Device state data management method and system applied to automated radix astragali production line

Abstract

This application relates to the field of data analysis technology, and provides a method and system for managing equipment status data in an automated astragalus extract production line. This system enables accurate prediction and proactive maintenance of equipment health through end-to-end data monitoring and management. The method includes: acquiring a set of production line equipment status data; extracting a set of production line equipment status features from the set of data, wherein the set of features includes multi-dimensional temporal features reflecting equipment operating trends and topological features indicating inter-equipment relationships; training an astragalus extract production line status prediction model based on the set of features, wherein the model predicts the probability of equipment anomalies and performance degradation trends within a selected time period based on the current equipment status features; and generating a set of equipment management strategies based on the output of the model, wherein the set includes production parameter adjustment instructions and maintenance node planning schemes for different equipment maintenance priorities.

Description

Technical Field

[0001] This application belongs to the field of data analysis technology, specifically relating to a method and system for managing equipment status data in an automated astragalus extract production line. Background Technology

[0002] In the field of automated production of traditional Chinese medicine preparations, equipment condition monitoring and maintenance decision-making technologies play a crucial role in ensuring product quality and production line stability. Currently, the equipment health management systems commonly used in the industry rely primarily on single-dimensional real-time data acquisition, such as obtaining equipment operating parameters through vibration sensors or temperature detection devices, and then issuing alarms based on preset thresholds. Furthermore, existing solutions often employ independent monitoring modes for each piece of equipment, failing to effectively integrate the synergistic parameters between production line equipment. Particularly in complex formulation processes, the dynamic correlations between equipment are often simplified to linear superposition. In addition, existing maintenance strategy generation systems mostly use decision-making mechanisms with fixed rule bases, failing to dynamically correlate prediction results with production process parameters. When faced with batch variations in raw materials or fluctuations in environmental parameters, existing systems struggle to generate adaptive maintenance plans that balance equipment health status with production quality requirements. Especially in scenarios like Astragalus extract production, which is temperature-sensitive and has a narrow process parameter window, existing maintenance decisions often create a conflict between equipment downtime and product quality control.

[0003] Therefore, the aforementioned technical defects result in existing technologies having problems such as insufficient timeliness of equipment anomaly warning, low accuracy of performance degradation prediction, and poor compatibility between maintenance strategies and production processes in continuous production scenarios of Astragalus membranaceus extracts, making it difficult to achieve accurate prediction of equipment health and proactive maintenance. Summary of the Invention

[0004] This application provides a method and system for managing equipment status data in an automated astragalus extract production line, which enables accurate prediction and proactive maintenance of equipment health through full-process data monitoring and management.

[0005] In a first aspect, embodiments of this application provide a method for managing equipment status data in an automated astragalus extract production line, applied to an equipment status data management system. The method includes: acquiring a set of production line equipment status data, the set including production parameters, operating status parameters, and environmental monitoring parameters of each piece of equipment in multiple consecutive production cycles of the automated astragalus extract production line; extracting a set of production line equipment status features from the set of production line equipment status data, the set including multi-dimensional temporal features reflecting equipment operating trends and topological features reflecting inter-equipment relationships; training an astragalus extract production line status prediction model based on the set of production line equipment status features, the model being used to predict the probability of equipment anomalies and performance degradation trends within a selected time period based on the current equipment status features; and generating a set of equipment management strategies based on the output of the astragalus extract production line status prediction model, the set including production parameter adjustment instructions and maintenance node planning schemes for different equipment maintenance priorities.

[0006] Secondly, embodiments of this application provide a device status data management system, which includes a processor and a memory, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor performs the steps of the above-described method.

[0007] Thirdly, embodiments of this application provide a computer-readable storage medium including a computer program that, when run on a device status data management system, causes the device status data management system to perform the steps of the above-described method.

[0008] In this application, accurate prediction and proactive maintenance of equipment health are achieved through closed-loop data management throughout the entire process. Specifically, a comprehensive status profile is constructed based on multi-source equipment parameters and environmental monitoring data, supporting multi-dimensional feature extraction. Furthermore, by integrating the temporal characteristics and associated topological features of equipment operation trends, the dynamic evolution patterns and synergistic mechanisms of the equipment are revealed, enhancing the model's learning efficiency. Further, the trained Astragalus extract production line status prediction model can simultaneously identify equipment anomaly risks and performance degradation trends, achieving early fault warnings and deterioration path predictions. Finally, based on the generated set of differentiated equipment management strategies, process parameters can be dynamically adjusted and maintenance nodes can be intelligently planned, reducing the risk of unplanned downtime and optimizing maintenance resource allocation. This application's embodiment forms a closed-loop control from data perception to decision execution, improving production line operational stability and equipment lifecycle management capabilities, ensuring the continuity and consistency of Astragalus extract production. Attached Figure Description

[0009] Figure 1This is a flowchart illustrating a method for managing equipment status data in an automated astragalus extract production line, as provided in an embodiment of this application.

