficiency of the production line.

As an embodiment of the present invention, referring to FIG. 1, a flow chart of a method for collecting and controlling production line data of an intelligent workshop of the present invention is shown. In this embodiment, the method for collecting and controlling production line data of an intelligent workshop comprises the following steps:

Step S1: collecting multi-source heterogeneous data of equipment status of a workshop production line according to a multimodal sensor to generate multi-source heterogeneous data of equipment status; performing multi-source heterogeneous data integration processing on the multi-source heterogeneous data of equipment status to generate multi-source data of equipment status;

In an embodiment of the present invention, multimodal sensors, such as temperature sensors, vibration sensors, sound sensors, etc., are deployed at key locations of the workshop production line. According to the characteristics and requirements of the production line, the configuration requirements of the sensors are preset to ensure that the coverage and monitoring capabilities of each sensor meet the comprehensiveness and accuracy of data collection. The multimodal sensors collect equipment status signals in real time, such as temperature changes, vibration frequencies, noise levels, etc., to generate multi-source heterogeneous data of equipment status. These data have different data formats and timestamps. The collected multi-source heterogeneous data are preprocessed, including data cleaning (removing errors and invalid data), data standardization (unifying the formats and measurement units of different data), and time synchronization (ensuring the alignment of data from different sources in time). Through data fusion technology, such as database joint query, data warehouse or data lake technology, the processed multi-source heterogeneous data are integrated into unified multi-source data of equipment status, involving data mapping and pre-designed data conversion rules to ensure data consistency and integrity.

Step S2: performing multi-source data classification processing on the multi-source data of the device status to generate single-source classified data of the device status; performing single-source subset distribution feature node analysis of the device status based on the single-source classified data of the device status to generate single-source subset distribution feature node data; performing multi-source feature data selection on the multi-source data of the device status according to the single-source subset distribution feature node data to generate multi-source feature data of the device status;

In an embodiment of the present invention, the multi-source data of the equipment status are divided into different single-source classified data according to the data type. For example, all temperature-related data are classified into one category, and vibration data are classified into another category. Select representative or critical single-source classified data, such as temperature data of the equipment that has the greatest impact on production, and use statistical analysis or machine learning models (such as decision trees, factor analysis) or KL divergence analysis to analyze the selected single-source classified data to identify key distribution feature nodes in the data, which are key factors in controlling or predicting the equipment status. Based on the identified single-source subset distribution feature nodes, data related to these feature nodes are selected from the multi-source data of the equipment status, including feature engineering techniques such as feature extraction and feature selection to ensure the relevance and effectiveness of the selected features, and the selected multi-source feature data are integrated into a new feature data set to generate a multi-source feature data type for the equipment status.

Step S3: integrating the multi-source characteristic data of the device status in time series to generate the multi-source characteristic data of the device status in time series; identifying the time series trend of abnormal device status based on the multi-source characteristic data of the device status in time series to generate the time series trend data of abnormal device status;

In the embodiment of the present invention, the timestamps of all time series data collected from multi-source feature data are aligned, data processing techniques such as interpolation or resampling are used to unify the data frequency, and time series analysis techniques such as sliding window technology are used to integrate the data into a time series multi-source feature data set with a unified time frame. A time series analysis model, such as ARIMA or long short-term memory network (LSTM), is applied to perform trend analysis on the time series multi-source feature data to identify trend data of the device status. Then, cluster analysis, such as the isolation forest algorithm, is used to identify abnormal states that deviate from normal behavior patterns from the trend data.

Step S4: Acquire equipment control parameters of the workshop production line; perform dynamic drive mapping relationship analysis of equipment control parameters and equipment status based on the equipment control parameters of the workshop production line and the equipment status time series multi-source feature data to generate an equipment dynamic drive data model; perform equipment adaptive control strategy analysis based on abnormal equipment status trend data and the equipment dynamic drive data model to generate an equipment adaptive control strategy;

In an embodiment of the present invention, the operating parameters of the equipment, such as temperature setting, speed, pressure, etc., are collected in real time from the production line control system to ensure that the collected control parameters are synchronized with the equipment status data in time for further analysis. In combination with the equipment control parameters and the equipment status time series multi-source feature data, data association analysis technology (such as correlation analysis or causal inference model) is used to explore the dynamic relationship between the two, and a dynamic driving data model between the equipment status and the control parameters is established through the dynamic association relationship between the equipment control parameters and the equipment status time series multi-source feature data and system identification or deep learning algorithm. Based on the abnormal equipment status trend data and the dynamic driving data model, an adaptive control strategy is developed, and control theory and optimization algorithms (such as PID control, fuzzy logic control or reinforcement learning) are used to adjust the control parameters to respond to changes in the equipment status, and control parameters related to abnormal operation of the equipment status are eliminated to generate an adaptive control strategy for the equipment.

Step S5: Perform scheduling optimization processing on the control parameters of the equipment adaptive control strategy to generate an optimized equipment adaptive control strategy, and feed back the optimized equipment adaptive control strategy to the terminal device to execute the intelligent control operation of the production line equipment.

In the embodiment of the present invention, an optimization algorithm (such as a genetic algorithm and a gradient descent method) is used to adjust control parameters according to production scheduling requirements and equipment performance to optimize production efficiency and product quality. The control strategy is iteratively adjusted according to simulation and actual feedback. If the production equipment shows a downward trend in efficiency after the running time reaches the 4H-5H interval, the scheduling priority of the production equipment is adjusted. The control parameters of the equipment adaptive control strategy are optimized based on the scheduling priority and the scheduling resource center to generate an optimized equipment adaptive control strategy. The optimized equipment adaptive control strategy is fed back to the terminal equipment of the production line through the control system to monitor the implementation effect to ensure that the actual application effect of the control strategy is consistent with expectations, and the operation data is collected for further analysis and optimization.

