An asset management system based on industrial internet identification

Abstract

This invention discloses an asset management system based on industrial internet identifiers, relating to the field of equipment management. It includes: an identifier reading module for real-time acquisition of parameter attribute information of target equipment via a PLC; a target identification module for deploying sensors at the target equipment to detect and identify several indicator data points of the target area inside and outside the equipment, identifying target areas that meet the attention threshold, generating a unique equipment identifier for the target area for marking and storage; and a data analysis unit for evaluating the health status of the target area represented by a certain equipment identifier, segmenting multiple current task indicators of the equipment into time periods to form digital indicator identifiers, and evaluating the correlation score between the indicator identifier and the equipment identifier. This real-time assessment of equipment health status ensures the system's efficiency and accuracy, improves the intelligence level of equipment management, and provides users with more personalized and precise management solutions.

Description

Technical Field

[0001] This invention relates to the field of equipment management technology, specifically to an asset management system based on industrial internet identification. Background Technology

[0002] With the deep integration of information technology and traditional manufacturing, the Industrial Internet is gradually emerging. By adopting technologies such as smart sensors, IoT, and big data analytics, it is possible to efficiently collect, analyze, and manage equipment data, thereby improving industrial production efficiency and equipment management. The development of technologies such as PLC, sensors, edge computing, and cloud computing enables equipment parameters to be collected in real time and to undergo complex data analysis and processing. Enterprises hope to improve production efficiency and market competitiveness through refined asset management. However, existing systems typically lack the ability to deeply analyze the health status of equipment, fail to detect potential problems in a timely manner, lack refined identification and management of local areas of equipment, and are unable to conduct health monitoring for specific areas. This results in low fault location accuracy, the absence of a dynamic correlation between task indicators and equipment health status, and a lack of systematic methods for equipment adjustments. They are unable to flexibly adjust based on real-time data and prediction results, and their response speed is slow in monitoring equipment operation status and implementing adjustment strategies, making it difficult to quickly adapt to changes. Summary of the Invention

[0003] (a) Technical problems to be solved In view of the above-mentioned shortcomings of the existing technology, the present invention provides an asset management system based on industrial internet identification, which can effectively solve the problems of the existing technology.

[0004] (II) Technical Solution To achieve the above objectives, the present invention provides the following technical solution: This invention discloses an asset management system based on industrial internet identification, comprising: The identification reading module is used to collect parameter attribute information of the target device in real time via PLC; The target recognition module is used to deploy sensors at the target device to detect and identify several indicators of the target area inside and outside the device, identify the target area that meets the attention threshold, and generate a unique device identifier for the target area for marking and storage. The data analysis unit is used to assess the health status of the target area represented by a device identifier, to segment multiple current task indicators of the device into time periods, to form digital indicator identifiers, and to evaluate the correlation score between the indicator identifiers and the device identifiers. The model building module is used to extract key features from the health status and index correlation scores of the target area and train the prediction model using a neural network. The loss prediction module is used to extract the prediction model trained by the model building module, input the feature data of the current period, and output the predicted value of the loss coefficient for the specified target area. The indicator adjustment module is used to obtain the predicted loss coefficient, identify the adjustment space of the task indicator in terms of duration and intensity, simulate different adjustment schemes through sensitivity analysis, calculate and sort the impact of each adjustment scheme on equipment loss, and generate adjustment suggestions including adjustment range and type. The strategy generation module is used to reorganize the adjustment plan in the time dimension based on the adjustment suggestions output by the indicator adjustment module, the adjustment time point, adjustment range and execution steps, to form a new indicator time series and generate the optimal process scheduling strategy.