[0010] Figure 2 This is a schematic diagram of the structure of a device status data management system provided in an embodiment of this application. Detailed Implementation

[0011] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of this application will be clearly and completely described below with reference to the accompanying drawings of the embodiments. Obviously, the described embodiments are only some, not all, of the embodiments of the technical solutions of this application. All other embodiments obtained by those skilled in the art based on the embodiments described in this application without creative effort are within the scope of protection of the technical solutions of this application. See also... Figure 1 This is a method for managing equipment status data in an automated astragalus extract production line provided in this application embodiment. This method can be applied to an equipment status data management system, and the specific process is as follows: steps 110-140.

[0012] Step 110: Obtain the production line equipment status data set, which includes the production parameters, operating status parameters, and environmental monitoring parameters of each piece of equipment in multiple consecutive production cycles of the automated Astragalus Extract production line.

[0013] In this embodiment, the equipment status data management system collects and stores multi-source heterogeneous data in real time through a sensor network deployed at key nodes of the automated Astragalus membranaceus production line. For example, in the Astragalus membranaceus raw material extraction section, a temperature sensor records the temperature change curve of the liquid medium in the extraction tank once per minute, a pressure transmitter continuously monitors the steam pressure value of the tank jacket, and a flow meter cumulatively counts the hourly injection volume of ethanol solvent. In the centrifugal separation equipment, a vibration sensor captures the triaxial vibration waveform of the drum bearing at a sampling rate of 500Hz, and a current transformer synchronously collects the three-phase operating current of the drive motor.

[0014] In addition, the deployed environmental monitoring module continuously records and extracts air temperature and humidity distribution data in different areas of the workshop through distributed temperature and humidity probes, and measures the concentration of suspended particles in the operating area using a laser particle counter. All data can be stored in a structured manner according to production batch number, forming a complete data set containing equipment serial number, timestamp, parameter type, and original value, covering the operation records of all equipment in multiple consecutive production cycles. For example, in the third production cycle, the equipment status data management system fully recorded the dynamic process of the vacuum pump in the concentration section gradually increasing its outlet pressure from the initial value of -0.095MPa to -0.088MPa during 72 hours of continuous operation, and simultaneously stored the inlet and outlet water temperature difference data of the cooling water circulation system during this period.

[0015] Step 120: Extract the production line equipment status feature set from the production line equipment status data set. The production line equipment status feature set includes multi-dimensional time-series features reflecting equipment operation trends and topological features reflecting the relationships between equipment.

[0016] In this embodiment, the equipment status data management system employs feature engineering algorithms to perform in-depth analysis of the raw data. For example, for the temperature time-series data of the extraction tank, the system uses a sliding window Fourier transform to extract the frequency domain features for each 15-minute period, while simultaneously calculating the standard deviation between adjacent windows as an index of temperature fluctuation intensity. For the vibration signal of the centrifuge, wavelet packet decomposition is used to extract the energy proportion features of specific frequency bands, and a histogram of vibration energy distribution along the X / Y / Z axes is generated.

[0017] Understandably, at the equipment correlation analysis level, the equipment status data management system establishes an equipment topology map based on the material transport path, calculates the cross-correlation function between the opening duration of the extraction tank's outlet valve and the rotational speed of the concentration tank's feed pump, and quantifies the linkage strength between the two. For example, during feature extraction, when the agitator speed in the alcohol precipitation process increases to 120 rpm, the pressure difference growth rate of its downstream tubular filter exhibits a significant non-linear relationship with the agitation duration; this feature is marked as a key correlation parameter and included in the feature set. Furthermore, the equipment status data management system can also align the key event timestamps of multiple devices using a dynamic time warping algorithm and generate a cross-device state transition matrix.

[0018] Step 130: Train the Astragalus Extract Production Line Status Prediction Model based on the set of production line equipment status features. The Astragalus Extract Production Line Status Prediction Model is used to predict the probability of equipment abnormality and performance degradation trend within a selected time period based on the current equipment status features.

[0019] In this embodiment, the equipment status data management system can construct a dual-channel prediction model comprising a long short-term memory network and a graph convolutional network as a status prediction model for the astragalus extract production line. The temporal feature channel uses a three-layer LSTM structure to process the temperature fluctuation sequence of the extraction tanks, with each layer containing 128 memory units, to learn the temporal propagation law of temperature anomaly patterns. The topological feature channel uses a graph convolutional network to embed the equipment association matrix, capturing potential patterns of multi-equipment collaborative failure.

[0020] For example, during the training process of the Astragalus Extract production line status prediction model, the equipment status data management system uses data from production cycles 1-30 as the training set and data from cycles 31-36 as the validation set, employing an early stop method to prevent overfitting. After 150 rounds of iterative training, the Astragalus Extract production line status prediction model can accurately predict the remaining life of centrifuge bearings. For instance, when the proportion of high-frequency energy in the vibration signal exceeds the threshold baseline for three consecutive time periods, the Astragalus Extract production line status prediction model determines that the probability of the bearing jamming in the following 48 hours of operation has increased to a high-risk level. Regarding the vacuum level index of the concentration tank, the Astragalus Extract production line status prediction model can predict the performance degradation rate within the next 72 hours based on the current rate of decline, providing a basis for preventative maintenance decisions.

[0021] Step 140: Generate a set of equipment management strategies based on the output of the Astragalus Extract production line status prediction model. The set of equipment management strategies includes production parameter adjustment instructions and maintenance node planning schemes for different equipment maintenance priorities.

