Manufacturing equipment monitoring method based on Internet of Things

Abstract

The invention relates to the technical field of equipment monitoring, in particular to a manufacturing equipment monitoring method based on the Internet of Things, which comprises the following steps of: based on a manufacturing equipment field environment, collecting multi-source data of manufacturing equipment through the Internet of Things, and constructing an equipment data heterogeneous network; according to the invention, through the constructed equipment data heterogeneous network, the operation state and environmental factors of the equipment can be comprehensively reflected, the equipment state can be analyzed more accurately, the process parameters can be predicted and adjusted in advance, the optimization of the equipment performance and the prevention of faults can be realized, the optimization analysis of the process parameters can be carried out, the depth and the breadth of the analysis can be enhanced, and the analysis efficiency can be improved. The technical parameter optimization strategy is more scientific and reasonable, the decision tree classification algorithm and the weighted average algorithm are adopted, the health state of the equipment is subjected to hierarchical evaluation and comprehensive health index calculation, the evaluation accuracy is improved, and a more effective maintenance optimization strategy can be made according to the actual influence degree of the equipment performance.

Description

Technical Field

[0001] The present invention relates to the technical field of equipment monitoring, and in particular to a manufacturing equipment monitoring method based on the Internet of Things. Background Art

[0002] The field of equipment monitoring technology involves real-time or periodic detection, analysis and management of the performance, status and working environment of manufacturing equipment to ensure the normal operation of the equipment, prevent failures, and optimize the working efficiency of the equipment. In modern manufacturing, equipment monitoring technology is particularly important, which is directly related to production efficiency, product quality, cost control, and equipment maintenance and life management. With the development of Internet of Things technology, the field of equipment monitoring has shifted from traditional manual detection and local network monitoring to more complex networked, intelligent and automated monitoring systems. By using sensors, communication technology and data analysis tools, data from many devices can be collected and processed in real time across regions to provide support for equipment management and decision-making.

[0003] Among them, the IoT-based manufacturing equipment monitoring method is a manufacturing equipment monitoring solution implemented using IoT technology. The core purpose of this method is to optimize equipment performance, prevent failures, and improve maintenance efficiency by collecting equipment operation data in real time and analyzing equipment status. This method aims to improve equipment utilization, reduce unexpected downtime, extend equipment life, and improve the efficiency and product quality of the entire manufacturing process through monitoring and analysis. Through the application of IoT technology, remote monitoring of equipment is achieved, integrated monitoring of multiple devices and across locations is supported, and the flexibility and scalability of the monitoring system are enhanced.

[0004] Traditional equipment monitoring methods lack effective integration and in-depth analysis capabilities when processing complex, multi-source equipment data. Traditional monitoring methods lack sufficient data support and intelligent analysis tools in the adjustment of process parameters and the formulation of equipment maintenance strategies, resulting in untimely maintenance responses or inaccurate strategies, affecting production efficiency and product quality. For example, if the health status of the equipment is not accurately assessed, maintenance may be performed too early or too late, increasing unnecessary costs or causing failures. Summary of the invention

[0005] The purpose of the present invention is to solve the shortcomings of the prior art and to propose a manufacturing equipment monitoring method based on the Internet of Things.

[0006] In order to achieve the above object, the present invention adopts the following technical solution: a manufacturing equipment monitoring method based on the Internet of Things, comprising the following steps:

[0007] S1: Based on the on-site environment of manufacturing equipment, multi-source data of manufacturing equipment is collected through the Internet of Things to build a heterogeneous network of equipment data;

[0008] S2: Based on the device data heterogeneous network, a node sequence is sampled by random walk, a vector representation is generated for each node and edge, and a network feature embedding vector is obtained;

[0009] S3: Based on the network feature embedding vector, the process parameters are optimized and analyzed, the node representation is updated by aggregating the information of the node and the adjacent nodes, the relationship between the nodes is captured, and the process parameters are predicted and adjusted using the updated node representation to generate a process parameter optimization strategy;

[0010] S4: Based on the operating data of the manufacturing equipment, the status of each component is evaluated layer by layer. The decision tree classification algorithm is used to analyze the data to determine the health status of each component and obtain the hierarchical equipment status evaluation results;

[0011] S5: Based on the hierarchical equipment status evaluation results, a weighted average algorithm is used to calculate the comprehensive health index according to the degree of influence of the differentiated levels on the equipment performance, and an equipment maintenance optimization strategy is formulated according to the change trend of the health index and the preset threshold;

[0012] S6: Implement the process parameter optimization strategy and equipment maintenance optimization strategy, evaluate the equipment performance after maintenance adjustment, and obtain continuous optimization feedback data.

