CN117434912B

CN117434912B - Method and system for monitoring production environment of non-woven fabric product

Info

Publication number: CN117434912B
Application number: CN202311772564.2A
Authority: CN
Inventors: 王琳
Original assignee: Ningjin Runboda Medical Protection Articles Co ltd
Current assignee: Ningjin Runboda Medical Protection Articles Co ltd
Priority date: 2023-12-21
Filing date: 2023-12-21
Publication date: 2024-02-20
Anticipated expiration: 2043-12-21
Also published as: CN117434912A

Abstract

The invention relates to the technical field of production monitoring, in particular to a method and a system for monitoring production environment of non-woven fabrics, which comprise the following steps: based on the characteristics of the non-woven fabric production line, a computer vision technology and a space analysis algorithm are adopted to divide production links and generate a partition monitoring map. In the invention, accurate production link division is realized through a computer vision technology and a spatial analysis algorithm, the monitoring accuracy is improved, a real-time environment data set is utilized to effectively monitor and dynamically adjust a production line, the reaction speed and the reaction efficiency are improved, the environmental data change trend can be deeply analyzed through the application of time sequence analysis and association rule learning, the environment risk assessment capability is enhanced through the application of a hidden Markov model and a decision tree algorithm, the preventive maintenance is effectively realized through combining reinforcement learning and predictive maintenance strategies, and the multi-environmental-friendly parameter is optimally adjusted through a recurrent neural network and reinforcement learning algorithm, so that the efficiency and the environmental protection of the production environment are improved.

Description

Method and system for monitoring production environment of non-woven fabric product

Technical Field

The invention relates to the technical field of production monitoring, in particular to a method and a system for monitoring production environment of non-woven fabric products.

Background

The technical field of production monitoring is a key component of industrial automation and intelligent manufacturing. The field of production monitoring technology relates to the use of various automated tools and techniques to monitor, control and optimize a production process. The aim in this field is to improve the production efficiency, to ensure the quality of the product, to reduce the costs and to ensure the safety of the working environment. Production monitoring techniques typically include real-time data collection, process monitoring, fault diagnosis, resource allocation, and scheduling optimization. In such specific applications of nonwoven production, quality control of raw materials, operating efficiency of production lines, maintenance of environmental parameters (e.g. temperature, humidity), and quality inspection of finished products are involved.

The method for monitoring the production environment of the non-woven fabric product is a technical method which is specially used for monitoring and managing the production process of the non-woven fabric product. The purpose of this method is to ensure the quality and efficiency of the nonwoven fabric during production, while reducing production costs and waste. By monitoring the production environment, problems in the production process, such as equipment failure or raw material defects, can be found and solved in time, so that the consistency and quality of the final product are ensured. In addition, the monitoring method is also beneficial to improving the safety of the working environment and preventing accidents. In order to achieve the purpose of monitoring the production environment of nonwoven fabric products, various means are generally adopted. This includes the installation of various types of sensors to monitor in real time critical parameters on the production line such as temperature, pressure, humidity and machine operating conditions. The data acquisition system will collect this data in real time, and predict equipment failure and production anomalies through advanced data analysis methods (e.g., machine learning algorithms). In addition, the monitoring system can be integrated into production management software to realize optimal allocation of resources and automatic adjustment of production flow.

The traditional monitoring method for the production environment of the non-woven fabric product has various defects. These methods often lack a fine division of the production links, resulting in generalized and inaccurate monitoring strategies. In terms of data processing, real-time analysis and processing often cannot be achieved, and therefore production efficiency and response speed are affected. Furthermore, conventional approaches suffer from deficiencies in the application of advanced data analysis and risk assessment tools, which make it difficult to effectively predict and address potential environmental risks. The lack of an effective preventive maintenance strategy is also an important drawback, often resulting in measures being taken after the problem has occurred. Moreover, the traditional method is relatively fixed and passive in terms of production environment adjustment, lacks dynamic adjustment capability based on real-time data, and limits optimization and environmental protection performance of production.

Disclosure of Invention

The invention aims to solve the defects in the prior art, and provides a method and a system for monitoring the production environment of a non-woven fabric product.

In order to achieve the above purpose, the present invention adopts the following technical scheme: the method for monitoring the production environment of the non-woven fabric product comprises the following steps:

s1: based on the characteristics of a non-woven fabric production line, carrying out production link division by adopting a computer vision technology and a space analysis algorithm, and generating a partition monitoring map;

S2: based on the subarea monitoring map, adopting an expert system and a rule reasoning technology to formulate an environment monitoring rule aiming at multiple links and form an environment monitoring rule set;

s3: based on the environment monitoring rule set, deploying monitoring equipment, and collecting environment data of multiple links in real time to generate a real-time environment data set;

s4: based on the real-time environment data set, adopting time sequence analysis and association rule learning to analyze the change trend of the environment data and generate an environment trend analysis report;

s5: based on the environmental trend analysis report, performing environmental risk assessment by using a hidden Markov model and a decision tree algorithm to generate an environmental risk assessment report;

s6: based on the environmental risk assessment report, adopting reinforcement learning and predictive maintenance strategies to automatically adjust monitoring parameters so as to realize preventive maintenance strategies;

s7: based on the preventive maintenance strategy, optimizing and adjusting environmental parameters of multiple loops of the production line by using a recurrent neural network and a reinforcement learning algorithm, and establishing optimized production environment configuration;

the regional monitoring map comprises a raw material processing region, a textile region and a finished product processing region, the environment monitoring rule set comprises a temperature range, a humidity level and a noise limit value of multiple links, the real-time environment data set comprises real-time monitoring data of the temperature, the humidity and the noise of multiple links, the environment trend analysis report comprises an environment change mode and potential association of the multiple links, the environment risk assessment report comprises key risk points and potential risk factors of production links, the preventive maintenance strategy comprises environment parameter adjustment and monitoring measure updating, and the optimized production environment configuration comprises optimized temperature, humidity and noise reduction settings of each link.