[0005] Furthermore, the data analysis unit is equipped with sub-modules, including a health assessment module, an indicator segmentation module, and a correlation scoring module. The indicator segmentation module interacts with the health assessment module and the correlation scoring module via a wireless network. The health assessment module compares several indicator data of the target area with health thresholds to determine the degree of deviation of the indicator parameters. Combined with historical monitoring data, it uses a trend analysis algorithm to calculate the health score of the target area. The indicator segmentation module is used to collect the task indicators of the equipment, segment the task indicators on the time axis according to the duration and intensity, divide the continuous indicators into multiple time periods to form discrete indicator segments, assign a unique indicator identifier to each segment, and statistically analyze the duration, average value and maximum value characteristics of each indicator segment. The correlation scoring module is used to evaluate the correlation between indicator segments and the corresponding target area device identifiers. It calculates and stores the correlation score between indicator identifiers and device identifiers by combining the importance weight of the indicators.

[0006] Furthermore, the criteria for determining the deviation threshold of the health assessment module include: the standard operating condition parameter range and the time-varying parameter baseline generated by the sliding window averaging method based on historical monitoring data. The degree of deviation is calculated using a segmented quantification strategy. When the indicator data exceeds the standard operating condition parameter range, a first-level deviation alarm is triggered. When the indicator data exceeds the preset scale of the dynamic threshold for several consecutive periods, a second-level trend deviation alarm is triggered. Linear regression analysis is performed on the indicator data within several consecutive periods. When the absolute value of the slope exceeds the preset slope threshold, a trend deviation is determined to exist.

[0007] Furthermore, in the trend analysis algorithm used by the health assessment module, the parameters for assessing the degree of deviation include: sliding window parameters, time series prediction parameters, and statistical test parameters; the parameters for calculating the health score include: dynamic weight assignment parameters, trend deterioration coefficient, and health score synthesis parameters.

[0008] Furthermore, the process by which the target recognition module generates a unique device identifier for the target area is as follows: Acquire spatial coordinates, geometric features, or spatial relative position data of the target area; Extract the attribute feature information of the internal parameters and external environmental parameters of the target area; Obtain the Industrial Internet Identifier Code of the target device; The spatial and attribute features of the target area are integrated with the industrial internet identification code information. The integrated multi-source information is then encoded to generate a unique identification code. The encoded result is used as a unique identifier for the target region and stored.

[0009] Furthermore, the evaluation method for the correlation score in the correlation scoring module includes the following steps: The duration, average value, and maximum value of the indicator segment are obtained, and the features are standardized to form a feature vector; According to the device type, the corresponding weight matrix is ​​extracted from the preset weight allocation rule library. The weight matrix includes duration weight, mean weight and extreme value weight, wherein: time weight + mean weight + extreme value weight = 1; Feature quantization processing is performed on the historical fault dataset associated with equipment identifiers in the target area to establish a baseline vector of equipment sensitive features; Calculate the cosine similarity between the current feature vector and the baseline vector as the basic correlation. An indicator importance correction factor is introduced to adjust the weight distribution according to the criticality level of the task indicators, thereby generating a dynamic weight matrix. The final association score is calculated using a weighted association degree algorithm; The final association score is normalized, mapped to the (0,1) interval, and then stored in the association score database.

[0010] Furthermore, the expression for the operational logic of the prediction model constructed by the model building module is as follows: ; In the formula, The target area represents the period The predicted value of the loss coefficient, Represents a non-linear activation function. Represents the number of neurons in the hidden layer. This represents the weight vector from the hidden layer to the output layer. This represents the total dimension of the input features. This represents the weight matrix from the input layer to the hidden layer. Represents the input feature vector. Represents the hidden layer The bias of each neuron represent, Represents the index of hidden layer neurons. Represents the input feature index.