[0013] The present invention has the following improvements: the equipment data heterogeneous network includes sensor nodes, operation log nodes, maintenance record nodes and interaction relationships between nodes; the network feature embedding vector includes the spatial position vector of the node and the strength index of the relationship between nodes; the process parameter optimization strategy includes the adjusted temperature setting, pressure level, and operation sequence; the hierarchical equipment status evaluation results include the degree of component wear, equipment system operating efficiency, and potential failure points; the equipment maintenance optimization strategy includes maintenance time, component replacement priority, and maintenance operation process; the continuous optimization feedback data includes equipment performance comparison information before and after parameter adjustment, optimization measure effect evaluation information, and subsequent optimization strategy.

[0014] The present invention is improved in that, based on the device data heterogeneous network, a node sequence is sampled by random walk, a vector representation is generated for each node and edge, and the specific steps of obtaining the network feature embedding vector are as follows:

[0015] S201: Based on the device data heterogeneous network, perform sequence sampling on nodes in the network, record the access order of the nodes, simulate the logical flow and interaction between the data, and obtain a node sequence sample set;

[0016] S202: Based on the node sequence sample set, analyzing the roles and functions of differentiated nodes in equipment operation and maintenance activities, identifying context associations between nodes, and constructing a node context relationship set;

[0017] S203: Based on the node context relationship set, a corresponding vector value is assigned to each node and the edge between nodes, and the vector reflects the characteristics of the node and the interaction mode between the nodes, including the functional category of the node and the interaction intensity between the nodes, to obtain a network feature embedding vector.

[0018] The present invention is improved in that, based on the network feature embedding vector, the process parameters are optimized and analyzed, the node representation is updated by aggregating the information of the node and the adjacent nodes, the relationship between the nodes is captured, and the process parameters are predicted and adjusted using the updated node representation. The specific steps of generating the process parameter optimization strategy are as follows:

[0019] S301: Based on the network feature embedding vector, the vector is sorted and analyzed, and the distribution and role of the nodes in the network are sorted out through analysis to reflect the key operation points of the equipment, so as to build a node role and information flow framework;

[0020] S302: Based on the node roles and information flow framework, analyzing the interactions between key nodes and adjacent nodes, identifying the interaction patterns between key nodes and adjacent nodes, and obtaining node interaction identification results;

[0021] S303: Based on the node interaction identification result, information aggregation is performed, and the node representation is updated to reflect the current state and function of the node. According to the updated node, the process parameters of the manufacturing equipment are predicted and adjusted to generate a process parameter optimization strategy.

[0022] The present invention is improved in that the state of each component is evaluated layer by layer based on the operation data of the manufacturing equipment, and the decision tree classification algorithm is used to judge the health state of each component by analyzing the data. The specific steps of obtaining the hierarchical equipment state evaluation result are as follows:

[0023] S401: Based on the operation data of the manufacturing equipment, identify various basic components of the manufacturing equipment and related performance indicators, including component wear degree and efficiency indicators, and obtain component performance indicator information;

[0024] S402: Based on the identified component performance indicators, a decision tree classification algorithm is used to analyze the data of each component, and the component health status is evaluated by comparing the historical performance data with the current operating status to obtain component status information;

[0025] S403: Based on the component status information, the health status assessment result of each component is aggregated from the bottom layer to the top layer by layer, and the status assessment results of multiple components are summarized and integrated to obtain a hierarchical device status assessment result.

[0026] The present invention is improved in that the decision tree classification algorithm is according to the formula:

[0027]

[0028] Calculate the entropy value of component operation data and evaluate the health status of components, where E′(S) is the entropy value of set S, S is the operation data set of multiple components, and p i is the original proportion of the i-th state in the set S, Δp i is the adjustment value of the i-th state ratio, c i The weight coefficient of the i-th state, n is the total number of states.

[0029] The present invention is improved in that, based on the hierarchical equipment status evaluation results, a weighted average algorithm is used to calculate the comprehensive health index according to the degree of influence of the differentiated levels on the equipment performance, and the specific steps of formulating the equipment maintenance optimization strategy according to the change trend of the health index and the preset threshold are as follows:

[0030] S501: Based on the hierarchical device status evaluation result, identify the influence degree of each hierarchical component on the device performance, assign a weight value to each hierarchical component, and obtain influence weight setting information;

[0031] S502: Based on the weight setting information, a weighted average algorithm is used to perform weighted calculation on the health status of each level, calculate the overall health status index of the manufacturing equipment, reflect the current operating status of the manufacturing equipment, and obtain a comprehensive health index;

[0032] S503: Based on the comprehensive health index, the preset manufacturing equipment operation threshold is compared to identify the equipment maintenance and optimization needs, and a matching equipment maintenance adjustment plan is formulated, including adjusting process parameters and planning component replacement, to obtain an equipment maintenance optimization strategy.