Preferably, step S1 comprises the following steps:

Step S11: configuring the multimodal sensor as a sensor node based on a preset sensor configuration requirement to obtain sensor node configuration data, and setting a sensor network topology structure through the sensor configuration data;

Step S12: Real-time monitoring of equipment status signals of the workshop production line is performed according to the multimodal sensor to generate real-time monitoring equipment status signals;

Step S13: Perform multi-source heterogeneous data analysis on the equipment status of the workshop production line according to the real-time monitoring equipment status signal to generate multi-source heterogeneous data on the equipment status;

Step S14: performing time sequence synchronization correction on the device state multi-source heterogeneous data to generate synchronized device state multi-source heterogeneous data;

Step S15: analyzing the device status integration queue according to the sensor network topology structure, and generating the device status integration queue;

Step S16: performing multi-source heterogeneous data integration processing on the synchronous device status multi-source heterogeneous data according to the device status integration queue to generate device status multi-source data.

The present invention configures the sensor nodes of the multimodal sensor according to the preset sensor configuration requirements, ensures the reasonable layout of the sensor nodes, maximizes the coverage and accuracy of data collection, configures the data through the sensor nodes, sets the topology of the sensor network, optimizes the communication efficiency and data transmission stability of the sensor network, and ensures the efficiency and reliability of the data collection process. The multimodal sensor is used to monitor the device status signal in real time, ensures the instant acquisition and update of the device status information, and provides real-time data support. The multimodal sensor covers a variety of device status parameters, realizes comprehensive monitoring of the equipment status of the workshop production line, and improves the breadth and depth of monitoring. According to the real-time monitoring of the device status signal, multi-source heterogeneous data analysis is performed to generate multi-source heterogeneous data of the device status, fully utilize the data diversity of the multimodal sensor, improve the comprehensiveness of the data analysis, and improve the accuracy and precision of the device status monitoring by comprehensive analysis of data from different sources and types, providing a reliable data basis for subsequent data processing and control. Perform time synchronization correction on multi-source heterogeneous data of device status to ensure the time consistency and accuracy of data and eliminate data deviation caused by differences in different data collection times. Time synchronization correction improves the integration quality of multi-source data and provides more accurate basic data for subsequent data analysis and processing. Perform device status integration queue analysis based on the sensor network topology, generate device status integration queue, integrate multi-source heterogeneous data, and improve the organization and system of data processing. The analysis and generation of device status integration queue will help optimize the data processing process and improve the efficiency of data analysis and integration. Perform integration processing on synchronized device status multi-source heterogeneous data based on the device status integration queue, generate device status multi-source data, improve data integrity and consistency, and integrate data of different data structures into a unified data structure, which is convenient for subsequent correlation analysis of multi-source data of device status.

In the embodiment of the present invention, the specific monitoring requirements of the workshop production line are analyzed and defined, including the types of parameters to be monitored (such as temperature, pressure, vibration, etc.) and the accuracy requirements of monitoring, and suitable multimodal sensor types are selected and deployed at key locations in the workshop according to the requirements, such as near the machine, above the conveyor belt, etc., to obtain sensor node configuration data, and the topological structure of the sensor network is designed based on the corresponding distributed network nodes in the sensor node configuration data to ensure the efficiency and stability of data transmission, as well as the corresponding data source of data collection in the system. The multimodal sensor is started to collect the operating status signals of the workshop equipment in real time, such as temperature signals, vibration frequency signals, etc., and the collected signals are preliminarily processed, including amplification, filtering and digitization, to facilitate subsequent analysis and transmission, and generate real-time monitoring equipment status signals. Data analysis and conversion are performed according to the sensor type and the real-time monitoring equipment status signal to be organized into the corresponding structured data format, and the data is classified and labeled to generate multi-source heterogeneous data of equipment status. Perform timestamp processing on multi-source heterogeneous data to ensure that all data has accurate time tags. Use time synchronization algorithms (such as linear interpolation and dynamic time warping (DTW)) to align data at different time intervals to generate synchronized device state multi-source heterogeneous data. Verify the synchronized data to ensure data integrity and consistency and eliminate data errors caused by time differences. According to the distributed network acquisition nodes in the sensor network topology and the corresponding sources of multi-source heterogeneous data of device status, build a device status integrated queue, queue the data according to certain rules and priorities, and use data analysis algorithms such as FIFO (first in, first out) and LIFO (last in, first out) to analyze the device status integrated queue to ensure the orderliness and efficiency of data processing. The synchronous device status multi-source heterogeneous data are integrated according to the device status integration queue, and the queue nodes of the device status integration queue corresponding to the synchronous device status multi-source heterogeneous data are stored. The integrated data are stored and managed using database technology (such as SQL or NoSQL database), such as data cleaning (removing errors and invalid data), data compression (reducing storage space), and the integrated unified data structure is converted to integrate the multi-source heterogeneous data and generate device status multi-source data.

Preferably, step S13 comprises the following steps:

Perform sensor self-noise interference signal analysis on multimodal sensors to generate sensor self-noise interference signals;

Perform effective signal analysis on the real-time monitoring device status signal based on the sensor self-noise interference signal to generate an effective monitoring device status signal;

According to the sensor network topology, the signal source of the effective monitoring device status signal is identified to generate an identification measuring device status signal;

The device status signal of the identification measurement device is converted into multi-source heterogeneous data of the equipment status of the workshop production line to generate multi-source heterogeneous data of the equipment status.

The present invention performs self-noise interference signal analysis on multimodal sensors, identifies and filters the noise signals generated by the sensors themselves, improves the purity and effectiveness of data, and effectively reduces the interference of noise on real-time monitoring equipment status signals through analysis and processing of self-noise interference signals, thereby improving the quality and accuracy of signals. Effective signal analysis is performed on real-time monitoring equipment status signals according to the self-noise interference signals of the sensors, ensuring that the collected data has a high signal-to-noise ratio. During the effective signal analysis process, key information of the equipment status can be accurately extracted, thereby improving the reliability and practicality of the data. According to the sensor network topology, the signal source of the effective monitoring equipment status signal is identified, and an identification measuring equipment status signal is generated, ensuring that the source of each signal is clear and traceable. The signal source identification can help to carry out refined data management, improve the organization and systematicness of data processing, and facilitate subsequent data analysis and application. The identification measuring equipment status signal is converted into multi-source heterogeneous data of the equipment status of the workshop production line, and multi-source heterogeneous data of the equipment status is generated, ensuring the effective conversion of data between different formats and types. The multi-source heterogeneous data conversion improves the compatibility of the data, facilitates subsequent integrated processing and analysis, and ensures that the data can flow efficiently between different systems and platforms.