[0022] In this embodiment, the equipment status data management system generates a graded response strategy based on the output of the Astragalus extract production line status prediction model. For example, when an abnormal fluctuation risk in the extraction tank jacket pressure is predicted, the system automatically generates an optimized temperature control parameter scheme: suggesting adjusting the third-stage heating rate from 2°C per minute to 1.5°C and extending the holding time by 15 minutes. For highly correlated equipment groups, such as the centrifuge and concentrate tank linkage system, the system coordinates and formulates maintenance windows, planning to prioritize the replacement of centrifuge seals during regular batch intervals, while simultaneously arranging vacuum testing of the concentrate tank.

[0023] For example, when the Astragalus Extract production line status prediction model identifies emergency risk equipment, such as a vibrating screen spring support showing a structural fatigue warning, the equipment status data management system immediately generates a red alarm command, triggering a production line speed-down mode and dynamically rescheduling the production tasks of the relevant sections to the standby equipment line. All strategies are virtually verified through a digital twin system before being sent to the on-site control system for execution, ensuring that strategy adjustments do not affect the overall production rhythm.

[0024] As can be seen, the equipment status data management method for automated astragalus extract production lines provided in this application embodiment achieves intelligent monitoring and predictive maintenance throughout the entire process. To further understand the technical solutions described in the above-described embodiments, a complete application scenario example will be used for illustration below.

[0025] Taking a typical application in the production line of Astragalus extract oral liquid of a pharmaceutical company as an example, during the 36th production cycle, the equipment status data management system collected multi-source heterogeneous data in real time through a sensor network deployed in the three major sections of extraction, separation, and concentration. In the raw material extraction stage, temperature sensors recorded the temperature curve of the liquid medium in the extraction tank with minute-level accuracy. When an abnormal fluctuation of ±1.2℃ was detected in the fifth batch of material at the 75℃ constant temperature stage, the equipment status data management system simultaneously retrieved the jacket steam pressure data and found that the pressure value dropped sharply from 0.35MPa to 0.28MPa during the corresponding period. Combined with the hourly injection deviation of 8.7% recorded by the ethanol solvent flow meter, this constituted a multi-dimensional data anomaly event. Simultaneously, the vibration sensor in the centrifugation section, with a sampling rate of 500Hz, captured the Z-axis vibration acceleration of the drum bearing, which decreased from 4.3m / s² in three consecutive batches. 2 Increased to 6.8 m / s 2 Its high-frequency band (8-12kHz) energy ratio exceeded the historical threshold baseline of 15%, triggering in-depth analysis of the feature extraction mechanism.

[0026] In this application scenario, the feature engineering module employs adaptive sliding window technology to perform joint time-frequency domain analysis on the temperature sequence of the extraction tank. For example, by using Fourier transform to identify the 0.05Hz low-frequency interference component hidden in the temperature fluctuations, and combining wavelet packet decomposition algorithm to extract 32 sets of frequency band energy features from the centrifuge vibration signal, it was found that the energy value of the 16th frequency band (corresponding to the bearing cage fault characteristic frequency) is 3.2 times higher than that under normal operating conditions. Furthermore, the equipment correlation analysis layer, based on the material transport topology, calculates that the Pearson correlation coefficient between the vacuum pump outlet pressure of the concentration tank and the cooling water circulation temperature difference reaches -0.87, revealing a strong negative correlation between the two. In addition, the dynamic time warping algorithm further aligns the time lag characteristics of the agitator speed increase event and the downstream filter pressure difference change in the alcohol precipitation process, establishing a nonlinear response model with a 23-second delay between the two, which is marked as a key process control node.

[0027] Next, based on the aforementioned feature set, multi-dimensional state inference can be performed using a dual-channel prediction model. For example, when processing the temperature fluctuation sequence of the extraction tank, the Long Short-Term Memory (LSTM) network captures the similarity between the abnormal fluctuation pattern and the failure case of the third historical production cycle in the memory unit, outputting a 78.6% probability of temperature runaway within the next 24 hours. As another example, the graph convolutional network, through device topology embedding representation, identifies the synergistic failure mode of abnormal centrifuge vibration and decreased vacuum in the concentration system, predicting that if operation continues for more than 48 hours, it will lead to a 42% degradation in the efficiency of the linkage system. During the model validation phase, data from the 31st to 35th production cycles confirmed that the prediction error for the remaining life of the centrifuge bearings was controlled within ±8 hours, and the coefficient of determination R for predicting the performance degradation trend of the vacuum pump was [missing information]. 2 It reached 0.93.

[0028] Then, in response to the high-risk warnings output by the dual-channel prediction model, the strategy generation module of the equipment status data management system initiates a multi-objective optimization algorithm. For the extraction tank temperature control issue, parameter adjustment instructions can be generated: reducing the third-stage heating rate from 2℃ / min to 1.3℃ / min and extending the 75℃ holding period to 1.5 times the original duration. This solution, verified by a digital twin system simulation, can reduce the standard deviation of temperature fluctuations by 64%. Regarding maintenance planning, the equipment status data management system, based on the equipment correlation matrix, coordinates the centrifuge bearing replacement and the concentrator cooler cleaning tasks within the batch interval, reducing production line downtime to 3.2 hours through a dynamic scheduling algorithm. When a structural fatigue warning appears for the vibrating screen spring support, the equipment status data management system immediately triggers a three-level response mechanism: reducing the production line operating speed to 70% of the rated value, activating the standby filter unit to take over production tasks, and pushing maintenance work orders to mobile terminals, guiding technicians to complete flaw detection of key components within 45 minutes. All strategies are seamlessly integrated with the field control system via the OPC-UA protocol, ensuring the stable maintenance of the Operational Continuity of Production (OEE).