As a further scheme of the invention, based on the characteristics of a non-woven fabric production line, adopting a computer vision technology and a space analysis algorithm to divide production links and generating a partition monitoring map comprises the following steps:

s101: based on a non-woven fabric production line, performing preliminary division by adopting a computer vision technology to generate a preliminary partition model;

s102: based on the preliminary partition model, carrying out refinement space identification and division by using a deep learning algorithm to form a refinement partition scheme;

s103: based on the refined partition scheme, the geographic information system technology is applied to optimize the space layout, and the space layout optimization design is completed;

s104: integrating the space layout optimization design, and generating a partition monitoring map by using a graphical user interface design tool;

the computer vision technology comprises a feature extraction algorithm and an image segmentation algorithm, the deep learning algorithm comprises hierarchical feature learning and spatial relation recognition of a convolutional neural network, the geographic information system technology comprises spatial database management and map visualization, and the graphical user interface design tool comprises interface layout design and interactive design.

As a further scheme of the invention, based on the subarea monitoring map, an expert system and a rule reasoning technology are adopted to formulate environment monitoring rules aiming at multiple links, and the steps of forming an environment monitoring rule set are specifically as follows:

S201: based on the partition monitoring map, collecting environmental requirements of a production link by adopting an expert system technology, and collecting environmental requirement data;

s202: based on the environmental requirement data, formulating a preliminary environmental monitoring rule by using a rule reasoning technology to form a preliminary environmental monitoring rule set;

s203: verifying and adjusting the preliminary environment monitoring rule set, and establishing an optimized environment monitoring rule set by adopting a simulation test and feedback adjustment technology;

s204: based on the optimized environment monitoring rule set, integrating and formatting rules by using a knowledge management system to form an environment monitoring rule set;

the expert system technology comprises knowledge base construction and an reasoning mechanism, the rule reasoning technology comprises logic reasoning and pattern matching, the simulation test comprises virtual environment construction and scene simulation, and the knowledge management system comprises knowledge extraction and content management.

As a further scheme of the invention, based on the environment monitoring rule set, monitoring equipment is deployed, environment data of multiple links are collected in real time, and the step of generating the real-time environment data set comprises the following steps:

s301: based on the environment monitoring rule set, selecting monitoring equipment, determining a monitoring position and an installation method, and performing equipment layout optimization by adopting a spatial layout optimization algorithm to generate a monitoring equipment configuration scheme;

S302: based on the monitoring equipment configuration scheme, equipment installation and network configuration are carried out, and a network connection stability test method is adopted to verify the connection stability of equipment and a central monitoring system and generate a monitoring network construction report;

s303: based on the monitoring network construction report, data acquisition is implemented, and a real-time data transmission protocol is adopted to verify the real-time performance and integrity of the data, so as to generate a real-time environment data stream;

s304: based on the real-time environment data stream, data integration and formatting are carried out, and a real-time environment data set is generated by adopting a data cleaning and preprocessing technology;

the space layout optimization algorithm comprises particle swarm optimization and genetic algorithm, the network connection stability test method comprises network delay test and data packet loss rate test, the real-time data transmission protocol comprises message queue telemetry transmission, and the data cleaning and preprocessing technology comprises missing value processing, outlier detection and data normalization.

As a further scheme of the invention, based on the real-time environment data set, the steps of analyzing the change trend of the environment data and generating an environment trend analysis report by adopting time sequence analysis and association rule learning are specifically as follows:

S401: based on the real-time environment data set, adopting a time sequence analysis algorithm to analyze time dependence and trend in the data and generate a time sequence analysis result;

s402: based on the time sequence analysis result, adopting an association rule learning algorithm to identify the association between environmental factors and generate an association rule learning report;

s403: synthesizing the time sequence analysis result and the association rule learning report, and displaying the data change trend and the correlation by adopting a data visualization technology to generate environment data visualization display;

s404: based on the visual display of the environmental data, adopting a report writing strategy to explain the analysis result and the meaning of the analysis result on environmental monitoring, and generating an environmental trend analysis report;

the time sequence analysis algorithm is specifically an autoregressive model and a moving average model, the association rule learning algorithm is specifically an Apriori algorithm and an FP-Growth algorithm, the data visualization technology comprises scatter diagram making, line diagram making and thermodynamic diagram making, and the report writing strategy comprises data interpretation, trend prediction and influence assessment.

As a further aspect of the present invention, based on the environmental trend analysis report, using a hidden markov model and a decision tree algorithm, performing environmental risk assessment, and generating an environmental risk assessment report specifically includes:

S501: based on the environmental trend analysis report, carrying out sequence analysis of environmental states by adopting a hidden Markov model, identifying a potential risk mode, and generating a hidden Markov model analysis result;

s502: based on the hidden Markov model analysis result, classifying and evaluating risk factors by adopting a decision tree algorithm to generate a decision tree risk evaluation result;

s503: synthesizing the hidden Markov model analysis result and the decision tree risk assessment result, carrying out risk point analysis, locking a key monitoring area and generating a risk point analysis report;

s504: based on the risk point analysis report, writing a comprehensive environmental risk assessment report, formulating a risk level and countermeasures, and generating an environmental risk assessment report;

the hidden Markov model specifically comprises state transition probability analysis and observation probability analysis, the decision tree algorithm comprises information gain calculation, tree construction and pruning, and the risk point analysis comprises risk classification and key risk factor identification.