[0011] Furthermore, the operating logic of the indicator adjustment module is as follows: Based on the discretization results of historical data of equipment operation task indicators, the control boundaries of each indicator in the dimensions of duration and intensity are extracted, and a parameter adjustment space with duration adjustment amplitude and intensity change gradient as the two axes is established. Monte Carlo sampling is performed on the parameter adjustment space to generate several sets of candidate adjustment schemes. A multiple regression model of equipment loss coefficient and adjustment parameters of each index is constructed to quantify the marginal effect weight of different parameter combinations on the loss coefficient. Each candidate adjustment scheme is input into a multiple regression model, and combined with the baseline value of the loss coefficient predicted in the current cycle, the relative change rate and absolute increment of the loss coefficient after the implementation of each scheme are obtained. Simultaneously, equipment health status constraints are superimposed to filter out illegal schemes that exceed the preset risk threshold. Based on the reduction in losses, deviation in task completion rate, and adjustment of operating costs, a comprehensive evaluation function containing weighted factors is established to normalize the scores of legitimate candidate solutions. The weighted factors are dynamically configured according to the equipment type. Candidate schemes are sorted in descending order of their scores. A set of adjustment suggestions is generated from the top few schemes with scores higher than a set threshold. The suggestions include the time window for adjusting the indicators, the intensity gradient change curve, and the expected percentage reduction in losses for each scheme.

[0012] Furthermore, the strategy generation module is interconnected with the execution feedback module via a wireless network. The execution feedback module detects the application path of the generated adjustment strategy in real time, provides real-time adjustment of indicator parameters through the industrial control interface, monitors the equipment operating status in real time, and collects actual indicator data, using the actual indicator data as input reference for the training and optimization of the prediction model in the next cycle.

[0013] Furthermore, the target identification module is interconnected with the identifier reading module and the data analysis unit via a wireless network; the model building module is interconnected with the data analysis unit and the loss prediction module via a wireless network; and the index adjustment module is interconnected with the loss prediction module and the strategy generation module via a wireless network.

[0014] (III) Beneficial Effects Compared with the known prior art, the technical solution provided by this invention has the following beneficial effects: By collecting parameter attribute information and sensor monitoring data of target equipment in real time and combining them with trend analysis algorithms, the system can assess the health status of equipment in real time, ensuring the efficiency and accuracy of the system. It can dynamically respond to changes in equipment status and generate health scores in real time through big data analysis and machine learning technology, thereby improving the level of intelligence in equipment management, significantly reducing equipment maintenance response time, and improving production efficiency.

[0015] By conducting in-depth correlation analysis between health indicators and equipment identifiers in the target area, the system segments task indicators according to duration and intensity to form discrete indicator fragments. By comparing and analyzing historical data, the system can accurately identify key equipment indicators and their influencing factors. This not only provides a scientific basis for equipment health assessment but also provides data support for subsequent loss prediction and adjustment plans. This gives the system a unique technological advantage in the field of asset management, enabling it to provide users with more personalized and precise management solutions.

[0016] By training the prediction model and calculating the loss coefficient, the system can identify the adjustment space of task indicators in terms of duration and intensity. Through sensitivity analysis, different adjustment schemes are simulated, the impact of each scheme on equipment loss is calculated and ranked, and adjustment suggestions are generated. This enhances the system's initiative in equipment management, provides data-driven decision support for enterprises, optimizes process scheduling strategies, and enables dynamic adjustments by reorganizing the indicator time series. This ensures that equipment is always in optimal operating condition in a changing production environment, giving the asset management system the potential for continuous optimization in a rapidly changing industrial environment. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are merely some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without any creative effort.

[0018] Figure 1 This is a schematic diagram of the overall framework of the present invention.

[0019] The labels in the diagram represent: 1. Identifier Reading Module; 2. Target Recognition Module; 3. Data Analysis Unit; 31. Health Assessment Module; 32. Indicator Segmentation Module; 33. Correlation Scoring Module; 4. Model Building Module; 5. Loss Prediction Module; 6. Indicator Adjustment Module; 7. Strategy Generation Module; 8. Execution Feedback Module. Detailed Implementation

[0020] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0021] The present invention will be further described below with reference to embodiments. Example 1

[0022] This embodiment presents an asset management system based on industrial internet identification, such as... Figure 1 As shown, it includes: Identification reading module 1 is used to collect parameter attribute information of the target device in real time via PLC.