[0033] The present invention is improved in that the weighted average algorithm is according to the formula:

[0034]

[0035] Calculate the comprehensive health index, where H′ is the improved comprehensive health index, w i is the weight of the i-th level, h i is the health status of the i-th level, a i is the adjustment coefficient of the ith level, and n is the total number of levels.

[0036] Compared with the prior art, the advantages and positive effects of the present invention are:

[0037] In the present invention, by constructing a heterogeneous network of equipment data, it is possible to comprehensively reflect the operating status and environmental factors of the equipment, analyze the equipment status more accurately, predict and adjust the process parameters in advance, optimize the equipment performance and prevent faults, and perform optimization analysis of the process parameters, thereby enhancing the depth and breadth of the analysis, making the process parameter optimization strategy more scientific and reasonable. The decision tree classification algorithm and the weighted average algorithm are used to perform a hierarchical evaluation of the health status of the equipment and calculate the comprehensive health index, which not only improves the accuracy of the evaluation, but also can formulate a more effective maintenance optimization strategy based on the actual degree of impact on the equipment performance, making the monitoring, maintenance and optimization of manufacturing equipment more efficient and accurate, and significantly improving the efficiency of the manufacturing process and product quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 A flowchart of a manufacturing equipment monitoring method based on the Internet of Things is proposed for the present invention;

[0039] Figure 2 The present invention proposes a detailed flow chart of step S1 in a manufacturing equipment monitoring method based on the Internet of Things;

[0040] Figure 3 The present invention proposes a detailed flow chart of step S2 in a manufacturing equipment monitoring method based on the Internet of Things;

[0041] Figure 4 The present invention proposes a detailed flow chart of step S3 in a manufacturing equipment monitoring method based on the Internet of Things;

[0042] Figure 5 The present invention proposes a detailed flow chart of step S4 in a manufacturing equipment monitoring method based on the Internet of Things;

[0043] Figure 6 The present invention proposes a detailed flow chart of step S5 in a manufacturing equipment monitoring method based on the Internet of Things;

[0044] Figure 7 The present invention proposes a detailed flow chart of step S6 in a manufacturing equipment monitoring method based on the Internet of Things. DETAILED DESCRIPTION

[0045] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0046] In the description of the present invention, it should be understood that the terms "length", "width", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", etc., indicating positions or positional relationships, are based on the positions or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, in the description of the present invention, the meaning of "multiple" is two or more, unless otherwise clearly and specifically defined.

[0047] Example

[0048] See also Figure 1 The present invention provides a technical solution: a manufacturing equipment monitoring method based on the Internet of Things, comprising the following steps:

[0049] S1: Based on the on-site environment of manufacturing equipment, the multi-source data of manufacturing equipment is collected through the Internet of Things, including sensor data, operation logs, and maintenance records. The elements in the data source are defined as nodes in the network, and the correlation between data is defined as edges between nodes to build a heterogeneous network of equipment data;

[0050] S2: Based on the heterogeneous network of device data, we sample node sequences through random walks, learn the context information of nodes, generate vector representations for each node and edge, capture the interactions between nodes and edges of different types, and obtain network feature embedding vectors;

[0051] S3: Based on the network feature embedding vector, the optimization analysis of process parameters is carried out. By aggregating the information of nodes and adjacent nodes, the representation of nodes is updated, the relationship between nodes is captured, and the updated node representation is used to predict and adjust the process parameters to generate a process parameter optimization strategy.

[0052] S4: Based on the operating data of manufacturing equipment, starting from the basic components of manufacturing equipment, the status of each component is evaluated layer by layer. The decision tree classification algorithm is used to analyze the data to determine the health status of each component and obtain the hierarchical equipment status evaluation results;

[0053] S5: Based on the hierarchical equipment status assessment results, a weighted average algorithm is used to calculate the comprehensive health index according to the degree of influence of the differentiated levels on equipment performance. Based on the change trend of the health index and the preset threshold, an equipment maintenance optimization strategy is formulated;

[0054] S6: Implement process parameter optimization strategies and equipment maintenance optimization strategies, adjust the process parameters of manufacturing equipment, monitor the implementation effects of adjustment and maintenance activities, evaluate the performance of equipment after maintenance and adjustment, and obtain continuous optimization feedback data.

[0055] The equipment data heterogeneous network includes sensor nodes, operation log nodes, maintenance record nodes and the interaction relationship between nodes. The network feature embedding vector includes the spatial position vector of the node and the strength index of the relationship between nodes. The process parameter optimization strategy includes the adjusted temperature setting, pressure level, and operation sequence. The hierarchical equipment status assessment results include the degree of component wear, equipment system operation efficiency, and potential failure points. The equipment maintenance optimization strategy includes maintenance time, component replacement priority, and maintenance operation process. The continuous optimization feedback data includes the equipment performance comparison information before and after parameter adjustment, the effect evaluation information of optimization measures, and the subsequent optimization strategy.