As a further scheme of the invention, based on the environmental risk assessment report, a reinforcement learning and predictive maintenance strategy is adopted to automatically adjust monitoring parameters, and the steps for realizing the preventive maintenance strategy are specifically as follows:

S601: based on the environmental risk assessment report, performing an automatic adjustment test of monitoring parameters by using a reinforcement learning algorithm, searching for optimal parameter configuration, and generating a reinforcement learning adjustment scheme;

s602: based on the reinforcement learning adjustment scheme, implementing automatic adjustment of the monitoring parameters, and adopting an analog feedback evaluation method to evaluate the adjustment effect and generate a monitoring parameter optimization report;

s603: based on the monitoring parameter optimization report, planning a future monitoring scheme by adopting a predictive maintenance strategy, including equipment maintenance and parameter adjustment planning, and generating a preventive maintenance plan;

s604: based on the preventive maintenance plan, preventive maintenance measures are implemented, long-term stable operation of the monitoring system is maintained, and a preventive maintenance strategy is generated;

the reinforcement learning algorithm comprises a reward function design, exploration and utilization strategy, the simulation feedback evaluation method comprises parameter adjustment simulation and performance evaluation, the predictive maintenance strategy comprises a fault prediction model and maintenance scheduling, and the preventive maintenance measures comprise maintenance operation execution, performance monitoring and feedback.

As a further scheme of the present invention, based on the preventive maintenance strategy, using a recurrent neural network and a reinforcement learning algorithm, optimizing and adjusting environmental parameters of multiple loops of a production line, the steps of establishing an optimized production environment configuration are specifically as follows:

S701: based on the preventive maintenance strategy, adopting a recurrent neural network to analyze historical production data and generating a key performance index analysis report;

s702: based on the key performance index analysis report, performing real-time environmental parameter optimization by using a reinforcement learning algorithm to generate optimized environmental parameter configuration;

s703: based on the optimized environment parameter configuration, applying a genetic algorithm to carry out multi-parameter combination and optimization to generate a genetic optimization production environment configuration scheme;

s704: based on the genetic optimization production environment configuration scheme, adopting a simulated annealing algorithm to carry out fine adjustment to generate an optimized production environment configuration;

the recurrent neural network comprises a long-term and short-term memory network and a gating circulation unit, the reinforcement learning algorithm comprises a Q learning and strategy gradient method, the genetic algorithm comprises crossing, mutation and selection operations, and the simulated annealing algorithm comprises a cooling plan design and a neighborhood searching strategy.

The non-woven fabric product production environment monitoring system is used for executing the non-woven fabric product production environment monitoring method, and comprises a production link dividing module, an environment monitoring rule making module, a monitoring equipment deployment module, an environment data analysis module, an environment risk assessment module and a preventive maintenance implementation module.

As a further scheme of the invention, the production link dividing module divides production links based on the characteristics of a non-woven fabric production line by adopting a computer vision technology and a space analysis algorithm and generates a partition monitoring map;

the environment monitoring rule making module is used for making environment monitoring rules and generating an environment monitoring rule set by adopting an expert system and rule reasoning technology based on the partition monitoring map of the production link dividing module;

the monitoring equipment deployment module deploys monitoring equipment based on an environment monitoring rule set of the environment monitoring rule formulation module, acquires multi-link environment data and generates a real-time environment data set;

the environment data analysis module analyzes the data change trend and generates an environment trend analysis report by adopting time sequence analysis and association rule learning based on a real-time environment data set of the monitoring equipment deployment module;

the environment risk assessment module is used for carrying out risk assessment by using a hidden Markov model and a decision tree algorithm based on an environment trend analysis report of the environment data analysis module and generating an environment risk assessment report;

the preventive maintenance implementation module automatically adjusts the monitoring parameters based on the environmental risk assessment report of the environmental risk assessment module using reinforcement learning and predictive maintenance strategies, and generates an optimized production environment configuration.

Compared with the prior art, the invention has the advantages and positive effects that:

according to the invention, more accurate production link division is realized through a computer vision technology and a space analysis algorithm, a regional monitoring map is generated, and the monitoring accuracy is improved. The production line is effectively monitored and dynamically adjusted by utilizing the real-time environment data set, and the reaction speed and the reaction efficiency are obviously improved. The application of time series analysis and association rule learning enables the environmental data change trend to be deeply analyzed, while the application of hidden Markov models and decision tree algorithms enhances the environmental risk assessment capability. By combining reinforcement learning and predictive maintenance strategies, preventive maintenance is effectively realized, and stable operation of the production line is ensured. And the recurrent neural network and the reinforcement learning algorithm are used for optimizing and adjusting the environmental parameters of the multiple loops, so that the efficiency and the environmental friendliness of the production environment are improved.

Drawings

FIG. 1 is a schematic workflow diagram of the present invention;

FIG. 2 is a S1 refinement flowchart of the present invention;

FIG. 3 is a S2 refinement flowchart of the present invention;

FIG. 4 is a S3 refinement flowchart of the present invention;

FIG. 5 is a S4 refinement flowchart of the present invention;

FIG. 6 is a S5 refinement flowchart of the present invention;

FIG. 7 is a S6 refinement flowchart of the present invention;

FIG. 8 is a S7 refinement flowchart of the present invention;

fig. 9 is a system flow diagram of the present invention.

Detailed Description

The present invention will be described in further detail with reference to the drawings and examples, in order to make the objects, technical solutions and advantages of the present invention more apparent. It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the scope of the invention.