[0023] Target identification module 2 is used to deploy sensors at target devices to detect and identify several indicators of the target area inside and outside the device. It identifies target areas that meet the attention thresholds, generates a unique device identifier for each target area, and stores it. The process of generating a unique device identifier for a target area involves: acquiring the spatial coordinates, geometric features, or spatial relative position data of the target area; extracting the attribute characteristics of the target area's internal parameters and external environmental parameters; acquiring the target device's Industrial Internet Identifier (III) code; fusing the spatial and attribute characteristics of the target area with the III code information; encoding the fused multi-source information to generate a unique identifier; and storing the encoded result as the unique identifier for the target area. By filtering target areas through the set attention thresholds, it ensures that the identified targets meet the preset monitoring requirements. Assigning a unique device identifier to each target area guarantees its uniqueness and traceability, providing basic data support for subsequent asset status assessment and monitoring.

[0024] Data analysis unit 3 is used to evaluate the health status of a target area represented by a device identifier. It segments multiple current task indicators of the device into time periods, forming digital indicator identifiers, and evaluates the correlation score between the indicator identifiers and the device identifier. This enables dynamic monitoring and status updates of the target area to support subsequent health assessments and maintenance decisions. Data analysis unit 3 has sub-modules, including: a health assessment module 31, an indicator segmentation module 32, and a correlation scoring module 33. The indicator segmentation module 32 interacts with the health assessment module 31 and the correlation scoring module 33 via a wireless network. The health assessment module 31 is used to compare several indicator data of the target area with health thresholds to determine the degree of deviation of the indicator parameters. Combined with historical monitoring data, a trend analysis algorithm is used to calculate the health score of the target area. The criteria for determining the indicator deviation threshold include: the standard operating condition parameter range and the time-varying parameter baseline generated by the sliding window averaging method based on historical monitoring data. The calculation of the degree of deviation adopts a segmented quantification strategy. When the indicator data exceeds the standard operating condition parameter range, a first-level deviation alarm is triggered. When the indicator data exceeds the preset scale of the dynamic threshold for several consecutive periods, a second-level trend deviation alarm is triggered. Linear regression analysis is performed on the indicator data within several consecutive periods. When the absolute value of the slope exceeds the preset slope threshold, it is determined that there is a trend deviation. The trend analysis algorithm used in the health assessment module 31 includes the following parameters for evaluating the degree of deviation: a sliding window parameter, used to extract historical fluctuation characteristics of indicator data in the target area, with the window size set to a number of consecutive collection cycles in the historical monitoring data; a time series prediction parameter, used to predict the trend of abnormal indicators in the target area over several future cycles; and a statistical test parameter, which sets a significant deviation threshold based on historical health status data and uses a two-sample test to calculate the probability of difference between the current indicator deviation value and the historical normal range. The calculation parameters for the health score include: a dynamic weight assignment parameter, which assigns different weight factors according to the indicator type (temperature, vibration, energy consumption), with the total weight factor being 1; a trend deterioration coefficient, which calculates the deterioration rate of indicator data over several consecutive cycles using a weighted moving average algorithm, where the number of cycles is related to the equipment maintenance cycle; and a health score synthesis parameter, which uses a piecewise linear function to map the degree of deviation and deterioration rate to a 0-100 score range and adds a safety attenuation coefficient provided by the equipment manufacturer for normalization. The sliding window parameter is linked to the statistical test parameter, and when the variance of the data within the window exceeds a preset critical value, the adjustment mechanism of the dynamic weight assignment parameter is triggered. The indicator segmentation module 32 is used to collect the task indicators of the equipment, segment the task indicators on the time axis according to the duration and intensity, divide the continuous indicators into multiple time periods to form discrete indicator segments, assign a unique indicator identifier to each segment, and statistically analyze the duration, average value and maximum value characteristics of each indicator segment. The correlation scoring module 33 is used to evaluate the correlation between indicator segments and the corresponding target area device identifiers, and calculates and stores the correlation score between indicator identifiers and device identifiers by combining the importance weight of the indicators.