[0029] This design significantly improves the reliability and maintenance efficiency of the Astragalus extract production line. For example, during the target production cycle, the equipment status data management system predicts the risk of seal failure of the alcohol precipitation tank agitator shaft dozens of hours in advance. By adjusting process parameters and replacing spare parts in advance, batch material loss can be effectively avoided. Furthermore, this embodiment can significantly shorten unplanned equipment downtime and improve the first-pass yield of products.

[0030] In an optional embodiment, step 120, extracting the production line equipment status feature set from the production line equipment status data set, includes:

[0031] Step 121: Segment the set of production line equipment status data to obtain a subset of equipment status data corresponding to each production cycle.

[0032] In this embodiment, the equipment status data management system performs segmented processing on the production line equipment status data set based on the production cycle. Specifically, it segments and collects the original data according to the start and end timestamps of each production cycle.

[0033] Taking the automated astragalus extract production line as an example, the equipment status data management system defines the data subset corresponding to the third production cycle as all equipment parameters collected from 08:00 on May 12, 2023 to 08:00 on May 15, 2023. This includes 2,880 temperature data points per minute recorded by the temperature sensor in the extraction section, 108 million sampling point data points generated by the centrifuge vibration sensor, and 216 sets of hourly pressure monitoring values ​​from the vacuum pump in the concentration section. Each data subset is associated with the equipment serial number and production batch code to ensure that the time axis of the data of each equipment within the same cycle is strictly aligned, and overlapping data points generated during the transition period between cycles are eliminated.

[0034] Step 122: Extract equipment operation trend features for each subset of equipment status data. The equipment operation trend features include the fluctuation range of production parameters within a continuous time window, the parameter change rate, and the cumulative duration of parameter deviation from the standard threshold.

[0035] In this embodiment, the equipment status data management system extracts multi-dimensional equipment operation trend features from the segmented subsets of equipment status data. For the temperature parameters of the extraction tank, the equipment status data management system uses a sliding time window analysis method to calculate the temperature range within each 15-minute period as the fluctuation amplitude feature. For example, in the 18th hour window of the third production cycle, the temperature rises from 85.3℃ to 87.9℃ and then falls back to 84.1℃, with a range feature value of 3.8℃.

[0036] Meanwhile, the equipment status data management system calculates the rate of change of parameters using first-order difference. For example, if the growth rate of the centrifuge drive motor current reaches 0.12 A / s within 5 consecutive minutes, a current change rate exceeding the limit is generated. For the ethanol solvent flow parameter, the equipment status data management system accumulates statistics on the duration for which the measured value deviates from the process standard value (150 L / h) by more than ±5% per hour. When the flow rate is insufficient for 3 consecutive hours during the concentration stage of the fifth production cycle, the accumulated deviation duration triggers an early warning indicator.

[0037] Step 123: Generate a production line topology network based on the physical connection relationship between the devices, and extract the inter-device association features based on the production line topology network. The inter-device association features include the synchronization index of parameter transmission between adjacent devices and the coupling index of the operating status of the device group.

[0038] In this embodiment of the application, the equipment status data management system constructs a production line topology network based on the material transmission path and signal interaction relationship between the equipment. The production line topology network takes the extraction tank, centrifuge, and concentration tank as nodes and the connecting pipelines and control system signal lines as edges, forming a directed weighted graph structure.

[0039] For example, the equipment status data management system calculates the synchronization index of parameter transmission between adjacent equipment through cross-correlation analysis. For instance, the time lag coefficient between the opening event of the liquid outlet valve of the extraction tank and the speed increase action of the feed pump of the concentration tank is 2.3 seconds, and its synchronization index reaches 0.92.

[0040] For example, regarding the coupling index of the operating status of equipment groups, the equipment status data management system uses principal component analysis to evaluate the collaborative operating efficiency of equipment such as agitators, temperature controllers, and level gauges in the alcohol precipitation section. When the agitator speed increases from 100 rpm to 120 rpm, the coupling index of the associated equipment decreases by 15%, indicating a risk of operating status mismatch.

[0041] Step 124: Fuse the equipment operation trend features with the inter-equipment correlation features to obtain the production line equipment status feature set.

[0042] In this embodiment, the equipment status data management system performs multimodal fusion of equipment operation trend characteristics and inter-equipment correlation characteristics. Specifically, the system normalizes the temperature fluctuation amplitude characteristics of the extraction tank and the vibration energy distribution characteristics of the centrifuge, and generates a cross-equipment joint feature vector through feature concatenation. Simultaneously, the system encodes the synchronization and coupling indices in the topology network into graph embedding vectors, spatially aligning them with the time series features.