In the embodiment of the present invention, during the actual operation, the multimodal sensor will inevitably be interfered by the noise generated by itself. First, the output signal of the sensor in a static or standard environment is recorded. These signals are mainly caused by the self-noise of the sensor. The collected noise data is analyzed by spectrum analysis and time domain analysis using signal processing technology (such as fast Fourier transform (FFT), wavelet transform), and the characteristic parameters of the noise are extracted. According to the extracted noise characteristic parameters, a sensor self-noise interference signal model is established for subsequent signal analysis and noise suppression. The real-time monitored device status signal is compared with the sensor self-noise interference signal, and a denoising algorithm (such as adaptive filtering, Kalman filtering) is used to eliminate or reduce noise interference, and an effective device status signal is extracted. The signal enhancement technology (such as average filtering, empirical mode decomposition) is applied to further improve the signal-to-noise ratio of the signal to ensure that the extracted effective signal has high accuracy and reliability. After denoising and enhancement processing, an effective monitoring device status signal is generated for subsequent analysis and processing. According to the topological structure of the sensor network, a corresponding relationship between the device status signal and the sensor position is established to ensure that each signal source can be accurately identified. The distributed network nodes corresponding to the device status signal collection in the topological structure of the sensor network are used to identify the source of the effective monitoring device status signal, determine the specific sensor node of the signal source, and attach source identification information to each effective monitoring device status signal to generate an identified device status signal, ensuring that the signal can be traced back to the specific sensor node in subsequent processing.

Preferably, step S2 comprises the following steps:

Step S21: performing multi-source data classification processing on the multi-source data of the device status to generate single-source classified data of the device status;

Step S22: selecting the single-source classified data of device status from the multi-source data of device status one by one as a single analysis variable, performing a single-source subset impact similarity evaluation of the device status on the single-source classified data of device status corresponding to the single analysis variable, and generating single-source subset impact similarity data;

Step S23: performing single-source subset distribution feature node analysis based on the single-source subset impact similarity data to generate single-source subset distribution feature node data;

Step S24: selecting device status multi-source feature data from the device status multi-source data according to the single-source subset distribution feature node data to generate the device status multi-source feature data.

The present invention classifies and processes the multi-source data of the equipment status, classifies the data by source and type, generates the single-source classified data of the equipment status, makes the data structure clearer, and after the classification and processing, can adopt appropriate processing and analysis methods for different categories of data, improve the accuracy and effect of data analysis. The single-source classified data in the multi-source data of the equipment status are selected one by one as a single analysis variable to perform the single-source subset impact similarity evaluation of the equipment status, identify the similarity and correlation between different data subsets, and improve the depth and breadth of data analysis. Based on the single-source subset impact similarity data, the distribution feature node analysis is performed, the key nodes in the equipment status data can be identified, and the distribution pattern and trend in the data can be discovered, which is helpful for equipment status monitoring and prediction. Feature data selection is performed on the multi-source data of the equipment status according to the single-source subset distribution feature node data, and the most valuable data for equipment status monitoring and control is extracted. The feature data selection process optimizes the data set, reduces redundant data, and improves data processing efficiency and system response speed.

As an embodiment of the present invention, referring to FIG. 2 , which is a schematic flow chart of detailed implementation steps of step S2 in FIG. 1 , in this embodiment, step S2 includes:

Step S21: performing multi-source data classification processing on the multi-source data of the device status to generate single-source classified data of the device status;

In the embodiment of the present invention, multi-source data of the device status are collected from various multimodal sensors, and these data include temperature, vibration, noise and other types, and the multi-source data of the device status are divided into different single-source classified data according to the data type. For example, all temperature-related data are classified into one category, and vibration data are classified into another category. Through classification processing, single-source classified data of the device status are generated, ensuring that each type of data has a clear classification label, which is convenient for subsequent analysis and processing.

Step S22: selecting the single-source classified data of device status from the multi-source data of device status one by one as a single analysis variable, performing a single-source subset impact similarity evaluation of the device status on the single-source classified data of device status corresponding to the single analysis variable, and generating single-source subset impact similarity data;

In an embodiment of the present invention, single-source classified data in the multi-source data of device status are selected one by one as a single analysis variable. For example, temperature data is selected as the analysis variable, and the impact similarity of the single-source classified data of device status corresponding to the single analysis variable is evaluated. The similarity between data is calculated using statistical analysis methods (such as Pearson correlation coefficient, cosine similarity) or machine learning methods (such as K nearest neighbor algorithm), and single-source subset impact similarity data is generated based on the similarity calculation results. These data reflect the similarity between different data points in the single-source classified data, which helps to identify association patterns in the data.

Step S23: performing single-source subset distribution feature node analysis based on the single-source subset impact similarity data to generate single-source subset distribution feature node data;

In the embodiment of the present invention, a detailed distribution analysis is performed based on the impact similarity data of the single-source subset. Cluster analysis (such as hierarchical clustering and K-means clustering) is used to identify clustering patterns and characteristic nodes in the data. The characteristic nodes in the data are identified through the cluster analysis results. The characteristic nodes refer to nodes that are representative or critical in the data distribution. These nodes can reflect the key changes in the device status. Based on the identified characteristic nodes, the single-source subset distribution characteristic node data is generated.

Step S24: selecting device status multi-source feature data from the device status multi-source data according to the single-source subset distribution feature node data to generate the device status multi-source feature data.

In an embodiment of the present invention, based on the single-source subset distribution feature node data, data related to these feature nodes are selected from the multi-source data of the device status, and the key features in the multi-source data of the device status are extracted using feature extraction methods (such as principal component analysis and factor analysis) to ensure that the selected features have high relevance and validity. The selected multi-source feature data are integrated into a new feature data set, and data integration technology (such as data fusion and data merging) is used to generate multi-source feature data of the device status.

Preferably, step S23 includes the following steps:

Perform single-source subset distribution impact probability evaluation based on single-source subset impact similarity data to generate single-source subset distribution impact probability data;

Perform KL divergence calculation on the single-source subset distribution impact probability data to generate single-source subset distribution impact KL divergence data;

According to the KL divergence data of single-source subset distribution influence, single-source subset distribution feature node analysis is performed to generate single-source subset distribution feature node data.