In the description of the present invention, it should be understood that the terms "length," "width," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," "outer," and the like indicate orientations or positional relationships based on the orientation or positional relationships shown in the drawings, merely to facilitate describing the present invention and simplify the description, and do not indicate or imply that the devices or elements referred to must have a specific orientation, be configured and operated in a specific orientation, and therefore should not be construed as limiting the present invention. Furthermore, in the description of the present invention, the meaning of "a plurality" is two or more, unless explicitly defined otherwise. Embodiment one:

referring to fig. 1, the present invention provides a technical solution: the method for monitoring the production environment of the non-woven fabric product comprises the following steps:

s2: based on the partition monitoring map, adopting an expert system and a rule reasoning technology to formulate an environment monitoring rule aiming at multiple links and form an environment monitoring rule set;

s4: based on a real-time environment data set, adopting time sequence analysis and association rule learning to analyze the change trend of the environment data and generate an environment trend analysis report;

s6: based on the environmental risk assessment report, adopting reinforcement learning and predictive maintenance strategies to automatically adjust monitoring parameters, and realizing the preventive maintenance strategies;

s7: based on a preventive maintenance strategy, optimizing and adjusting environmental parameters of multiple loops of a production line by using a recurrent neural network and a reinforcement learning algorithm, and establishing optimized production environment configuration;

The regional monitoring map comprises a raw material processing region, a textile region and a finished product processing region, the environment monitoring rule set comprises a temperature range, a humidity level and a noise limit value of multiple links, the real-time environment data set comprises real-time monitoring data of the temperature, the humidity and the noise of the multiple links, the environment trend analysis report comprises an environment change mode and potential association of the multiple links, the environment risk assessment report comprises key risk points and potential risk factors of production links, the preventive maintenance strategy comprises environment parameter adjustment and monitoring measure update, and the optimized production environment configuration comprises optimized temperature, humidity and noise reduction setting of each link.

The production links are divided by a computer vision technology and a space analysis algorithm, and a partition monitoring map is generated, so that the accuracy and the efficiency of monitoring are greatly improved. The step ensures that each key link of the production line is clearly identified and specially focused, and the comprehensiveness and pertinence of monitoring are ensured. The formulation of the monitoring map also helps to understand the production flow more clearly, thereby improving the operability and effectiveness of the overall monitoring system.

The expert system and the rule reasoning technology are applied, so that the environment monitoring rule set is more perfect and careful, and specific monitoring standards can be set for different production links. The method not only improves the accuracy of environment monitoring, but also enhances the control capability of the production environment. Through the refined monitoring rules, potential problems in the production link can be effectively prevented, and the production risk is reduced.

The monitoring equipment is deployed, multi-link environment data are collected in real time, a real-time environment data set is generated, and a solid foundation is provided for subsequent data analysis and risk assessment. The collection and analysis of real-time data ensures real-time knowledge of the production environment so that any potential problems can be discovered and addressed at an early stage.

Through time series analysis and association rule learning, environmental data change trend is analyzed, and environmental trend analysis reports are generated, which enhances the prediction capability of production environmental change. The deep data analysis method can reveal complex relations among environmental parameters, forecast future change trend and provide scientific basis for production decision.

The hidden Markov model and the decision tree algorithm are used for carrying out environmental risk assessment, and an environmental risk assessment report is generated, so that risk management is more systematic and scientific. Through the advanced algorithms, risk points in the production link can be more accurately identified and evaluated, so that more effective risk prevention and relief measures are formulated.

The application of reinforcement learning and predictive maintenance strategies automatically adjusts monitoring parameters to realize the preventive maintenance strategies, and the step improves the adaptability and flexibility of the whole production system. The strategy can be automatically adjusted according to the environmental data and the risk assessment report, so that potential problems are effectively prevented, and unexpected shutdown and production loss are reduced.

The application of the recurrent neural network and the reinforcement learning algorithm optimizes and adjusts environmental parameters of multiple loops of the production line, and establishes the optimized production environmental configuration, which provides powerful support for the optimization of the whole production flow. The method not only improves the production efficiency, but also optimizes the use of resources and reduces the production cost.

Referring to fig. 2, based on the characteristics of the non-woven fabric production line, the steps of dividing production links and generating a partition monitoring map by adopting a computer vision technology and a spatial analysis algorithm are specifically as follows:

In step S101, a nonwoven fabric production line is initially divided using a computer vision technique. This process involves feature extraction algorithms and image segmentation algorithms for identifying critical areas and equipment of the production line. Through the algorithms, useful information can be extracted from the image data of the production line, and the production line is divided into different functional areas to form a preliminary partition model.

In step S102, the preliminary partition model is refined using a deep learning algorithm. The deep learning algorithm, and in particular the convolutional neural network, is used here to further learn and identify the hierarchical features and spatial relationships of the production line. By these advanced algorithms, each critical area and equipment on the production line can be more accurately divided and identified, resulting in a more refined and accurate partitioning scheme.

In step S103, a Geographic Information System (GIS) technique is applied to optimize the spatial layout of the refined partition scheme. The GIS technology is mainly applied to spatial database management and map visualization, and helps to more effectively manage and display the spatial layout of each region on a production line and optimize the spatial arrangement of a production flow.

In step S104, the spatial layout optimization design is integrated, and the partition monitor map is generated using the graphical user interface design tool. The application of the graphical user interface design tool comprises interface layout design and interactive design, so that the finally generated partition monitoring map is clear and understandable in vision, simple and visual in operation and convenient for daily monitoring and management of the production line.

Referring to fig. 3, based on the partition monitoring map, the steps of formulating the environment monitoring rule for multiple links and forming the environment monitoring rule set by adopting the expert system and rule reasoning technology are specifically as follows:

s204: based on the optimized environment monitoring rule set, integrating and formatting rules by using a knowledge management system to form the environment monitoring rule set;

In step S201, an expert system technology is used to investigate the environmental requirements in detail for each partition on the nonwoven fabric production line. This process involves building a knowledge base containing environmental criteria for each production link, such as temperature and humidity ranges, noise levels, etc. By the method, accurate environmental requirement data can be collected, and a solid foundation is laid for subsequent rule formulation.

In step S202, a rule reasoning technique is used to formulate a preliminary environmental monitoring rule based on the collected environmental requirement data. Through logical reasoning and pattern matching technology, key information can be extracted from environment requirement data, and a set of preliminary environment monitoring rule sets are formed.