[0056] See also Figure 2 ,Based on the on-site environment of manufacturing equipment, the multi-source data of manufacturing equipment,including sensor data, operation logs, and maintenance records, are collected through the Internet of Things.,The elements in the data source are defined as nodes in the network, and the association between data is defined as the edge between nodes.,The specific steps of building a heterogeneous network of equipment data are as follows:

[0057] S101: Based on the on-site environment of manufacturing equipment, IoT sensors are deployed to collect data, including temperature data, pressure data, equipment operation logs, maintenance and fault records of manufacturing equipment. The specific process of obtaining the original data set is as follows;

[0058] In sub-step S101, based on the on-site environment of manufacturing equipment, temperature and pressure sensors are deployed through the Internet of Things technology to collect data. The Python Pandas library is used to format the data, integrate temperature data, pressure data, equipment operation logs, and maintenance fault records, and store the collected data in the SQLite database to ensure data organization and archiving, and generate the original data set.

[0059] S102: Based on the original data set, identify data elements, including sensor identification and log timestamp, define the elements as network nodes, and define the logical and physical connections between data elements as edges between nodes. The specific process of building the device data network structure is as follows;

[0060] In sub-step S102, based on the original data set, the device data network is constructed using Python's NetworkX library. The data elements are identified as network nodes through the functions provided by Ne tworkX. The logical and physical connections between data elements are defined as edges between nodes using node attributes such as sensor identification and log timestamps. These connections are automatically established through the graph construction function of NetworkX to construct a device data network structure that reflects the device status and activity.

[0061] S103: Based on the device data network structure, the correlation between data elements is analyzed, including the time relationship between operation logs and maintenance records, and the definition of edges between nodes is optimized. The specific process of obtaining the device data heterogeneous network is as follows;

[0062] In sub-step S103, based on the network structure of equipment data, Python's Scikit-learn library is used to perform correlation analysis between data elements. The edge definition between nodes is optimized through the correlation coefficient calculation method in Scikit-learn, and the time relationship between equipment operation logs and maintenance records is enhanced. The clustering algorithm of Scikit-learn is used to identify and optimize the relationship between nodes. The strength of the interaction between nodes is expressed by adjusting the weight of the edge. The graph algorithm provided by Scikit-learn, such as the spectral clustering algorithm, is used to analyze the network structure, refine the network connection mode, and obtain a heterogeneous network of equipment data.

[0063] See also Figure 3 ,Based on the heterogeneous network of device data, we sample node sequences through random walks, learn the context information of nodes, generate vector representations for each node and edge, capture the interactions between nodes and edges of differentiated types, and obtain the network feature embedding vectors in the following specific steps:

[0064] S201: Based on the device data heterogeneous network, perform sequence sampling on the nodes in the network, record the access order of the nodes, simulate the logical flow and interaction between the data, and obtain the specific process of the node sequence sample set as follows;

[0065] In sub-step S201, node sequence sampling is performed based on the heterogeneous network of device data. The random walk algorithm is used to set the step size and sampling times to simulate the logical flow and interaction between data. Through Python's NetworkX library, the random_walk function is defined. The parameters include the network graph object, the starting node, the step size and the sampling times. The node to be visited next is randomly selected, the node visit order is recorded, and a node sequence sample set is generated.

[0066] S202: Based on the node sequence sample set, the roles and functions of the differentiated nodes in the equipment operation and maintenance activities are analyzed, the context associations between the nodes are identified, and the specific process of constructing the node context relationship set is as follows;

[0067] In sub-step S202, based on the node sequence sample set, differentiated node role and function analysis is performed, natural language processing technology is adopted, and Python's NLTK library is used to analyze the context association in the node sequence. By defining the context_analysis function, the parameters include the node sequence sample set and the window size for determining the node context range, and the word2vec model is used to learn the node context relationship. The model parameters, such as vector size and window size, are set to build a context relationship set for each node.

[0068] S203: Based on the node context relationship set, a corresponding vector value is assigned to each node and the edge between nodes. The vector reflects the characteristics of the node and the interaction mode between nodes, including the functional category of the node and the interaction strength between nodes. The specific process of obtaining the network feature embedding vector is as follows;

[0069] In sub-step S203, vector allocation is performed based on the node context relationship set. A graph embedding algorithm is adopted using Python's Gensim library. The graph_embedding function is defined, and the parameters include the node context relationship set and the vector dimension. The model is trained using the Node2Vec algorithm. The model training parameters include the dimension size, walk length, and window size. Vector values ​​are assigned to each node and the edge between nodes. The vector values ​​reflect the functional category of the node and the interaction pattern between nodes, including the node functional category vector and the interaction strength index between nodes, to obtain the network feature embedding vector.