The present invention performs distribution influence probability evaluation based on single-source subset influence similarity data, generates single-source subset distribution influence probability data, can quantify the mutual influence and correlation between different subsets, more accurately identify and describe the distribution characteristics and change laws of single-source subsets, and provide a reliable data basis for subsequent analysis. KL divergence calculation is performed on the single-source subset distribution influence probability data to quantify the differences between different subset distributions. KL divergence calculation can effectively measure the differences between two probability distributions, provide a method for measuring the amount of information, and help to deeply understand the data distribution characteristics. Single-source subset distribution feature node analysis is performed based on the single-source subset distribution influence KL divergence data, single-source subset distribution feature node data is generated, key feature nodes in the data are identified, and important patterns and trends in the single-source subset distribution are discovered and extracted through feature node analysis, thereby enhancing the understanding and prediction capabilities of the device status.

In an embodiment of the present invention, a pre-designed single-source subset impact probability assessment model (such as a Gaussian mixture model (GMM) or a Bayesian network) is selected to perform a single-source subset distribution impact probability assessment on the single-source subset impact similarity data, and the single-source subset impact similarity data is input into the selected probability assessment model for training, and the distribution probability of each data point is obtained. According to the model output, the single-source subset distribution impact probability data is generated. The data represents the distribution probability of each data point in the single-source subset. The KL divergence formula is used to perform KL divergence calculation on the single-source subset distribution impact probability data, and a standard reference distribution (such as a uniform distribution or a historical normal distribution) is selected as a comparison object, and the KL divergence value between the single-source subset distribution impact probability data and the reference distribution is calculated one by one. This calculation is implemented using a programming language (such as the scipy library in Python), and according to the calculation results, the single-source subset distribution impact KL divergence data is generated, which represents the degree of difference between each data point and the reference distribution. Cluster analysis methods (such as K-means clustering and DBSCAN) are used to perform nonlinear analysis on the KL divergence data affected by the single-source subset distribution, identify the main features in the data, and identify representative or critical single-source subset data in the clustering results. For example, redundant single-source subset data is replaced with similar single-source subset data. The identified feature nodes are verified through statistical analysis to ensure their accuracy and effectiveness. Based on the results of identification and verification, single-source subset distribution feature node data is generated, and the position and feature value of each feature node in the data set are marked.

Preferably, step S3 comprises the following steps:

Step S31: integrating the time series data of the multi-source characteristic data of the device status to generate the time series multi-source characteristic data of the device status;

Step S32: performing device state time series trend analysis on the device state time series multi-source feature data based on a preset long short-term memory neural network algorithm to generate device state time series trend data;

Step S33: Selecting the equipment status single-source classification data corresponding to the equipment status time series trend data to perform equipment status time series trend evaluation binary tree design, so as to establish an equipment status time series trend evaluation binary tree;

Step S34: identifying abnormal device state timing trend on the device state timing trend data according to the device state timing trend evaluation binary tree, and generating abnormal device state timing trend data.

The present invention integrates the time series data of multi-source characteristic data of equipment status to ensure the consistency of data from different sources and types in the time dimension. The time series data integration makes the data have a time dimension and can dynamically monitor the changes in equipment status. Based on the preset long short-term memory neural network (LSTM) algorithm, the time series trend analysis is performed on the multi-source characteristic data of equipment status time series. The LSTM algorithm can handle the long-term dependencies in the time series data and improve the accuracy and depth of the analysis. Through the time series trend analysis, the future change trend of the equipment status can be predicted, potential problems can be identified in advance, and the predictability and initiative of equipment management can be improved. The single-source classification data of the equipment status corresponding to the equipment status time series trend data are selected respectively, and the equipment status time series trend evaluation binary tree is designed to make the time series trend analysis more structured and systematic. Through the binary tree structure, the time series trend data can be evaluated from multiple dimensions, and the influence of different factors can be comprehensively considered to improve the comprehensiveness and accuracy of the trend analysis. Abnormal equipment status timing trend is identified based on the equipment status timing trend evaluation binary tree, which can accurately identify abnormal changes in equipment status. Abnormal trend identification provides a basis for equipment failure warning, discovers abnormal trends and potential failures of equipment in advance, and takes timely measures to reduce production downtime and maintenance costs.

As an embodiment of the present invention, referring to FIG. 2 , which is a schematic flow chart of detailed implementation steps of step S3 in FIG. 1 , in this embodiment, step S3 includes:

Step S31: integrating the time series data of the multi-source characteristic data of the device status to generate the time series multi-source characteristic data of the device status;

In an embodiment of the present invention, multi-source feature data of the device state is obtained from each multimodal sensor to ensure that the data has a consistent timestamp. Time series data of different sensors are aligned using time alignment techniques such as linear interpolation or spline interpolation to ensure the temporal consistency of the data. The aligned multi-source feature data are time-series fused, and a data fusion algorithm (such as weighted average, Bayesian fusion) is used to generate device state time series multi-source feature data.

Step S32: performing device state time series trend analysis on the device state time series multi-source feature data based on a preset long short-term memory neural network algorithm to generate device state time series trend data;

In an embodiment of the present invention, multi-source feature data of device status time series are extracted from a time series database, and a long short-term memory neural network (LSTM) is selected as a time series trend analysis model. The LSTM model is good at processing long-term dependencies of time series data. The LSTM model is trained using historical data corresponding to the multi-source feature data of device status time series. The training process includes data normalization, model architecture design (such as selecting the number of LSTM layers and the number of neurons), loss function selection (such as mean square error) and optimization algorithm (such as Adam optimizer). The trained LSTM model is applied to the multi-source feature data of device status time series to generate device status time series trend data.

Step S33: Selecting the equipment status single-source classification data corresponding to the equipment status time series trend data to perform equipment status time series trend evaluation binary tree design, so as to establish an equipment status time series trend evaluation binary tree;

In the embodiment of the present invention, from the generated equipment status time series trend data, the equipment status single-source classification data corresponding to the representative time series data are respectively selected, and the equipment status time series trend evaluation binary tree is designed, and the hierarchical structure and splitting criteria of the binary tree are determined. The decision tree algorithm (such as CART, ID3) is used to automatically generate the binary tree structure. Since the equipment status of each type is considered separately, the equipment status time series trend evaluation binary tree is an isolated binary tree. Based on the selected equipment status single-source classification data, the binary tree nodes are split layer by layer, and each node corresponds to a range of characteristic values or classification criteria to ensure that the binary tree structure is clear and interpretable. The accuracy and robustness of the binary tree model are verified by cross-validation or leave-out validation methods to ensure that it can effectively evaluate the time series trend of the equipment status.