In step S203, the set of preliminary environmental monitoring rules is validated and adjusted. This includes optimization of rules through simulation testing and feedback adjustment techniques. Simulation tests test the validity of rules by building virtual environments and scenes, and feedback adjustment adjusts and optimizes rules based on test results to ensure the accuracy and applicability of rule sets.

In step S204, the optimized environmental monitoring rule set is integrated and formatted by the knowledge management system. Applications of the knowledge management system include knowledge extraction and content management such that the resulting set of environmental monitoring rules is not only comprehensive but also easy to understand and implement.

Referring to fig. 4, based on an environment monitoring rule set, a monitoring device is deployed, and environmental data of multiple links is collected in real time, so that a real-time environment data set is generated specifically by the steps of:

S302: based on a monitoring equipment configuration scheme, equipment installation and network configuration are carried out, and a network connection stability test method is adopted to verify the connection stability of equipment and a central monitoring system and generate a monitoring network construction report;

s304: based on the real-time environment data flow, data integration and formatting are carried out, and a real-time environment data set is generated by adopting a data cleaning and preprocessing technology;

In step S301, appropriate monitoring devices are carefully selected according to the requirements of the environmental monitoring rule set, and the installation positions and methods thereof in each link of the production line are determined. In this process, the spatial layout of the monitoring device is optimized using spatial layout optimization algorithms, such as particle swarm optimization and genetic algorithms. By the aid of the method, monitoring equipment can cover all key links comprehensively, and monitoring efficiency and effectiveness can be improved. After optimization, an exhaustive monitoring equipment configuration scheme is generated.

In step S302, installation of devices and network configuration are performed according to the monitoring device configuration scheme. After installation, network connection stability testing methods, such as network delay testing and packet loss rate testing, are used to verify the stability of the connection between the monitoring device and the central monitoring system. This step is critical to ensure that the monitoring data is accurately and stably transmitted to the monitoring center. And after the test is finished, generating a monitoring network construction report.

In step S303, real-time data collection starts to be performed based on the monitoring network construction report. At this stage, real-time data transfer protocols, such as Message Queue Telemetry Transport (MQTT), are employed to verify the real-time and integrity of the data. This ensures that the data collected from the monitoring device can be transferred to the data processing system in a timely and complete manner, forming a real-time environmental data stream.

In step S304, data integration and formatting is performed based on the real-time environment data stream. This includes the application of data cleansing and preprocessing techniques such as outlier processing, outlier detection and data normalization to ensure accuracy and consistency of the data. After the processing is completed, a comprehensive and accurate real-time environment data set is generated, and a solid data base is provided for subsequent environment analysis and risk assessment.

Referring to fig. 5, based on a real-time environmental data set, the steps of analyzing environmental data change trend and generating an environmental trend analysis report by adopting time sequence analysis and association rule learning are specifically as follows:

s401: based on a real-time environment data set, adopting a time sequence analysis algorithm to analyze time dependence and trend in the data and generate a time sequence analysis result;

s403: synthesizing a time sequence analysis result and an association rule learning report, displaying a data change trend and correlation by adopting a data visualization technology, and generating an environment data visualization display;

s404: based on visual display of environmental data, adopting a report writing strategy to explain analysis results and significance of the analysis results on environmental monitoring, and generating an environmental trend analysis report;

In step S401, time-series analysis methods such as an autoregressive model (AR) and a moving average Model (MA) are applied to analyze time-dependence and trend in the data based on the real-time environmental data set. These timing analysis algorithms can reveal the laws of the environmental data over time, helping to understand the dynamic nature of the data. From these analyses, time series analysis results can be generated, which can reflect key trends and patterns in the production environment.

In step S402, based on the time series analysis result, an association rule learning algorithm, such as an Apriori algorithm and an FP-Growth algorithm, is used to further identify the association between the environmental factors. These algorithms can mine meaningful correlation rules from complex environmental data, revealing correlations between different environmental factors such as temperature, humidity, noise, etc. The resulting association rule learning report provides a deep perspective for understanding how the various factors interact in the production environment.

In step S403, the time series analysis result and the association rule learning report are integrated, and the data change trend and the correlation are intuitively displayed by using data visualization techniques such as scatter diagram, line diagram and thermodynamic diagram. The application of data visualization allows complex data and analysis results to be more easily understood and interpreted, helping to quickly identify key trends and potential problems.

In step S404, based on the visual presentation of the environmental data, a report writing strategy is adopted to explain the analysis result and its significance for environmental monitoring. This step includes data interpretation, trend prediction and impact assessment, ultimately forming a comprehensive environmental trend analysis report. This report not only details the analysis of the environmental data, but also provides predictions and impact assessment of future environmental trends.

Referring to fig. 6, based on the environmental trend analysis report, using a hidden markov model and a decision tree algorithm, the environmental risk assessment is performed, and the steps for generating the environmental risk assessment report are specifically as follows:

s501: based on the environmental trend analysis report, carrying out sequence analysis of the environmental state by adopting a hidden Markov model, identifying a potential risk mode, and generating a hidden Markov model analysis result;

s503: synthesizing a hidden Markov model analysis result and a decision tree risk assessment result, carrying out risk point analysis, locking a key monitoring area, and generating a risk point analysis report;

S504: based on the risk point analysis report, writing a comprehensive environmental risk assessment report, formulating a risk level and a countermeasure, and generating an environmental risk assessment report;

the hidden Markov model is specifically state transition probability analysis and observation probability analysis, the decision tree algorithm comprises information gain calculation, tree construction and pruning, and the risk point analysis comprises risk classification and key risk factor identification.

In step S501, by collecting and sorting historical data related to environmental trends, including environmental variables, event records, and the like, then building a hidden markov model, defining the likelihood of environmental states, analyzing transition probabilities and observation probability distributions between the states, finally using the hidden markov model to perform sequence analysis of the environmental states, identify potential risk patterns, and generate a hidden markov model analysis result.