[0070] See also Figure 4 ,Based on the network feature embedding vector, the optimization analysis of process parameters is carried out.,By aggregating the information of nodes and adjacent nodes, the representation of nodes is updated, the relationship between nodes is captured, and the updated node representation is used to predict and adjust the process parameters.,The specific steps for generating the process parameter optimization strategy are as follows:

[0071] S301: Based on the network feature embedding vector, the vector is sorted and analyzed, and the distribution and role of the nodes in the network are sorted out through analysis to reflect the key operation points of the equipment, and the specific process of building the node role and information flow framework is as follows;

[0072] In sub-step S301, vectors are organized and analyzed based on the network feature embedding vectors. The principal component analysis method is adopted and Python's Scikit-learn library is used. The parameter settings include the number of components retained to reflect the key operating points of the equipment. The feature vectors are reduced in dimension through the pca.fit_transform method to sort out the distribution and role of nodes in the network and build a node role and information flow framework.

[0073] S302: Based on the node role and information flow framework, the interaction between the key nodes and the adjacent nodes is analyzed, and the interaction mode between the key nodes and the adjacent nodes is identified. The specific process of obtaining the node interaction identification result is as follows;

[0074] In sub-step S302, based on the node roles and information flow framework, the key node interaction pattern analysis is performed, and the clustering algorithm is applied. The K-means method is used. Through the Python Scikit-learn library, the number of cluster centers is set equal to the pre-defined number of key node categories. According to the node roles and information flow framework, the interaction pattern between the key nodes and the adjacent nodes is identified to obtain the node interaction identification result.

[0075] S303: Based on the node interaction identification results, information aggregation is performed, and the node representation is updated to reflect the current state and function of the node. According to the updated node, the process parameters of the manufacturing equipment are predicted and adjusted. The specific process of generating the process parameter optimization strategy is as follows;

[0076] In sub-step S303, based on the node interaction recognition results, information aggregation and node representation update are performed. A graph convolutional network model is adopted, and the model is built using Python's PyTorch-Geometric library. Model definition includes adding a graph convolutional layer to aggregate node information. Parameter settings include the output dimension and activation function type of each layer. The representation of each node is updated using the graph convolutional network model to reflect the current state and function of the node. Based on the updated node representation, a linear regression model is applied to predict and adjust the process parameters of the manufacturing equipment. Linear regression is implemented through Python's Scikit-learn library. Parameter settings include the regularization strength of the fitting model, and a process parameter optimization strategy is generated.

[0077] See also Figure 5 , based on the operating data of manufacturing equipment, starting from the basic components of manufacturing equipment, the status of each component is evaluated layer by layer, and the decision tree classification algorithm is used to judge the health status of each component by analyzing the data. The specific steps to obtain the hierarchical equipment status evaluation results are as follows:

[0078] S401: Based on the operation data of the manufacturing equipment, identify various basic components of the manufacturing equipment and related performance indicators, including component wear degree and efficiency indicators, and obtain the specific process of component performance indicator information as follows;

[0079] In sub-step S401, component performance indicators are identified based on the operating data of the manufacturing equipment. The time series analysis method is adopted, and the Python Pandas library is used to process the operating data. By defining the time_series_analysis function, the parameters include the data frame, the timestamp column, and the indicator column. The sliding window technology is used to calculate the time series changes of the component wear degree and the efficiency index. For each basic component and its associated performance indicators, including the component wear degree and the efficiency index, the calculated index value is used as the numerical representation of the performance index to obtain the component performance index information.

[0080] S402: Based on the identified component performance indicators, a decision tree classification algorithm is used to analyze the data of each component, and the component health status is evaluated by comparing the historical performance data with the current operating status. The specific process of obtaining the component status information is as follows;

[0081] In sub-step S402, the component health status is evaluated based on the identified component performance indicators, and a decision tree classification algorithm is applied. This is implemented through Python's Scikit-learn library, and the depth of the decision tree and the minimum number of leaf node samples are set as parameters. By comparing the historical performance data set and the current operating status, the health status of each component is evaluated, including normal, warning, fault and other states, to obtain component status information.

[0082] Decision tree classification algorithm, according to the formula:

[0083]

[0084] Calculate the entropy value of component operation data and evaluate the health status of components, where E′(S) is the entropy value of set S, which represents the uncertainty measure of set S after considering state changes and weights. S is the operation data set of multiple components, where the data of each component contains performance indicators and health status labels, p i is the original proportion of the i-th state in the set S, indicating the proportion of this state in all states, Δp i is the adjustment value of the i-th state ratio, reflecting the latest change of the state ratio or the expected change under specific conditions, c i The weight coefficient of the i-th state is determined according to the importance of the state or the degree of impact on the health of the device. n is the total number of states, which represents the number of different states considered in the set S.