Step S34: identifying abnormal device state timing trend on the device state timing trend data according to the device state timing trend evaluation binary tree, and generating abnormal device state timing trend data.

In an embodiment of the present invention, the equipment status time series trend data is input into a designed time series trend evaluation binary tree for evaluation. The binary tree model recursively evaluates the input data, determines the classification or feature range to which the data point belongs layer by layer, generates an evaluation result, and uses a clustering algorithm such as K-means clustering or DBSCAN to perform cluster analysis on the binary tree recursive evaluation result to identify the cluster center and abnormal points in the data. Based on the clustering results, outlier point identification is performed to determine which data points deviate from the normal cluster center. These points are abnormal equipment status time series trend data. The identified abnormal data points are marked to generate the final abnormal equipment status time series trend data.

Preferably, step S34 includes the following steps:

Transmitting the equipment status timing trend data to the equipment status timing trend evaluation binary tree for binary tree recursive evaluation to generate binary tree recursive evaluation data;

Perform cluster analysis on the binary tree recursive evaluation data to generate clustered binary tree recursive evaluation data; perform outlier identification based on the clustered binary tree recursive evaluation data to generate binary tree recursive evaluation outlier data;

The abnormal nodes of the device status time series trend data are identified based on the outlier data recursively evaluated by the binary tree to generate abnormal device status time series trend data.

The present invention transmits the equipment status time series trend data to the equipment status time series trend evaluation binary tree for binary tree recursive evaluation, which can analyze the data in layers and refinements, and go deeper layer by layer from global to local, so as to improve the comprehensiveness and accuracy of data analysis. The recursive evaluation process is structured and systematized, ensuring that the data at each level is analyzed in detail, which helps to fully understand the changing trend of the equipment status. Cluster analysis is performed on the binary tree recursive evaluation data, which can identify and group similar patterns and laws in the data, and help to understand the common characteristics of equipment status changes. Cluster analysis effectively reduces the dimensionality of high-dimensional data, extracts the main features, reduces the complexity of data processing, and improves processing efficiency. Outlier identification is performed based on the clustered binary tree recursive evaluation data. Outlier identification can accurately detect abnormal points in the data, help to timely discover potential problems, timely warn of abnormal equipment status, take measures in advance, and reduce losses caused by equipment failures. Abnormal nodes in the equipment status time series trend data are identified based on the binary tree recursive evaluation of outlier data, and the abnormal nodes in the equipment status are accurately located, which helps to quickly identify and handle equipment failures. Abnormal node identification provides a basis for predictive maintenance of equipment, improves the efficiency and pertinence of maintenance work, reduces unplanned downtime, and reduces maintenance costs.

In an embodiment of the present invention, a preset device status time series trend evaluation binary tree model is used to transfer device status time series trend data to the model. The binary tree model recursively evaluates layer by layer based on the previously constructed hierarchical structure and splitting criteria. For each data point, the binary tree model starts from the root node and evaluates each node in turn according to the splitting criteria until it reaches the leaf node, records the result of the binary tree recursive evaluation, and generates binary tree recursive evaluation data, which includes the path of each data point in the binary tree and the final evaluation result. The generated binary tree recursive evaluation data is input into the clustering analysis algorithm, and an appropriate clustering algorithm is selected, such as K-means, DBSCAN or a hierarchical clustering algorithm. The clustering algorithm is used to perform cluster analysis on the binary tree recursive evaluation data, identify the cluster center and boundary in the data, and generate clustered binary tree recursive evaluation data. The data includes the cluster to which each data point belongs and the location of the cluster center. Apply outlier identification algorithms, such as LOF (local outlier factor), Z-score, or distance-based outlier detection algorithms, to perform outlier calculations on clustered data, identify data points that are significantly deviated from the cluster center or cluster boundary, generate binary tree recursive evaluation outlier data, and mark the location and eigenvalue of each outlier. Integrate the binary tree recursive evaluation outlier data with the equipment status time series trend data, identify abnormal nodes on the equipment status time series trend data based on the binary tree recursive evaluation outlier data, mark the data points identified as abnormal, and finally generate abnormal equipment status time series trend data.

Preferably, step S4 comprises the following steps:

Step S41: Obtaining equipment control parameters of the workshop production line;

Step S42: performing device driver association analysis based on the equipment control parameters of the workshop production line and the multi-source characteristic data of the equipment status time series to generate device driver association data;

Step S43: analyzing the dynamic drive mapping relationship of the device control parameters and the device status according to the device drive association data to generate a device dynamic drive data model;

Step S44: performing a preferred device driving characteristic data analysis on the device dynamic driving data model based on the abnormal device status trend data to generate preferred device driving characteristic data;

Step S45: Perform device adaptive control strategy analysis based on the preferred device driving characteristic data to generate a device adaptive control strategy.

The present invention obtains the equipment control parameters of the workshop production line, ensures that all relevant control parameters are comprehensively collected, provides detailed basic data, obtains control parameters in real time, ensures the timeliness and accuracy of data, and provides reliable input for subsequent analysis. According to the equipment control parameters of the workshop production line and the multi-source characteristic data of the equipment state sequence, the equipment drive association analysis is performed to identify the association relationship between the equipment control parameters and the equipment state, provide insight into the association level, and improve the comprehensiveness and comprehensiveness of data analysis by fusing the control parameters and the state characteristic data, laying the foundation for the establishment of dynamic drive mapping relationship. According to the equipment drive association data, the dynamic drive mapping relationship analysis of the equipment control parameters and the equipment state is performed, and a dynamic relationship data model between the equipment control and the state is established to improve the accuracy of the model. The dynamic drive mapping relationship analysis can adapt to the changes of the equipment under different working conditions and improve the flexibility and adaptability of the equipment control. Based on the abnormal equipment state trend data, the equipment dynamic drive data model is optimized for equipment drive feature data analysis, and the equipment drive feature data that will not have abnormal conditions in the equipment drive feature data is analyzed to identify the drive features that have the greatest impact on the equipment state, thereby improving the pertinence and effectiveness of the control strategy. The adaptive control strategy of the equipment is analyzed according to the preferred equipment driving characteristic data to realize the adaptive control of the equipment and improve the intelligence level of the production line. The adaptive control strategy can be dynamically adjusted according to the real-time changes in the equipment status, improve the stability and efficiency of the equipment operation, reduce manual intervention, and reduce the risk of abnormal downtime of the production line for the smart workshop.