In step S502, a data set including an environmental state and related risk factors is prepared, then an information gain calculation method is used to select an optimal risk factor, branches of a decision tree are constructed, the risk factors are divided into different subclasses, and pruning of the decision tree is performed when needed, so that a decision tree risk assessment result including tree structure and risk level allocation is finally generated.

In step S503, by defining different risk levels and identifying key risk factors, the hidden markov model analysis result and the decision tree risk assessment result are synthesized, and risk point analysis is performed to determine the region needing special attention, and finally a risk point analysis report is generated, including risk level classification and key risk factor identification.

In step S504, by writing a comprehensive environmental risk assessment report, integrating the hidden markov model analysis result, the decision tree risk assessment result and the risk point analysis result, making a risk level, classifying different risks into high, medium and low levels, providing corresponding countermeasures to reduce the risk of a high risk area, maintaining the sustainability of the environment, and finally generating a final environmental risk assessment report for relevant stakeholders to review and decide. These steps will ensure detailed implementation of the environmental risk assessment scheme, providing clear guidance and methods for environmental risk management.

Referring to fig. 7, based on the environmental risk assessment report, the monitoring parameters are automatically adjusted by adopting reinforcement learning and predictive maintenance strategies, and the steps for implementing the preventive maintenance strategies are specifically as follows:

s603: based on the monitoring parameter optimization report, planning a future monitoring scheme including equipment maintenance and parameter adjustment planning by adopting a predictive maintenance strategy to generate a preventive maintenance plan;

s604: based on the preventive maintenance planning, preventive maintenance measures are implemented, long-term stable operation of the monitoring system is maintained, and a preventive maintenance strategy is generated;

In step S601, based on the environmental risk assessment report, an automatic adjustment test of the monitoring parameters is performed using a reinforcement learning algorithm, an optimal parameter configuration is found, a reward function is designed, and based on the result of the risk assessment report, the risk minimization is taken as a main target.

Code example (assuming Q-learning algorithm is used):

import numpy as np

# assume a certain number of states and actions

n_states = 10

n_actions = 4

q_table = np.zeros((n_states, n_actions))

learning_rate = 0.1

discount_factor = 0.9

epsilon = 0.1

# environmental simulation function, return to next state and reward

def simulate_environment(state, action):

# Here, the environment simulation logic should be designed according to the actual situation

next_state = np.random.choice(n_states)

Reward= -risk_score # reward function assuming risk_score is a risk value obtained from the environmental risk assessment report

return next_state, reward

Training process of #Q-learning

for episode in range(1000):

state = np.random.choice(n_states)

while True:

if np.random.uniform(0, 1) < epsilon:

action = np.random.choice(n_actions)

else:

action = np.argmax(q_table[state, :])

next_state, reward = simulate_environment(state, action)

q_predict = q_table[state, action]

q_target = reward + discount_factor * np.max(q_table[next_state, :])

q_table[state, action] += learning_rate * (q_target - q_predict)

state = next_state

In step S602, automatic adjustment of the monitoring parameters is performed, the adjustment effect is evaluated by using an analog feedback evaluation method, new monitoring parameter configuration is tested by using an analog environment, and performance evaluation is performed.

Code example: setting a q_table optimized by Q-learning, and applying an optimal strategy to perform simulation test.

Select optimal action #

optimal_actions = np.argmax(q_table, axis=1)

Simulation test and evaluation of Performance

def simulate_test(state):

action = optimal_actions[state]

# return test results, e.g. performance scores

return performance_score

Test all states #)

performance_scores = [simulate_test(state) for state in range(n_states)]

In step S603, based on the monitoring parameter optimization report, a future monitoring scheme is planned by adopting a predictive maintenance strategy, and planning is performed by using a fault prediction model and a maintenance scheduling strategy.

In step S604, corresponding maintenance measures are implemented based on the preventive maintenance plan, the maintenance operation is performed, the performance is continuously monitored, and feedback is collected.

Referring to fig. 8, based on a preventive maintenance strategy, using a recurrent neural network and a reinforcement learning algorithm, the steps of optimizing and adjusting environmental parameters of multiple loops of a production line and establishing an optimized production environment configuration are specifically as follows:

s701: based on a preventive maintenance strategy, adopting a recurrent neural network to analyze historical production data and generating a key performance index analysis report;

s703: based on the optimized environment parameter configuration, applying a genetic algorithm to carry out multiparameter combination and optimization to generate a genetic optimization production environment configuration scheme;

s704: based on a genetic optimization production environment configuration scheme, adopting a simulated annealing algorithm to carry out fine adjustment, and generating an optimized production environment configuration;

the recurrent neural network comprises a long-term memory network and a gating circulation unit, the reinforcement learning algorithm comprises a Q learning and strategy gradient method, the genetic algorithm comprises crossover, mutation and selection operations, and the simulated annealing algorithm comprises a cooling plan design and a neighborhood searching strategy.

In step S701, by performing historical production data analysis using a recurrent neural network, first, historical production data including environmental parameters, performance indexes, maintenance records, and the like of each link is prepared. The historical data is then modeled using recurrent neural networks, such as long short term memory networks (LSTM) and gated loop units (GRUs), to generate key performance indicator analysis reports, including identification and extraction of key performance indicators. This step helps to understand the historical trends and performance issues of the production environment.

In step S702, real-time environmental parameter optimization is performed using a reinforcement learning algorithm by analyzing the report based on the key performance indicators. First, an environmental model is built, including the state and action space of environmental parameters, and an appropriate reward function is defined. Then, training is performed using reinforcement learning algorithms, such as Q-learning or strategy gradient methods, to achieve optimization of real-time environmental parameters. Finally, generating optimized environment parameter configuration to improve production performance and reduce fault risk, and supporting real-time optimization of the production process.