[0025] Model building module 4 is used to extract key features from the health status and index correlation scores of the target area and train the prediction model using a neural network. This invention optimizes model parameters and enhances the generalization ability and prediction accuracy of the model through an end-to-end training method.

[0026] The loss prediction module 5 is used to extract the prediction model trained by the model building module 4, input the feature data of the current period, and output the predicted value of the loss coefficient for the specified target area.

[0027] The indicator adjustment module 6 is used to obtain the predicted loss coefficient, identify the adjustment space of the task indicator in terms of duration and intensity, simulate different adjustment schemes through sensitivity analysis, calculate and sort the impact of each adjustment scheme on equipment loss, and generate adjustment suggestions including adjustment range and type. The operating logic of indicator adjustment module 6 is as follows: Based on the discretization results of historical data of equipment operation task indicators, the control boundaries of each indicator in the dimensions of duration and intensity are extracted, and a parameter adjustment space with duration adjustment amplitude and intensity change gradient as the two axes is established. Monte Carlo sampling is performed on the parameter adjustment space to generate several sets of candidate adjustment schemes. A multiple regression model of equipment loss coefficient and adjustment parameters of each index is constructed to quantify the marginal effect weight of different parameter combinations on the loss coefficient. Each candidate adjustment scheme is input into a multiple regression model, and combined with the baseline value of the loss coefficient predicted in the current cycle, the relative change rate and absolute increment of the loss coefficient after the implementation of each scheme are obtained. Simultaneously, equipment health status constraints are superimposed to filter out illegal schemes that exceed the preset risk threshold. Based on the reduction in losses, deviation in task completion rate, and adjustment of operating costs, a comprehensive evaluation function containing weighted factors is established to normalize the scores of legitimate candidate solutions. The weighted factors are dynamically configured according to the equipment type. Candidate schemes are sorted in descending order of their scores. A set of adjustment suggestions is generated from the top few schemes with scores higher than a set threshold. The suggestions include the time window for adjusting the indicators, the intensity gradient change curve, and the expected percentage reduction in losses for each scheme.

[0028] The strategy generation module 7 is used to reorganize the adjustment plan in the time dimension according to the adjustment suggestions output by the indicator adjustment module 6, based on the adjustment time point, adjustment range, and execution steps, to form a new indicator time series and generate the optimal process scheduling strategy. The strategy generation module 7 is connected to the execution feedback module 8 via a wireless network. The execution feedback module 8 detects the application path of the generated adjustment strategy in real time, provides real-time adjustment parameter permissions through the industrial control interface, monitors the equipment operating status in real time, and collects actual indicator data, which is used as the input reference for the training and optimization of the prediction model in the next cycle.

[0029] The target identification module 2 is interconnected with the identifier reading module 1 and the data analysis unit 3 via a wireless network. The model building module 4 is interconnected with the data analysis unit 3 and the loss prediction module 5 via a wireless network. The index adjustment module 6 is interconnected with the loss prediction module 5 and the strategy generation module 7 via a wireless network.

[0030] Compared with existing technologies, this embodiment achieves innovative integration of the entire process through the collaboration of various modules, including real-time identification of assets and equipment, target area recognition, health status assessment, digital segmentation of indicators, correlation scoring, model training, loss prediction, indicator adjustment analysis, and automatic generation and execution of process strategies. This embodiment has the advantages of reasonable structure, high efficiency, and strong scalability, providing reliable technical support for the intelligent management of industrial assets. Example 2

[0031] At other levels, this embodiment also provides another optimization mechanism based on Embodiment 1, specifically a predictive model's operating logic, expressed as follows: ; In the formula, The target area represents the period The predicted value of the loss coefficient, Represents a non-linear activation function. Represents the number of neurons in the hidden layer. This represents the weight vector from the hidden layer to the output layer. This represents the total dimension of the input features. This represents the weight matrix from the input layer to the hidden layer. Represents the input feature vector. Represents the hidden layer The bias of each neuron represent, Represents the index of hidden layer neurons. Represents the input feature index.