[0085] The specific implementation process of the improved formula is as follows:

[0086] For each state i, calculate its original proportion p in the set S i , set the adjustment value Δp according to the specific characteristics of the component or changes in external conditions i, to reflect the recent changes or trends of the components, and set the weight coefficient c for each state i The coefficient can be determined based on the severity, frequency, or impact of the condition on the overall performance of the equipment. For example, the c of a fault condition i higher than the normal state to reflect its significant impact on the health of the equipment. i Add the adjustment value Δp i , and multiply it by the corresponding weight coefficient c i , and then calculate the weighted entropy value E′(S), which not only improves the accuracy of the evaluation, but also increases the consideration of the importance of different states, making the entropy value more representative and practical. The entropy value E′(S) is used to select the split point in the decision tree algorithm to optimize the evaluation of the component health status.

[0087] S403: Based on the component status information, the health status assessment result of each component is aggregated from the bottom layer to the top layer by layer, and the status assessment results of multiple components are summarized and integrated to obtain the hierarchical equipment status assessment result. The specific process is as follows;

[0088] In sub-step S403, based on the component status information, the hierarchical equipment status assessment results are integrated. The weighted average method is adopted and coded in Python. By defining the weighted_average_aggregation function, the parameters include the status information of each component and its corresponding weight. The health status assessment results of each component are aggregated from the bottom up according to the set hierarchical structure. By summarizing and integrating the status assessment results of multiple components, the numerical representation of the hierarchical equipment status assessment results is calculated.

[0089] See also Figure 6 ,Based on the hierarchical equipment status assessment results, a weighted average algorithm is used to calculate the comprehensive health index according to the degree of influence of the differentiated level on the equipment performance. According to the change trend of the health index and the preset threshold, the specific steps for formulating the equipment maintenance optimization strategy are as follows:

[0090] S501: Based on the hierarchical equipment status evaluation results, identify the impact of each level component on equipment performance, assign a weight value to each level, and obtain the specific process of the impact weight setting information as follows;

[0091] In sub-step S501, based on the hierarchical equipment status assessment results, impact analysis and weight allocation are performed. The hierarchical analysis method is used to perform calculations through Python's NumPy library. The parameter settings include the relative importance matrix of components at each level. The pairwise_comparison function is used to compare components at each level, calculate weight values, assign weight values ​​to each level, and generate impact weight setting information.

[0092] S502: Based on the weight setting information, a weighted average algorithm is used to perform weighted calculation on the health status of each level, and the overall health status index of the manufacturing equipment is calculated to reflect the current operating status of the manufacturing equipment. The specific process of obtaining the comprehensive health index is as follows;

[0093] In sub-step S502, based on the weight setting information, weighted calculation is performed to obtain a comprehensive health index. The weighted average algorithm is applied and implemented in Python. By defining the weighted_average function, the parameters include the health status of each level and its weight. The health status of each level is weighted calculated according to the set weight to reflect the current operating status of the manufacturing equipment and obtain a comprehensive health index.

[0094] Weighted average algorithm, according to the formula:

[0095]

[0096] Calculate the comprehensive health index, where H′ is the improved comprehensive health index, representing the overall health status of the device under specific conditions, and w i is the weight of the i-th level, indicating the impact of this level on the overall equipment health status, h i is the health status of the i-th level, reflecting the current operation quality of the level, a i is the adjustment coefficient of the i-th level, which is set based on historical performance data and performance variability under specific operating conditions and is used to adjust the contribution of the level in the comprehensive health index. n is the total number of levels, indicating the number of levels considered in the equipment analysis.

[0097] The specific implementation process of the improved formula is as follows:

[0098] For each level, its health status h i Indicates the current operating quality of the level, weight w i It reflects the impact of this level on the overall equipment health status. The adjustment coefficient a is set through historical data analysis and expert experience. i , represents the performance variability of the layer under specific operating conditions, ensuring that the characteristics and operating environment of each layer are accurately considered, and the health status of each layer h i With its weight w i and adjustment factor a i A product calculation is performed to accumulate the calculation results of all levels to obtain an improved comprehensive health index, which more accurately reflects the overall health status of the device under specific conditions.

[0099] S503: Based on the comprehensive health index, the preset manufacturing equipment operation threshold is compared to identify the equipment maintenance and optimization requirements, and a matching equipment maintenance adjustment plan is formulated, including adjusting process parameters and planning component replacement. The specific process of obtaining the equipment maintenance optimization strategy is as follows;

[0100] In sub-step S503, maintenance needs are identified and optimization strategies are formulated based on the comprehensive health index. The comprehensive health index is compared with the preset manufacturing equipment operation threshold. Through Python's if-else logic judgment, the parameters include the comprehensive health index and the threshold. This process identifies the maintenance and optimization needs of the equipment. Based on the identification results, an equipment maintenance adjustment plan is formulated, including adjusting process parameters and planning component replacement, to generate an equipment maintenance optimization strategy.