In step S703, a genetic algorithm is applied for multiparameter combining and optimization by configuring based on the optimized environmental parameters. Firstly, modeling a production environment configuration problem as a multi-parameter optimization problem, and definitely defining a value range and an optimization target of each parameter. The multiparameter combinations are then searched using genetic algorithms, including crossover, mutation and selection operations, to generate a set of potential optimal production environment configurations. This step helps explore the potential optimization space of the various configurations.

In step S704, fine tuning is performed by using a simulated annealing algorithm based on the genetically optimized production environment configuration scheme, and finally an optimized production environment configuration is generated. First, an optimized environmental model simulating an annealing algorithm is established, including initial configuration, energy function (fitness function), and cooling plan design. Then, a simulated annealing algorithm is applied, and the performance of the different configurations is evaluated using a fitness function to find the optimal production environment configuration. Ultimately, an optimal configuration is generated to ensure optimal performance and sustainable production. This step facilitates fine tuning and further improves the performance of the production environment.

Referring to fig. 9, the system for monitoring the production environment of a nonwoven fabric product is used for executing the method for monitoring the production environment of a nonwoven fabric product, and comprises a production link dividing module, an environment monitoring rule making module, a monitoring equipment deployment module, an environment data analysis module, an environment risk assessment module and a preventive maintenance implementation module.

The production link dividing module divides production links based on the characteristics of a non-woven fabric production line by adopting a computer vision technology and a space analysis algorithm and generates a partition monitoring map;

the environment data analysis module analyzes the data change trend and generates an environment trend analysis report by adopting time sequence analysis and association rule learning based on the real-time environment data set of the monitoring equipment deployment module;

The environmental risk assessment module is used for carrying out risk assessment by using a hidden Markov model and a decision tree algorithm based on an environmental trend analysis report of the environmental data analysis module and generating an environmental risk assessment report;

the preventive maintenance implementation module automatically adjusts the monitoring parameters based on the environmental risk assessment report of the environmental risk assessment module using reinforcement learning and predictive maintenance strategies and generates an optimized production environment configuration.

The production link dividing module precisely divides production links through a computer vision technology and a space analysis algorithm to generate a partition monitoring map. The division greatly improves the accuracy and efficiency of monitoring, and ensures that each key link in the production process is fully concerned and monitored. This precise partitioning helps identify critical control points in the manufacturing process, providing a solid foundation for subsequent monitoring and maintenance.

The environment monitoring rule making module adopts expert system and rule reasoning technology to make specific environment monitoring rule and generate environment monitoring rule set. Such expert systems fully take into account the particularities and complexity of the production environment during rule formulation, thereby ensuring that the rule set is both comprehensive and targeted. The establishment of the rules is helpful for timely finding and coping with potential risks in the production link, and improves the production safety and stability.

By implementing the monitoring equipment deployment module and deploying advanced monitoring equipment, real-time environment data of a production link are effectively acquired. The real-time data provides real-time and accurate information support for subsequent analysis and decision making, so that the production environment can be monitored in real time, and potential problems can be found and dealt with in time.

The environmental data analysis module uses time sequence analysis and association rule learning to conduct deep analysis on the collected environmental data. Such analysis helps to identify trends and potential correlation patterns of environmental data, providing data support for the adjustment and optimization of production environments.

The application of the environment risk assessment module carries out risk assessment on the environment trend analysis result by using a hidden Markov model and a decision tree algorithm. The assessment method can accurately identify key risk points and potential risk factors in the production environment, and provides basis for formulating effective risk countermeasures.

The preventive maintenance implementation module automatically adjusts the monitoring parameters through reinforcement learning and predictive maintenance strategies to realize the preventive maintenance strategies. The intelligent level of the monitoring system is improved, the production environment configuration is optimized in an automatic mode, and the stability and efficiency of the whole production process are improved.

The present invention is not limited to the above embodiments, and any equivalent embodiments which can be changed or modified by the technical disclosure described above can be applied to other fields, but any simple modification, equivalent changes and modification made to the above embodiments according to the technical matter of the present invention will still fall within the scope of the technical disclosure.

Claims

1. The method for monitoring the production environment of the non-woven fabric product is characterized by comprising the following steps of:

based on the characteristics of a non-woven fabric production line, carrying out production link division by adopting a computer vision technology and a space analysis algorithm, and generating a partition monitoring map;

based on the subarea monitoring map, adopting an expert system and a rule reasoning technology to formulate an environment monitoring rule aiming at multiple links and form an environment monitoring rule set;

based on the environment monitoring rule set, deploying monitoring equipment, and collecting environment data of multiple links in real time to generate a real-time environment data set;

based on the real-time environment data set, adopting time sequence analysis and association rule learning to analyze the change trend of the environment data and generate an environment trend analysis report;

Based on the environmental trend analysis report, performing environmental risk assessment by using a hidden Markov model and a decision tree algorithm to generate an environmental risk assessment report;

based on the environmental risk assessment report, adopting reinforcement learning and predictive maintenance strategies to automatically adjust monitoring parameters so as to realize preventive maintenance strategies;

based on the preventive maintenance strategy, optimizing and adjusting environmental parameters of multiple loops of the production line by using a recurrent neural network and a reinforcement learning algorithm, and establishing optimized production environment configuration;

2. The method for monitoring the production environment of a non-woven fabric product according to claim 1, wherein the steps of dividing production links and generating a partition monitoring map are specifically performed by adopting a computer vision technology and a spatial analysis algorithm based on the characteristics of a non-woven fabric production line:

based on a non-woven fabric production line, performing preliminary division by adopting a computer vision technology to generate a preliminary partition model;

based on the preliminary partition model, carrying out refinement space identification and division by using a deep learning algorithm to form a refinement partition scheme;

based on the refined partition scheme, the geographic information system technology is applied to optimize the space layout, and the space layout optimization design is completed;

integrating the space layout optimization design, and generating a partition monitoring map by using a graphical user interface design tool;