[0032] In this embodiment, a two-layer neural network structure is adopted. The input layer receives features, and high-order features are extracted through nonlinear transformation in the hidden layer. Finally, a normalized loss prediction value is output. By combining the two-layer neural network structure with a nonlinear activation function, efficient nonlinear mapping of multi-dimensional features of the equipment is achieved. While preserving the temporal characteristics of the input data, the accuracy of loss prediction is significantly improved. Its input features integrate static health scores and dynamic task indicator fragment statistics, taking into account both the long-term degradation law of the equipment and the impact of short-term operating condition fluctuations. The weight training mechanism ensures that the model adapts to the data distribution characteristics of the industrial scenario, and finally provides quantitative prediction results with both interpretability and generalization ability for equipment maintenance decisions. Example 3

[0033] This embodiment provides a method for evaluating correlation scores, including the following steps: Obtain the duration, average value, and maximum value features of the indicator segments, and standardize the features to form a feature vector; Based on the device type, the corresponding weight matrix is ​​extracted from the preset weight allocation rule library. The weight matrix includes duration weight, mean weight, and extreme value weight, where: time weight + mean weight + extreme value weight = 1. Feature quantization processing is performed on the historical fault dataset associated with equipment identifiers in the target area to establish a baseline vector of equipment sensitive features; The cosine similarity between the current feature vector and the reference vector is calculated as the basic correlation. The standardized current feature vector and the reference vector generated from the historical fault data of the equipment are subjected to vector dot product operation. Then the magnitude of the two vectors is calculated respectively. The dot product result is divided by the product of the two magnitudes to obtain the cosine similarity between the two as the basic correlation. An indicator importance correction factor is introduced to adjust the weight distribution according to the criticality level of the task indicators, thereby generating a dynamic weight matrix. The final correlation score is calculated using a weighted correlation algorithm. The weights are dynamically adjusted based on the criticality level of the task indicators. The adjusted weights are multiplied by the basic correlation score, and all the products are summed before being multiplied by the environmental factor correction coefficient. The final association score is normalized, mapped to the 0,1 interval, and then stored in the association score database. Compared with existing technologies, this method overcomes the limitations of traditional static evaluation based on a single indicator by constructing multi-dimensional feature vectors and using a dynamic weight adjustment mechanism. It adopts a cosine similarity metric to quantify the correlation between equipment operating characteristics and historical failure modes, and introduces an environmental correction coefficient and a task criticality adaptive factor. This enables the evaluation model to have both failure mode matching accuracy and real-time operating condition adaptability. Compared with existing methods based on fixed threshold alarms or univariate trend analysis, this method significantly improves the dynamic correlation and predictive reliability of the correlation analysis between equipment health status and task load.

[0034] In summary, this invention consists of an identification reading module 1 that collects equipment parameter attributes in real time via a PLC; a target identification module 2 that deploys sensors inside and outside the equipment to detect target area indicator data, identify abnormal areas, and assign unique equipment identifiers for storage; a data analysis unit 3 that initiates a health assessment, calculating a health score for the target area through a health assessment module 31; an indicator segmentation module 32 that segments the task indicator time series into discrete segments and generates indicator identifiers; an association scoring module 33 that evaluates the association score between the indicator identifier and the equipment identifier; a model building module 4 that extracts features based on health and association scores to train a neural network prediction model; a loss prediction module 5 that calls the model to predict the loss coefficient of the target area; an indicator adjustment module 6 that performs a task indicator sensitivity analysis on the prediction results and generates an optimization adjustment plan; a strategy generation module 7 that reorganizes the indicator time series according to the plan to form an optimal scheduling strategy; and finally, an execution feedback module 8 that deploys the strategy in real time, adjusts equipment parameters through an industrial interface, and feeds back the operating data in a closed loop to the model building module 4 for continuous model optimization. This invention achieves digital management of equipment assets throughout their entire lifecycle through industrial internet identification, deeply integrating real-time monitoring and intelligent analysis. By integrating operating parameters, health status, and task indicators, it constructs a dynamic task scheduling mechanism, combines neural network prediction models with sensitivity analysis, accurately predicts equipment wear trends, and generates process optimization strategies. At the same time, relying on the identification system, it achieves efficient association and traceability of equipment, indicators, and strategies, balancing equipment health assurance with dynamic adaptation of production resources, significantly improving asset management efficiency, reducing operation and maintenance costs, and preventing potential failure risks.