[0043] For example, when integrating the feature data of the third production cycle, the equipment status data management system weights and combines the feature of the temperature change rate of the extraction tank with the feature of the vacuum decrease of the concentration tank to form a composite feature that reflects the linkage effect of the extraction-concentration section. The weight coefficient is dynamically adjusted according to the mutual information calculation results.

[0044] In an optional embodiment, step 130, training the Astragalus Extract production line status prediction model based on the set of production line equipment status features, includes:

[0045] Step 131: Divide the set of production line equipment status features into a training dataset and a validation dataset, wherein the training dataset contains historical equipment status features and corresponding equipment maintenance records.

[0046] In this embodiment, the equipment status data management system performs data partitioning on the set of production line equipment status features. 2400 sets of equipment operation trend features and associated topological features extracted from production cycles 1 to 30 are used as the training dataset, including 1856 equipment anomaly labels strictly corresponding to historical maintenance records. 480 sets of feature data generated from production cycles 31 to 36 are used as the validation dataset to evaluate the model's generalization ability. The equipment status data management system ensures consistent distribution ratios of each equipment type in the training and validation sets through stratified sampling. For example, centrifuge-related features account for 23.7% in the training set and 23.5% in the validation set.

[0047] Step 132: Generate a multi-layer neural network model that includes a temporal attention mechanism, which is used to capture the weight distribution of device state features at different time steps.

[0048] In this embodiment of the application, the device status data management system constructs a multi-layer neural network model that includes a temporal attention mechanism. The temporal feature processing channel of the multi-layer neural network model adopts a bidirectional long short-term memory network structure, and the number of its hidden layer units is set to 128.

[0049] For example, the temporal attention mechanism dynamically assigns importance weights to features at each time step using a learnable parameter matrix. For instance, when processing a 72-hour temperature sequence of an extraction tank, the multilayer neural network model assigns an attention weight of 0.68 to the abnormal temperature rise period occurring in the 48th hour, significantly higher than the 0.05-0.12 weights for normal periods. The attention mechanism's output vector and the device-associated feature vector are concatenated in the fusion layer to form the final state representation.

[0050] Step 133: The inter-device association features are processed using a dynamic graph convolution module. The dynamic graph convolution module is used to update the node connection weights of the production line topology network according to the real-time changes in the device operating status.

[0051] In this embodiment, the equipment status data management system deploys a dynamic graph convolution module to process the inter-equipment association features. This module dynamically updates the edge weight parameters of the production line topology network based on real-time operating data. For example, when the vibration amplitude of the centrifuge exceeds a threshold, its association weight with the vacuum pump of the downstream concentrator automatically increases from 0.75 to 0.91. The dynamic graph convolution operation adopts a message passing mechanism. The state update function of each graph node integrates the feature vectors of its neighboring nodes and the edge weights. For example, the state of the concentrator node is determined by a weighted combination of its own vacuum degree feature and the vibration feature of the upstream centrifuge.

[0052] Step 134: Iteratively adjust the model parameters of the multilayer neural network model by jointly optimizing the prediction loss function and the correlation constraint function. The prediction loss function is used to measure the prediction error of the device anomaly probability, and the correlation constraint function is used to constrain the contribution of the correlation features between devices to the prediction result to meet the preset threshold.

[0053] In this embodiment, the equipment status data management system trains a prediction model using a joint optimization strategy. The prediction loss function uses weighted cross-entropy to calculate the prediction error of equipment anomaly probability, applying a 3x weighting coefficient to high-risk equipment categories. The association constraint function uses the L2 norm to limit the contribution of inter-equipment association features to the prediction result to a preset threshold of 0.45. During training, the optimizer synchronously updates the temporal attention parameters and graph convolution weights. For example, in the 120th iteration, the association contribution of centrifuge vibration features is constrained to 0.43, satisfying the preset condition.

[0054] Step 135: When the prediction accuracy of the verification dataset reaches the convergence condition, stop training the multilayer neural network model and output the Astragalus extract production line status prediction model.

[0055] In this embodiment, the equipment status data management system monitors the prediction accuracy trend of the validation dataset. When the accuracy fluctuation range is less than 0.5% for 20 consecutive iterations, the model is deemed to have reached convergence. For example, during the 145th training iteration, the prediction accuracy of the validation set for vacuum leakage in the concentration tank stabilized in the range of 93.7%-94.2%, and the equipment status data management system immediately terminated the training process and exported the model parameters. The trained Astragalus extract production line status prediction model achieved an F1 score of 0.89 for centrifuge bearing jamming events, an improvement of 17% compared to the baseline model.

[0056] In an optional embodiment, step 140, generating a set of equipment management strategies based on the output of the Astragalus extract production line status prediction model, includes:

[0057] Step 141: Based on the distribution of equipment anomaly probabilities output by the Astragalus Extract production line status prediction model, determine the maintenance priority ranking of each piece of equipment.