3. The method for monitoring the production environment of a nonwoven fabric product according to claim 1, wherein the steps of formulating environmental monitoring rules for multiple links and forming an environmental monitoring rule set based on the partition monitoring map by adopting an expert system and rule reasoning technology are specifically as follows:

based on the partition monitoring map, collecting environmental requirements of a production link by adopting an expert system technology, and collecting environmental requirement data;

based on the environmental requirement data, formulating a preliminary environmental monitoring rule by using a rule reasoning technology to form a preliminary environmental monitoring rule set;

verifying and adjusting the preliminary environment monitoring rule set, and establishing an optimized environment monitoring rule set by adopting a simulation test and feedback adjustment technology;

based on the optimized environment monitoring rule set, integrating and formatting rules by using a knowledge management system to form an environment monitoring rule set;

4. The method for monitoring the production environment of a nonwoven fabric product according to claim 1, wherein based on the environment monitoring rule set, monitoring equipment is deployed, environmental data of multiple links are collected in real time, and the step of generating a real-time environment data set specifically comprises the following steps:

Based on the environment monitoring rule set, selecting monitoring equipment, determining a monitoring position and an installation method, and performing equipment layout optimization by adopting a spatial layout optimization algorithm to generate a monitoring equipment configuration scheme;

based on the monitoring equipment configuration scheme, equipment installation and network configuration are carried out, and a network connection stability test method is adopted to verify the connection stability of equipment and a central monitoring system and generate a monitoring network construction report;

based on the monitoring network construction report, data acquisition is implemented, and a real-time data transmission protocol is adopted to verify the real-time performance and integrity of the data, so as to generate a real-time environment data stream;

based on the real-time environment data stream, data integration and formatting are carried out, and a real-time environment data set is generated by adopting a data cleaning and preprocessing technology;

5. The method for monitoring the production environment of a nonwoven fabric product according to claim 1, wherein the steps of analyzing the trend of environmental data and generating an environmental trend analysis report based on the real-time environmental data set by using time series analysis and association rule learning are specifically as follows:

Based on the real-time environment data set, adopting a time sequence analysis algorithm to analyze time dependence and trend in the data and generate a time sequence analysis result;

based on the time sequence analysis result, adopting an association rule learning algorithm to identify the association between environmental factors and generate an association rule learning report;

synthesizing the time sequence analysis result and the association rule learning report, and displaying the data change trend and the correlation by adopting a data visualization technology to generate environment data visualization display;

based on the visual display of the environmental data, adopting a report writing strategy to explain the analysis result and the meaning of the analysis result on environmental monitoring, and generating an environmental trend analysis report;

6. The method according to claim 1, wherein the step of performing an environmental risk assessment using a hidden markov model and a decision tree algorithm based on the environmental trend analysis report, and generating an environmental risk assessment report is specifically:

Based on the environmental trend analysis report, carrying out sequence analysis of environmental states by adopting a hidden Markov model, identifying a potential risk mode, and generating a hidden Markov model analysis result;

based on the hidden Markov model analysis result, classifying and evaluating risk factors by adopting a decision tree algorithm to generate a decision tree risk evaluation result;

synthesizing the hidden Markov model analysis result and the decision tree risk assessment result, carrying out risk point analysis, locking a key monitoring area and generating a risk point analysis report;

based on the risk point analysis report, writing a comprehensive environmental risk assessment report, formulating a risk level and countermeasures, and generating an environmental risk assessment report;

7. The method for monitoring the production environment of a nonwoven fabric product according to claim 1, wherein based on the environmental risk assessment report, a reinforcement learning and predictive maintenance strategy is adopted to automatically adjust monitoring parameters, and the step of implementing the preventive maintenance strategy is specifically as follows:

Based on the environmental risk assessment report, performing an automatic adjustment test of monitoring parameters by using a reinforcement learning algorithm, searching for optimal parameter configuration, and generating a reinforcement learning adjustment scheme;

based on the reinforcement learning adjustment scheme, implementing automatic adjustment of the monitoring parameters, and adopting an analog feedback evaluation method to evaluate the adjustment effect and generate a monitoring parameter optimization report;

based on the monitoring parameter optimization report, planning a future monitoring scheme by adopting a predictive maintenance strategy, including equipment maintenance and parameter adjustment planning, and generating a preventive maintenance plan;

based on the preventive maintenance plan, preventive maintenance measures are implemented, long-term stable operation of the monitoring system is maintained, and a preventive maintenance strategy is generated;

8. The method for monitoring the production environment of a nonwoven fabric product according to claim 1, wherein based on the preventive maintenance strategy, the steps of optimizing and adjusting environmental parameters of multiple loops of a production line by using a recurrent neural network and a reinforcement learning algorithm, and establishing an optimized production environment configuration are specifically as follows:

Based on the preventive maintenance strategy, adopting a recurrent neural network to analyze historical production data and generating a key performance index analysis report;

based on the key performance index analysis report, performing real-time environmental parameter optimization by using a reinforcement learning algorithm to generate optimized environmental parameter configuration;

based on the optimized environment parameter configuration, applying a genetic algorithm to carry out multi-parameter combination and optimization to generate a genetic optimization production environment configuration scheme;

based on the genetic optimization production environment configuration scheme, adopting a simulated annealing algorithm to carry out fine adjustment to generate an optimized production environment configuration;

9. The non-woven fabric product production environment monitoring system is characterized in that the non-woven fabric product production environment monitoring method according to any one of claims 1-8 comprises a production link dividing module, an environment monitoring rule making module, a monitoring equipment deployment module, an environment data analysis module, an environment risk assessment module and a preventive maintenance implementation module.

10. The system for monitoring the production environment of the non-woven fabric product according to claim 9, wherein the production link dividing module divides the production links and generates a partition monitoring map by adopting a computer vision technology and a space analysis algorithm based on the characteristics of a non-woven fabric production line;