[0035] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions will not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. An asset management system based on industrial internet identification, characterized in that, include: The identification reading module is used to collect parameter attribute information of the target device in real time via PLC; The target recognition module is used to deploy sensors at the target device to detect and identify several indicators of the target area inside and outside the device, identify the target area that meets the attention threshold, and generate a unique device identifier for the target area for marking and storage. The data analysis unit is used to assess the health status of the target area represented by a device identifier, to segment multiple current task indicators of the device into time periods, to form digital indicator identifiers, and to evaluate the correlation score between the indicator identifiers and the device identifiers. The model building module is used to extract key features from the health status and index correlation scores of the target area and train the prediction model using a neural network. The loss prediction module is used to extract the prediction model trained by the model building module, input the feature data of the current period, and output the predicted value of the loss coefficient for the specified target area. The indicator adjustment module is used to obtain the predicted loss coefficient, identify the adjustment space of the task indicator in terms of duration and intensity, simulate different adjustment schemes through sensitivity analysis, calculate and sort the impact of each adjustment scheme on equipment loss, and generate adjustment suggestions including adjustment range and type. The strategy generation module is used to reorganize the adjustment plan in the time dimension based on the adjustment suggestions output by the indicator adjustment module, the adjustment time point, adjustment range and execution steps, to form a new indicator time series and generate the optimal process scheduling strategy. The data analysis unit has sub-modules deployed below it, including: a health assessment module, an indicator segmentation module, and a correlation scoring module. The indicator segmentation module interacts with the health assessment module and the correlation scoring module via a wireless network. The health assessment module compares several indicator data of the target area with health thresholds to determine the degree of deviation of the indicator parameters. Combined with historical monitoring data, it uses a trend analysis algorithm to calculate the health score of the target area. The indicator segmentation module is used to collect the task indicators of the equipment, segment the task indicators on the time axis according to the duration and intensity, divide the continuous indicators into multiple time periods to form discrete indicator segments, assign a unique indicator identifier to each segment, and statistically analyze the duration, average value and maximum value characteristics of each indicator segment. The correlation scoring module is used to evaluate the correlation between indicator segments and the corresponding target area device identifiers, and calculates and stores the correlation score between indicator identifiers and device identifiers by combining the importance weight of the indicators. The process by which the target identification module generates a unique device identifier for the target area is as follows: Acquire spatial coordinates, geometric features, or spatial relative position data of the target area; Extract the attribute feature information of the internal parameters and external environmental parameters of the target area; Obtain the Industrial Internet Identifier Code of the target device; The spatial and attribute features of the target area are integrated with the industrial internet identification code information. The integrated multi-source information is then encoded to generate a unique identification code. The encoded result is used as a unique identifier for the target region and stored. The evaluation method for correlation scores in the correlation scoring module includes the following steps: The duration, average value, and maximum value of the indicator segment are obtained, and the features are standardized to form a feature vector; According to the device type, the corresponding weight matrix is ​​extracted from the preset weight allocation rule library. The weight matrix includes duration weight, mean weight and extreme value weight, wherein: time weight + mean weight + extreme value weight = 1; Feature quantization processing is performed on the historical fault dataset associated with equipment identifiers in the target area to establish a baseline vector of equipment sensitive features; Calculate the cosine similarity between the current feature vector and the baseline vector as the basic correlation. An indicator importance correction factor is introduced to adjust the weight distribution according to the criticality level of the task indicators, thereby generating a dynamic weight matrix. The final association score is calculated using a weighted association degree algorithm; The final association score is normalized, mapped to the (0,1) interval, and then stored in the association score database; The expression for the running logic of the prediction model constructed by the model building module is as follows: ； In the formula, The target area represents the period The predicted value of the loss coefficient, Represents a non-linear activation function. Represents the number of neurons in the hidden layer. This represents the weight vector from the hidden layer to the output layer. This represents the total dimension of the input features. This represents the weight matrix from the input layer to the hidden layer. Represents the input feature vector. Represents the hidden layer The bias of each neuron represent, Represents the index of hidden layer neurons. Represents the input feature index.