[0058] In this embodiment, the equipment status data management system generates a maintenance priority ranking based on the anomaly probability distribution output by the prediction model. For example, it uses a risk matrix assessment method, multiplying the anomaly probability value by the equipment criticality coefficient to obtain a priority score. When the anomaly probability of the concentration tank vacuum pump is 0.82 and the criticality coefficient is 1.5, its priority score reaches 1.23, significantly higher than the extraction tank agitator with a score of 0.76. Furthermore, the equipment status data management system generates a list of equipment arranged in descending order of score, with the top three being the centrifuge spindle (1.35), the vacuum pump (1.23), and the ethanol transfer pump (1.12).

[0059] Step 142: Based on the analysis results of the performance degradation trend, classify the equipment maintenance types, including emergency shutdown maintenance, production gap maintenance, and preventive maintenance.

[0060] In this embodiment, the equipment status data management system categorizes maintenance types based on performance degradation trends. For centrifuge bearing components with a predicted remaining lifespan of less than 8 hours, the system marks them as emergency shutdown maintenance, requiring an immediate interruption of production. When the performance degradation curve of the extraction tank seal shows a leakage risk growth rate below a threshold within the next 48 hours, the system classifies it as intermittent maintenance. For equipment with regular degradation characteristics, such as vacuum pumps, the system sets preventative maintenance plans based on historical maintenance cycles, for example, automatically triggering a lubrication maintenance command every 500 hours of operation.

[0061] Step 143: Match different equipment maintenance types with the preset production parameter adjustment rule library to generate parameter adjustment instructions that are adapted to the current production line operating status.

[0062] In this embodiment, the equipment status data management system matches maintenance types with a parameter adjustment rule base. The rule base contains 213 adjustment logic rules defined by process experts. For example, when a centrifuge requires emergency maintenance, the system automatically matches rule CTL-09 to reduce the feed flow rate from 300 L / min to 200 L / min to alleviate equipment load. For preventative maintenance of the alcohol precipitation tank, the system invokes rule ALC-17 to adjust the stirring speed setting from 120 rpm to 110 rpm and extend the mixing time by 8 minutes to compensate for the temporary decrease in stirring efficiency.

[0063] Step 144: Combining the maintenance priority ranking and production plan data, allocate maintenance resources within the preset maintenance time window and generate a maintenance node planning scheme.

[0064] In this embodiment, the equipment status data management system combines production plan data to generate maintenance node planning schemes. For example, it can access the enterprise resource planning system to obtain production batch information for the next 72 hours and prioritize inserting maintenance tasks during scheduled equipment idle periods. For another example, based on the 37th production cycle plan, the system schedules centrifuge seal replacement work during the equipment cleaning period from 02:00 to 04:00 on May 18th, with the required human resources and spare parts inventory locked in advance through the material requirements planning module. For complex maintenance needs across work sections, the equipment status data management system uses critical path analysis to determine the optimal execution sequence, scheduling vacuum pump maintenance and extraction tank overhaul in parallel, reducing overall downtime by 2.3 hours.

[0065] Step 145: Logically verify the parameter adjustment command with the maintenance node planning scheme, and output the equipment management strategy set after eliminating strategies with resource conflicts or parameter contradictions.

[0066] In this embodiment, the equipment status data management system performs a strategy logic verification function. It verifies the compatibility of parameter adjustment commands and maintenance plans through discrete event simulation. For example, when a production batch requires simultaneous cooling of the extraction tank and speed-up of the centrifuge, the system detects a capacity conflict in the cooling water circulation system and automatically postpones the centrifuge speed-up command until the cooling phase. All strategies must be verified through 3D visualization using a digital twin system to ensure no interference between the robotic arm's movement trajectory and the maintenance personnel's workspace. The final output set of equipment management strategies includes 17 parameter adjustment commands and 9 maintenance plan schemes, which are then issued to the field control system for execution after a safety review.

[0067] In an optional embodiment, the method further includes:

[0068] Step 210: Real-time update the latest equipment status data of the automated astragalus extract production line, and extract the current equipment status features corresponding to the latest equipment status data and input them into the astragalus extract production line status prediction model.

[0069] In step 210, the equipment status data management system collects the latest equipment status data in real time through a sensor network deployed in each section of the automated astragalus extract production line, and performs feature extraction and model input preprocessing on the data. Taking the vacuum pump in the concentration section as an example, the equipment status data management system acquires the real-time reading of the vacuum pump outlet pressure sensor at a frequency of seconds, and simultaneously collects the drive motor winding temperature data and the cooling water circulation flow count. For the extraction tank temperature parameters, the equipment status data management system inputs the latest 30-minute liquid phase temperature sequence into the sliding window Fourier transform module to generate a frequency domain feature vector containing the fundamental amplitude and harmonic components.

[0070] Meanwhile, the equipment status data management system calls the wavelet packet decomposition algorithm to process the real-time waveform data uploaded by the centrifuge vibration sensor, extracts the energy proportion feature of the 200-500Hz frequency band, forms a set of current equipment status features with timestamps, and ensures through the data verification module that the feature dimensions strictly match the input layer structure of the Astragalus extract production line status prediction model.

[0071] Step 220: When the probability of an anomaly in the target equipment is predicted to exceed the dynamic threshold by the Astragalus Extract production line status prediction model, a real-time alarm is triggered and the maintenance priority ranking in the equipment management strategy set is updated.