In an embodiment of the present invention, a data interface is set in the control system of the workshop production line for real-time acquisition of control parameters of the equipment, such as temperature setting value, running speed, pressure, etc., and the equipment control parameters are acquired in real time using data acquisition software (such as SCADA system, PLC system) through the data interface. An appropriate association analysis model (such as Pearson correlation coefficient, Spearman rank correlation coefficient, etc.) is selected, and the equipment control parameters and equipment state time series multi-source feature data are input into the association analysis model to perform equipment drive association analysis, analyze the correlation between the control parameters and the equipment state, and generate equipment drive association data according to the model output to reflect the relationship between the equipment control parameters and the equipment state. Use multivariate regression analysis, neural network model or Bayesian network and other technologies to construct a dynamic drive mapping relationship model between equipment control parameters and equipment states, such as the corresponding equipment state time series change value under A equipment control parameter, and finally generate an equipment dynamic drive data model to describe how the equipment control parameters dynamically drive the equipment state change. According to the abnormal equipment status trend data, the feature data of the equipment dynamic drive data model is screened, and the key features are selected by feature engineering technology (such as feature extraction and feature selection). For example, the multi-source feature data of the equipment status time series in the equipment dynamic drive data model corresponding to the abnormal equipment status trend data is eliminated or the parameter control is adjusted to keep the equipment status from matching the abnormal equipment status trend data when the equipment is running, so as to generate the preferred equipment drive feature data that will not cause the production line equipment to stop. Use control theory (such as PID control, fuzzy control, and reinforcement learning) to establish an adaptive control strategy model, and design an adaptive control strategy that can adjust the control parameters in real time. For example, through reinforcement learning, the reward and punishment function is designed, and the equipment parameters that match the preferred equipment drive feature data correspond to the reward parameters of the reward and punishment function, while the equipment parameters that match the non-selected equipment drive feature data correspond to the punishment parameters of the reward and punishment function. Finally, the equipment adaptive control strategy is generated and prepared for implementation in the control system of the actual production line.

Preferably, step S51 includes the following steps:

Step S51: performing simulation equipment scheduling driving efficiency analysis according to the equipment adaptive control strategy and the equipment dynamic driving data model, and generating simulation equipment scheduling driving efficiency data;

Step S52: performing scheduling resource priority analysis according to the simulation device scheduling drive efficiency data to generate scheduling resource priority data;

Step S53: Perform scheduling optimization processing on the control parameters of the equipment adaptive control strategy according to the scheduling resource priority data, generate an optimized equipment adaptive control strategy, and feed back the optimized equipment adaptive control strategy to the terminal device to execute the intelligent control operation of the production line equipment.

The present invention performs simulation equipment scheduling drive efficiency analysis based on equipment adaptive control strategy and equipment dynamic drive data model, evaluates the efficiency and effect of equipment scheduling strategy in a virtual environment, reduces risks and costs in actual operation, and simulation analysis can find bottlenecks and deficiencies in equipment scheduling, provide a basis for optimization design, and improve the overall efficiency of equipment scheduling. According to the simulated equipment scheduling drive efficiency data, scheduling resource priority analysis is performed to identify the resources with the highest priority scheduling in different scenarios, improve the scientificity and rationality of scheduling decisions, and priority analysis helps to reasonably allocate and utilize resources, avoid resource waste, and improve resource utilization and the overall efficiency of the production line. According to the scheduling resource priority data, the equipment adaptive control strategy is optimized to schedule the control parameters, and the control parameters can be accurately adjusted to improve the execution effect of the control strategy. The optimized equipment adaptive control strategy is fed back to the terminal device to execute the intelligent control operation of the production line equipment. Intelligent feedback and closed-loop control are realized to improve the automation level and execution efficiency of equipment control.

In an embodiment of the present invention, the equipment adaptive control strategy and the equipment dynamic drive data model are imported into a simulation platform (such as MATLAB/Simulink, ANSYS or AnyLogic), and a virtual workshop is built according to the actual workshop production line, including equipment layout, production process and resource allocation. The initial conditions of the simulation are set, including the initial state of the equipment, production tasks and time steps, etc., and the initial parameters of the equipment adaptive control strategy are imported to ensure that the strategy can be adjusted in real time during the simulation process. The simulation program is started to simulate the operation process of the equipment under different control strategies, and the operation state, resource usage and production efficiency of the equipment during the simulation process are recorded. According to the data recorded by the simulation, the driving efficiency of the simulated equipment scheduling is calculated, including the production capacity, resource utilization and energy consumption of the production line, and the simulated equipment scheduling driving efficiency data is generated. The priority analysis algorithm (such as AHP (hierarchical analysis method) or TOPSIS (topic ideal solution sorting method)) and the pre-designed evaluation index system are used to perform resource priority analysis on the simulated equipment scheduling driving efficiency data, determine the priority evaluation data of the equipment scheduling, and dynamically adjust the priority evaluation data according to the simulation operation process. According to various resource parameters of the production line and the dynamically adjusted priority evaluation data, the priority of the equipment scheduling is analyzed to generate the scheduling resource priority data. Using genetic algorithms, particle swarm optimization algorithms or simulated annealing algorithms and pre-designed scheduling optimization objectives, the scheduling resource priority data and the equipment adaptive control strategy are optimized for the scheduling of control parameters, the optimal control parameter combination is iteratively searched, and the optimal control parameter combination is evaluated and adjusted according to the pre-designed scheduling optimization objectives, and finally the optimized equipment adaptive control strategy is generated. The optimized equipment adaptive control strategy is fed back to the control system of the production line, the optimized control strategy is implemented, and its execution effect is monitored in real time to ensure the smooth progress of the intelligent control operation of the production line and the continuous optimization of the equipment control strategy.