[0101] See also Figure 7 , implement process parameter optimization strategy and equipment maintenance optimization strategy, adjust the process parameters of manufacturing equipment, monitor the implementation effect of adjustment and maintenance activities, evaluate the performance of equipment after maintenance adjustment, and obtain continuous optimization feedback data. The specific steps are as follows:

[0102] S601: According to the process parameter optimization strategy and the equipment maintenance optimization strategy, the temperature setting, pressure level and operation sequence parameters of the equipment are adjusted, and matching maintenance operations are performed on the equipment. The specific process of optimizing the equipment performance through parameter adjustment and maintenance operations and obtaining the equipment optimization results is as follows;

[0103] In sub-step S601, according to the process parameter optimization strategy and equipment maintenance optimization strategy, the equipment parameters are adjusted and maintenance operations are implemented. Through the automated control system, the temperature, pressure and operation sequence parameters are set. The Python and PLC communication interfaces are used to adjust the equipment parameters. Maintenance operations such as cleaning and parts replacement are performed. Through parameter adjustment and maintenance operations, the equipment performance is optimized and the equipment optimization results are obtained.

[0104] S602: Based on the equipment optimization results, by real-time monitoring and recording of manufacturing equipment operation data, including output, energy consumption and product quality performance indicators, comparing the changes before and after the adjustment, evaluating the actual effect of the adjustment, and obtaining the optimization effect analysis results, the specific process is as follows;

[0105] In sub-step S602, based on the equipment optimization results, operation data monitoring and effect evaluation are carried out. A real-time data acquisition system is used to collect output, energy consumption and product quality performance indicators. Through data analysis software, such as Python's Pandas and Matplo tlib libraries, the data changes before and after the adjustment are compared, the actual effects of the adjustment measures are evaluated, and the optimization effect analysis results are obtained.

[0106] S603: Based on the optimization effect analysis results, determine whether the process parameter adjustment and equipment maintenance optimization have achieved the expected improvement effect, and update and adjust the process parameters and maintenance measures to improve equipment performance. The specific process of obtaining continuous optimization feedback data is as follows;

[0107] In sub-step S603, based on the optimization effect analysis results, the effects of the optimization measures are judged and feedback is updated. Statistical analysis methods, such as t-tests, are used to determine whether the process parameter adjustments and equipment maintenance optimization measures have achieved the expected improvement effects. Python is used for data analysis, and parameter settings include significance levels and sample data sets. Process parameters and maintenance measures are updated according to the analysis results to improve equipment performance and obtain continuous optimization feedback data.

[0108] The above are only preferred embodiments of the present invention and are not intended to limit the present invention in other forms. Any technician familiar with the profession may use the technical contents disclosed above to change or modify them into equivalent embodiments with equivalent changes and apply them to other fields. However, any simple modification, equivalent change and modification made to the above embodiments based on the technical essence of the present invention without departing from the technical solution of the present invention still falls within the protection scope of the technical solution of the present invention.

Claims

1. A manufacturing equipment monitoring method based on the Internet of Things, characterized in that: The following steps are involved: Based on the on-site environment of manufacturing equipment, the multi-source data of manufacturing equipment is collected through the Internet of Things to build a heterogeneous network of equipment data; Based on the device data heterogeneous network, a node sequence is sampled by random walk, a vector representation is generated for each node and edge, and a network feature embedding vector is obtained; Based on the network feature embedding vector, an optimization analysis of the process parameters is performed, the information of the nodes and adjacent nodes is aggregated, the representation of the nodes is updated, the relationship between the nodes is captured, and the process parameters are predicted and adjusted using the updated node representation to generate a process parameter optimization strategy; Based on the operating data of manufacturing equipment, the status of each component is evaluated layer by layer. The decision tree classification algorithm is used to analyze the data to determine the health status of each component and obtain the hierarchical equipment status evaluation results. Based on the hierarchical equipment status assessment results, a weighted average algorithm is used to calculate the comprehensive health index according to the degree of influence of the differentiated levels on equipment performance, and an equipment maintenance optimization strategy is formulated based on the change trend of the health index and the preset threshold; Implement the process parameter optimization strategy and equipment maintenance optimization strategy, evaluate the equipment performance after maintenance adjustment, and obtain continuous optimization feedback data.

2. The method for monitoring manufacturing equipment based on the Internet of Things according to claim 1, characterized in that: The equipment data heterogeneous network includes sensor nodes, operation log nodes, maintenance record nodes and the interaction relationship between nodes. The network feature embedding vector includes the spatial position vector of the node and the strength index of the relationship between nodes. The process parameter optimization strategy includes the adjusted temperature setting, pressure level, and operation sequence. The hierarchical equipment status evaluation results include the degree of component wear, equipment system operation efficiency, and potential failure points. The equipment maintenance optimization strategy includes maintenance time, component replacement priority, and maintenance operation process. The continuous optimization feedback data includes equipment performance comparison information before and after parameter adjustment, optimization measure effect evaluation information, and subsequent optimization strategy.