[0072] In step 220, the equipment status data management system inputs the current equipment status characteristics into the trained Astragalus extract production line status prediction model and calculates the abnormal probability index of each piece of equipment in real time. When the abnormal probability value of the vibration energy of the centrifuge drum bearing reaches 0.87, exceeding the dynamic threshold of 0.85, the equipment status data management system immediately triggers a level 3 alarm command, pushes the alarm information to the human-machine interface in the central control room, and raises the centrifuge maintenance level from P2 to P0 in the maintenance priority ranking queue.

[0073] Simultaneously, the equipment status data management system automatically scans the topology network of associated equipment, identifies the vacuum pump of the concentrator tank that has a strong coupling relationship with the centrifuge, and dynamically adjusts its anomaly probability threshold from 0.80 to 0.75, achieving synchronous updates of the risk level of associated equipment. After an alarm event is triggered, the maintenance work order generation module in the equipment management strategy set immediately interrupts the current low-priority task and recalculates the equipment maintenance sequence.

[0074] Step 230: Dynamically adjust the ongoing maintenance node planning scheme according to the updated maintenance priority sorting, and reallocate inactive maintenance resources to the target equipment.

[0075] In step 230, the equipment status data management system dynamically adjusts the maintenance resource allocation scheme according to the updated maintenance priority ranking. When a centrifuge is marked as a P0-level maintenance object, the equipment status data management system automatically allocates the spare lubricant inventory originally planned for the maintenance of the extraction tank agitator to the centrifuge maintenance point in the third workshop, and re-plans the maintenance personnel shift schedule.

[0076] Understandably, for incomplete but initiated concentrator seal replacement work, the equipment status data management system calculates the remaining time for the task using a progress assessment algorithm. If it determines that the task can be completed within 1.5 hours, the current work is maintained; otherwise, the task is paused and the vacuum testing equipment is transferred to the centrifuge section. Furthermore, the equipment status data management system simultaneously updates the maintenance resource status database, marking the assigned electric torque wrenches and vibration analyzers as occupied to prevent other work orders from repeatedly calling them.

[0077] Step 240: Synchronize and calibrate the adjusted maintenance node planning scheme with the current production parameter adjustment instructions to ensure that the production line throughput after parameter adjustment is not lower than the preset guaranteed value.

[0078] In step 240, the equipment status data management system performs a coordinated calibration operation of maintenance plan and production parameters, including: when the centrifuge maintenance plan requires the centrifuge to run at reduced speed, calling the CTL-12 instruction in the production parameter adjustment rule library to reduce the feed pump speed from 300 rpm to 220 rpm, and at the same time extending the stirring time of the alcohol precipitation process by 25 minutes to compensate for the loss of mixing efficiency.

[0079] After adjusting the parameters, the equipment status data management system simulated the production line throughput over the next 8 hours using a digital twin model to verify whether the adjusted configuration could maintain a minimum output of 135 kg of astragalus extract per hour. When the simulation results showed that the throughput dropped to 128 kg during a certain period, the equipment status data management system automatically activated the backup plan: while maintaining the centrifuge speed reduction, a backup concentration tank was temporarily activated to participate in production, and the material flow was redistributed through a dynamic load balancing algorithm to ensure that the overall output was not lower than the preset guaranteed value.

[0080] In an optional embodiment, the method for determining the dynamic threshold includes:

[0081] Step 221: Statistically analyze the distribution of equipment status characteristics during the target time period before the occurrence of historical equipment anomaly events.

[0082] In this embodiment, the equipment status data management system statistically analyzes the characteristic distribution data of historical equipment anomalies and establishes a correlation model between characteristic values ​​and anomalies. Taking a vacuum leak in a concentrator tank as an example, the equipment status data management system retrieves the operating data from the 72 hours preceding eight leak events that occurred within the past 12 months, extracting characteristic parameters such as the rate of vacuum decrease, the increase in cooling water temperature difference, and the amplitude of motor current fluctuations. A kernel density estimation algorithm is used to construct the probability distribution curves of each characteristic during the anomaly precursor period. The average value of the vacuum decrease rate characteristic in the 6 hours before the event is recorded as 0.015 MPa / h, significantly higher than the maximum allowable deviation of 0.008 MPa / h during a normal production cycle.

[0083] Step 222: Calculate the feature value offset of each equipment status feature in the equipment status feature distribution before the occurrence of the historical equipment abnormal event and the maximum allowable offset of the normal production cycle.

[0084] In this embodiment, the equipment status data management system calculates a quantitative index of feature value offset. For the centrifuge vibration energy feature, the maximum allowable offset of this feature within a normal production cycle can be calculated as ±12% of the baseline value. When the actual offset reaches +18% two hours before a historical anomaly occurs, its relative offset coefficient is calculated to be 1.5 (18% / 12%). By traversing all associated features, an offset coefficient matrix can be established, in which the average offset coefficient of the vacuum pump outlet pressure is 1.3, and the peak value of the extraction tank temperature fluctuation coefficient is 2.1, providing data support for subsequent threshold setting.

[0085] Step 223: Determine the critical value of the anomaly probability under different confidence levels based on the cumulative distribution function of the feature value offset.