This specification provides a production line data acquisition control system for an intelligent workshop, which is used to execute the production line data acquisition control method for the intelligent workshop as described above. The production line data acquisition control system for the intelligent workshop includes:

The equipment status multi-source data acquisition module is used to collect multi-source heterogeneous data of the equipment status of the workshop production line based on multi-modal sensors to generate multi-source heterogeneous data of the equipment status; perform multi-source heterogeneous data integration processing on the multi-source heterogeneous data of the equipment status to generate multi-source data of the equipment status;

The device status multi-source feature analysis module is used to perform multi-source data classification processing on the device status multi-source data to generate the device status single-source classification data; perform single-source subset distribution feature node analysis on the device status based on the device status single-source classification data to generate single-source subset distribution feature node data; perform device status multi-source feature data selection on the device status multi-source data based on the single-source subset distribution feature node data to generate the device status multi-source feature data;

The equipment status time series multi-source feature analysis module is used to integrate the time series data of the equipment status multi-source feature data to generate the equipment status time series multi-source feature data; based on the equipment status time series multi-source feature data, the abnormal equipment status time series trend is identified to generate the abnormal equipment status time series trend data;

The equipment adaptive control strategy analysis module is used to obtain the equipment control parameters of the workshop production line; based on the equipment control parameters of the workshop production line and the multi-source characteristic data of the equipment status time series, the dynamic drive mapping relationship analysis of the equipment control parameters and the equipment status is performed to generate the equipment dynamic drive data model; based on the abnormal equipment status trend data and the equipment dynamic drive data model, the equipment adaptive control strategy analysis is performed to generate the equipment adaptive control strategy;

The equipment adaptive control strategy optimization module is used to schedule and optimize the control parameters of the equipment adaptive control strategy, generate an optimized equipment adaptive control strategy, and feed back the optimized equipment adaptive control strategy to the terminal device to execute the intelligent control operation of the production line equipment.

The beneficial effect of the present application is that the production line data acquisition and control method of the intelligent workshop of the present invention significantly improves the intelligent control level and operation efficiency of the production line by comprehensively using multimodal sensor data acquisition, classification processing and feature analysis of multi-source data, integration of time series data and abnormal trend identification, and dynamic mapping and optimization of adaptive control strategies. Through efficient data acquisition and integrated processing, the equipment status can be fully monitored and accurately analyzed to ensure real-time update and high consistency of data. The use of long short-term memory networks and binary tree structures for trend analysis and abnormal state identification greatly improves the accuracy of fault prediction and early warning capabilities, thereby reducing the risk of equipment failure and downtime. Through simulation and optimization analysis, the optimal allocation of resources and fine adjustment of control strategies are achieved, the response speed and adaptability of the production line are enhanced, and the efficiency and stability of the production process are ensured.

Therefore, the embodiments should be regarded as illustrative and non-restrictive from all points, and the scope of the present invention is limited by the appended claims rather than the above description, and it is intended that all changes falling within the meaning and range of equivalent elements of the application documents are included in the present invention.

The above description is only a specific embodiment of the present invention, so that those skilled in the art can understand or implement the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but should conform to the widest scope consistent with the principles and novel features invented herein.

Claims (10)