2. The asset management system based on industrial internet identification as described in claim 1, characterized in that, The criteria for determining the deviation threshold of the health assessment module include: the standard operating condition parameter range and the time-varying parameter baseline generated by the sliding window averaging method based on historical monitoring data. The degree of deviation is calculated using a segmented quantification strategy. When the indicator data exceeds the standard operating condition parameter range, a first-level deviation alarm is triggered. When the indicator data exceeds the preset scale of the dynamic threshold for several consecutive periods, a second-level trend deviation alarm is triggered. Linear regression analysis is performed on the indicator data within several consecutive periods. When the absolute value of the slope exceeds the preset slope threshold, a trend deviation is determined to exist.

3. The asset management system based on industrial internet identification as described in claim 1, characterized in that, The trend analysis algorithm used in the health assessment module includes the following parameters for evaluating the degree of deviation: sliding window parameters, time series prediction parameters, and statistical test parameters; the calculation parameters for the health score include: dynamic weight assignment parameters, trend deterioration coefficient, and health score synthesis parameters.

4. The asset management system based on industrial internet identification as described in claim 1, characterized in that, The operating logic of the indicator adjustment module is as follows: Based on the discretization results of historical data of equipment operation task indicators, the control boundaries of each indicator in the dimensions of duration and intensity are extracted, and a parameter adjustment space with duration adjustment amplitude and intensity change gradient as the two axes is established. Monte Carlo sampling is performed on the parameter adjustment space to generate several sets of candidate adjustment schemes. A multiple regression model of equipment loss coefficient and adjustment parameters of each index is constructed to quantify the marginal effect weight of different parameter combinations on the loss coefficient. Each candidate adjustment scheme is input into a multiple regression model, and combined with the baseline value of the loss coefficient predicted in the current cycle, the relative change rate and absolute increment of the loss coefficient after the implementation of each scheme are obtained. Simultaneously, equipment health status constraints are superimposed to filter out illegal schemes that exceed the preset risk threshold. Based on the reduction in losses, deviation in task completion rate, and adjustment of operating costs, a comprehensive evaluation function containing weighted factors is established to normalize the scores of legitimate candidate solutions. The weighted factors are dynamically configured according to the equipment type. Candidate schemes are sorted in descending order of their scores. A set of adjustment suggestions is generated from the top few schemes with scores higher than a set threshold. The suggestions include the time window for adjusting the indicators, the intensity gradient change curve, and the expected percentage reduction in losses for each scheme.

5. An asset management system based on industrial internet identification as described in claim 1, characterized in that, The strategy generation module is interconnected with the execution feedback module via a wireless network. The execution feedback module detects the application path of the generated adjustment strategy in real time, provides real-time adjustment of indicator parameters through the industrial control interface, monitors the equipment operating status in real time, and collects actual indicator data, which is used as the input reference for the training and optimization of the prediction model in the next cycle.

6. An asset management system based on industrial internet identification as described in claim 1, characterized in that, The target identification module, the identifier reading module, and the data analysis unit are interconnected via a wireless network. The model building module, the data analysis unit, and the loss prediction module are interconnected via a wireless network. The index adjustment module, the loss prediction module, and the strategy generation module are interconnected via a wireless network.

Images