3. The method for monitoring manufacturing equipment based on the Internet of Things according to claim 1, characterized in that: Based on the heterogeneous network of device data, the specific steps of generating vector representations for each node and edge by randomly walking and sampling node sequences and obtaining network feature embedding vectors are as follows: Based on the device data heterogeneous network, perform sequence sampling on the nodes in the network, record the access order of the nodes, simulate the logical flow and interaction between the data, and obtain a node sequence sample set; Based on the node sequence sample set, analyzing the roles and functions of differentiated nodes in equipment operation and maintenance activities, identifying contextual associations between nodes, and constructing a node contextual relationship set; Based on the node context relationship set, a corresponding vector value is assigned to each node and the edge between nodes. The vector reflects the characteristics of the node and the interaction mode between nodes, including the functional category of the node and the interaction intensity between nodes, to obtain the network feature embedding vector.

4. The method for monitoring manufacturing equipment based on the Internet of Things according to claim 1, characterized in that: Based on the network feature embedding vector, the process parameters are optimized and analyzed. By aggregating the information of nodes and adjacent nodes, the representation of nodes is updated, the relationship between nodes is captured, and the process parameters are predicted and adjusted using the updated node representation. The specific steps of generating the process parameter optimization strategy are as follows: Based on the network feature embedding vector, the vector is sorted and analyzed, and the key operation points of the equipment are reflected through analysis, the distribution and role of the nodes in the network are sorted out, and a node role and information flow framework is constructed; Based on the node roles and information flow framework, the interaction between key nodes and adjacent nodes is analyzed, the interaction patterns between key nodes and adjacent nodes are identified, and the node interaction identification results are obtained; Based on the node interaction identification results, information aggregation is performed, and the node representation is updated to reflect the current state and function of the node. Based on the updated node, the process parameters of the manufacturing equipment are predicted and adjusted to generate a process parameter optimization strategy.

5. The method for monitoring manufacturing equipment based on the Internet of Things according to claim 1, characterized in that: Based on the operating data of manufacturing equipment, the status of each component is evaluated layer by layer. The decision tree classification algorithm is used to analyze the data to determine the health status of each component. The specific steps to obtain the hierarchical equipment status evaluation results are as follows: Based on the operation data of manufacturing equipment, identify various basic components of manufacturing equipment and related performance indicators, including component wear degree and efficiency indicators, and obtain component performance indicator information; Based on the identified component performance indicators, a decision tree classification algorithm is used to analyze the data of each component, and the component health status is evaluated by comparing the historical performance data with the current operating status to obtain component status information; Based on the component status information, the health status assessment result of each component is aggregated from the bottom layer to the top layer by layer, and the hierarchical equipment status assessment result is obtained by summarizing and integrating the status assessment results of multiple components.

6. The method for monitoring manufacturing equipment based on the Internet of Things according to claim 5, characterized in that: The decision tree classification algorithm is based on the formula: Calculate the entropy value of component operation data and evaluate the health status of components, where E′(S) is the entropy value of set S, S is the operation data set of multiple components, and p i is the original proportion of the i-th state in the set S, Δp i is the adjustment value of the i-th state ratio, c i The weight coefficient of the i-th state, n is the total number of states.

7. The method for monitoring manufacturing equipment based on the Internet of Things according to claim 1, characterized in that: Based on the hierarchical equipment status assessment results, a weighted average algorithm is used to calculate the comprehensive health index according to the degree of influence of the differentiated levels on equipment performance. According to the change trend of the health index and the preset threshold, the specific steps for formulating the equipment maintenance optimization strategy are as follows: Based on the hierarchical equipment status evaluation results, identifying the degree of influence of each level component on equipment performance, assigning a weight value to each level, and obtaining influence weight setting information; Based on the weight setting information, a weighted average algorithm is used to perform weighted calculation on the health status of each level, calculate the overall health status index of the manufacturing equipment, reflect the current operating status of the manufacturing equipment, and obtain a comprehensive health index; The comprehensive health index is compared with the preset manufacturing equipment operation threshold to identify the equipment maintenance and optimization needs, and a matching equipment maintenance adjustment plan is formulated, including adjusting process parameters and planning component replacement, to obtain an equipment maintenance optimization strategy.

8. The method for monitoring manufacturing equipment based on the Internet of Things according to claim 7, characterized in that: The weighted average algorithm is based on the formula: Calculate the comprehensive health index, where H′ is the improved comprehensive health index, w i is the weight of the i-th level, h i is the health status of the i-th level, a i is the adjustment coefficient of the ith level, and n is the total number of levels.

Images