1. A production line data acquisition control method for an intelligent workshop, characterized in that it comprises the following steps: Step S1: collecting multi-source heterogeneous data of equipment status of a workshop production line according to a multimodal sensor to generate multi-source heterogeneous data of equipment status; performing multi-source heterogeneous data integration processing on the multi-source heterogeneous data of equipment status to generate multi-source data of equipment status; Step S2: performing multi-source data classification processing on the multi-source data of the device status to generate single-source classified data of the device status; performing single-source subset distribution feature node analysis of the device status based on the single-source classified data of the device status to generate single-source subset distribution feature node data; performing multi-source feature data selection on the multi-source data of the device status according to the single-source subset distribution feature node data to generate multi-source feature data of the device status; Step S3: integrating the time series data of the multi-source characteristic data of the device status to generate the time series multi-source characteristic data of the device status; identifying the time series trend of the abnormal device status based on the time series multi-source characteristic data of the device status to generate the time series trend data of the abnormal device status; Step S4: Acquire equipment control parameters of the workshop production line; perform dynamic drive mapping relationship analysis of equipment control parameters and equipment status based on equipment control parameters of the workshop production line and equipment status time series multi-source feature data to generate a dynamic drive data model for the equipment; perform equipment adaptive control strategy analysis based on abnormal equipment status time series trend data and the equipment dynamic drive data model to generate an adaptive control strategy for the equipment; Step S5: Perform scheduling optimization processing on the control parameters of the equipment adaptive control strategy, generate an optimized equipment adaptive control strategy, and feed back the optimized equipment adaptive control strategy to the terminal device to execute the intelligent control operation of the production line equipment. 2. The production line data acquisition control method of the intelligent workshop according to claim 1 is characterized in that step S1 comprises the following steps: Step S11: configuring the multimodal sensor as a sensor node based on a preset sensor configuration requirement to obtain sensor node configuration data, and setting a sensor network topology structure through the sensor configuration data; Step S12: Real-time monitoring of equipment status signals of the workshop production line is performed according to the multimodal sensor to generate real-time monitoring equipment status signals; Step S13: Perform multi-source heterogeneous data analysis on the equipment status of the workshop production line according to the real-time monitoring equipment status signal to generate multi-source heterogeneous data on the equipment status; Step S14: performing time sequence synchronization correction on the device state multi-source heterogeneous data to generate synchronized device state multi-source heterogeneous data; Step S15: analyzing the device status integration queue according to the sensor network topology structure, and generating the device status integration queue; Step S16: performing multi-source heterogeneous data integration processing on the synchronous device status multi-source heterogeneous data according to the device status integration queue to generate device status multi-source data. 3. The production line data acquisition control method of the intelligent workshop according to claim 2 is characterized in that step S13 comprises the following steps: Perform sensor self-noise interference signal analysis on multimodal sensors to generate sensor self-noise interference signals; Perform effective signal analysis on the real-time monitoring device status signal based on the sensor self-noise interference signal to generate an effective monitoring device status signal; According to the sensor network topology, the signal source of the effective monitoring device status signal is identified to generate an identification measuring device status signal; The device status signal of the identification measurement device is converted into multi-source heterogeneous data of the equipment status of the workshop production line to generate multi-source heterogeneous data of the equipment status. 4. The production line data acquisition control method of the intelligent workshop according to claim 1 is characterized in that step S2 comprises the following steps: Step S21: performing multi-source data classification processing on the multi-source data of the device status to generate single-source classified data of the device status; Step S22: selecting the single-source classified data of device status in the multi-source data of device status one by one as a single analysis variable, performing a single-source subset impact similarity evaluation of the device status on the single-source classified data of device status corresponding to the single analysis variable, and generating single-source subset impact similarity data; Step S23: performing single-source subset distribution feature node analysis based on the single-source subset impact similarity data to generate single-source subset distribution feature node data; Step S24: selecting device status multi-source feature data from the device status multi-source data according to the single-source subset distribution feature node data to generate the device status multi-source feature data. 5. The production line data acquisition control method of the intelligent workshop according to claim 4 is characterized in that step S23 comprises the following steps: Perform single-source subset distribution impact probability evaluation based on single-source subset impact similarity data to generate single-source subset distribution impact probability data; Perform KL divergence calculation on the single-source subset distribution impact probability data to generate single-source subset distribution impact KL divergence data; According to the KL divergence data of single-source subset distribution influence, single-source subset distribution feature node analysis is performed to generate single-source subset distribution feature node data. 6. The production line data acquisition control method of the intelligent workshop according to claim 1, characterized in that step S3 comprises the following steps: Step S31: integrating the time series data of the multi-source characteristic data of the device status to generate the time series multi-source characteristic data of the device status; Step S32: performing device state time series trend analysis on the device state time series multi-source feature data based on a preset long short-term memory neural network algorithm to generate device state time series trend data; Step S33: Selecting the equipment status single-source classification data corresponding to the equipment status time series trend data to perform equipment status time series trend evaluation binary tree design, so as to establish an equipment status time series trend evaluation binary tree; Step S34: identifying abnormal device state timing trend on the device state timing trend data according to the device state timing trend evaluation binary tree, and generating abnormal device state timing trend data. 7. The production line data acquisition control method of the intelligent workshop according to claim 1, characterized in that step S34 comprises the following steps: Transmitting the equipment status timing trend data to the equipment status timing trend evaluation binary tree for binary tree recursive evaluation to generate binary tree recursive evaluation data; Perform cluster analysis on the binary tree recursive evaluation data to generate clustered binary tree recursive evaluation data; perform outlier identification based on the clustered binary tree recursive evaluation data to generate binary tree recursive evaluation outlier data; The abnormal nodes of the device status time series trend data are identified based on the outlier data recursively evaluated by the binary tree to generate abnormal device status time series trend data. 8. The production line data acquisition control method of an intelligent workshop according to claim 1, characterized in that step S4 comprises the following steps: Step S41: Obtaining equipment control parameters of the workshop production line; Step S42: performing device driver association analysis based on the equipment control parameters of the workshop production line and the multi-source characteristic data of the equipment status time series to generate device driver association data; Step S43: analyzing the dynamic drive mapping relationship of the device control parameters and the device status according to the device drive association data to generate a device dynamic drive data model; Step S44: performing device driving feature analysis on the device dynamic driving data model based on the abnormal device state time series trend data to generate device driving feature data, wherein the device driving feature data is the driving feature that has the greatest impact on the device state among the driving features that will not cause abnormal conditions; Step S45: Perform device adaptive control strategy analysis based on the device driving characteristic data to generate a device adaptive control strategy. 9. The production line data acquisition control method of an intelligent workshop according to claim 1, characterized in that step S51 comprises the following steps: Step S51: performing simulation equipment scheduling driving efficiency analysis according to the equipment adaptive control strategy and the equipment dynamic driving data model, and generating simulation equipment scheduling driving efficiency data; Step S52: performing scheduling resource priority analysis according to the simulation device scheduling drive efficiency data to generate scheduling resource priority data; Step S53: Perform scheduling optimization processing on the control parameters of the equipment adaptive control strategy according to the scheduling resource priority data, generate an optimized equipment adaptive control strategy, and feed back the optimized equipment adaptive control strategy to the terminal device to execute the intelligent control operation of the production line equipment. 10. A production line data acquisition control system for an intelligent workshop, characterized in that it is used to execute the production line data acquisition control method for an intelligent workshop as claimed in claim 1, and the production line data acquisition control system for an intelligent workshop comprises: The equipment status multi-source data acquisition module is used to collect multi-source heterogeneous data of the equipment status of the workshop production line based on multi-modal sensors to generate multi-source heterogeneous data of the equipment status; perform multi-source heterogeneous data integration processing on the multi-source heterogeneous data of the equipment status to generate multi-source data of the equipment status; The device status multi-source feature analysis module is used to perform multi-source data classification processing on the device status multi-source data to generate the device status single-source classification data; perform single-source subset distribution feature node analysis on the device status based on the device status single-source classification data to generate single-source subset distribution feature node data; perform device status multi-source feature data selection on the device status multi-source data based on the single-source subset distribution feature node data to generate the device status multi-source feature data; The equipment status time series multi-source feature analysis module is used to integrate the time series data of the equipment status multi-source feature data to generate the equipment status time series multi-source feature data; based on the equipment status time series multi-source feature data, the abnormal equipment status time series trend is identified to generate the abnormal equipment status time series trend data; The equipment adaptive control strategy analysis module is used to obtain the equipment control parameters of the workshop production line; based on the equipment control parameters of the workshop production line and the multi-source characteristic data of the equipment status time series, the dynamic drive mapping relationship analysis of the equipment control parameters and the equipment status is performed to generate the equipment dynamic drive data model; based on the abnormal equipment status time series trend data and the equipment dynamic drive data model, the equipment adaptive control strategy analysis is performed to generate the equipment adaptive control strategy; The equipment adaptive control strategy optimization module is used to schedule and optimize the control parameters of the equipment adaptive control strategy, generate an optimized equipment adaptive control strategy, and feed back the optimized equipment adaptive control strategy to the terminal device to execute the intelligent control operation of the production line equipment.