[0086] In this embodiment, the equipment status data management system determines the anomaly probability threshold based on the cumulative distribution function of feature offsets. Taking the current characteristics of the agitator in an alcohol precipitation tank as an example, the system analyzes current offset data from 1200 normal production periods and plots a cumulative probability distribution curve. When the confidence level is set to 95%, the corresponding current offset threshold is +14.7% of the baseline value, which is mapped to an anomaly probability of 0.82. Furthermore, the system establishes a confidence-threshold comparison table for each feature; for example, the vacuum rate of decrease corresponds to an anomaly probability of 0.91 at 99% confidence and 0.85 at 90% confidence.

[0087] Step 224: Based on the key indicator requirements of the current production stage, dynamically select a confidence level that matches the production line stability requirements, and set the anomaly probability threshold corresponding to the matched confidence level as the dynamic threshold.

[0088] In this embodiment, the equipment status data management system dynamically selects the confidence level based on current production needs. When the Astragalus extract production line is performing high-purity product production tasks, the equipment status data management system raises the production line stability requirement index to Level-4 and automatically selects the anomaly probability threshold corresponding to a 99% confidence level. At this time, the centrifuge vibration energy threshold is adjusted from the conventional 0.83 to 0.78, significantly improving the sensitivity of risk identification. When the production line switches to normal batch production mode, the equipment status data management system restores the 90% confidence level to balance the false alarm rate and detection efficiency. The dynamic threshold adjustment module polls the production planning system every 15 minutes to ensure that the threshold setting is synchronized with the process requirements in real time.

[0089] In an optional embodiment, the method further includes:

[0090] Step 310: Generate a knowledge graph for equipment status management, which includes the relationships between equipment structural attributes, historical maintenance records, and fault solutions.

[0091] In step 310, the equipment status data management system constructs an equipment status management knowledge graph, which integrates structured data and unstructured documents of equipment in the automated Astragalus extract production line to form a multi-dimensional knowledge network.

[0092] Specifically, the equipment status data management system extracts the centrifuge's component structure tree from the equipment asset management module, clarifying the assembly relationships and functional dependencies of sub-components such as the drum, spindle, and seals. For example, seal failure directly leads to a decrease in the vacuum level of the drum cavity. Simultaneously, the system analyzes text records from maintenance work orders over the past three years, employing a bidirectional long short-term memory network model to identify key entities in the fault descriptions. For instance, from work order number CT-2023-045, which describes "centrifuge vibration abnormality accompanied by abnormal noise," the system extracts the faulty equipment component "spindle bearing" and the corresponding repair measure "replace SKF6318 bearing." The system maps these entity relationships into a triplet structure of a knowledge graph, forming associated paths such as "centrifuge - occurrence - spindle wear" and "spindle wear - repair method - bearing replacement," and adds attribute fields for the average repair time of the repair measures (e.g., average time for seal replacement is 2.3 hours) and spare parts consumption (e.g., 2 seals are consumed per replacement).

[0093] Step 320: Match and retrieve the output of the Astragalus Extract production line status prediction model with the equipment status management knowledge graph to obtain a set of candidate maintenance schemes related to the current predicted equipment status.

[0094] In step 320, the equipment status data management system matches the output of the Astragalus membranaceus production line status prediction model with the knowledge graph. When the model predicts an abnormal probability of 0.91 for the centrifuge spindle bearing, the equipment status data management system uses "centrifuge bearing" as the query subject and performs a multi-condition search in the knowledge graph: first, it matches the "bearing wear" node with a fault probability > 0.85, and then retrieves three candidate solutions along the "repair method" edge, including "immediately stop the machine and replace the bearing", "run at reduced speed and inject grease", and "adjust the load and monitor the operation". Further, the equipment status data management system calculates the matching degree between each solution and the current equipment status through a graph neural network. Among them, the "immediate replacement" solution has the highest matching degree score (0.92) because its failure conversion rate when the bearing abnormal probability exceeds 0.9 in historical application scenarios is 98%.

[0095] Step 330: Based on historical evaluation data of the performance of maintenance plans, sort the candidate maintenance plans and prioritize those with higher execution efficiency than the preset benchmark and lower resource consumption than the preset upper limit.

[0096] In step 330, the equipment status data management system optimizes the ranking of candidate solutions based on historical execution data of maintenance plans. Specifically, it can access 1208 execution records from the maintenance work order database to calculate the comprehensive efficiency index of each solution. For the "replacing centrifuge bearing" solution, its average repair time of 2.3 hours corresponds to an efficiency index of 75 points, and its spare parts cost of 2850 yuan corresponds to a resource consumption index of 60 points, resulting in a comprehensive score of (75 × 0.6) + (60 × 0.4) = 69 points. While the "grease injection" solution only takes 0.5 hours (efficiency index of 90 points), its quality coefficient decreases due to a 42% failure recurrence rate, resulting in a final score of 68 points. For example, the equipment status data management system prioritizes solutions with a comprehensive score higher than 70 points. When no suitable option is found, a cross-solution combination optimization process is initiated, such as recommending a combined solution of "grease injection + vibration monitoring," which improves the comprehensive score to 73 points.

[0097] Step 340: Integrate the sorted candidate maintenance schemes with the set of equipment management strategies to generate an enhanced equipment management strategy that includes multi-dimensional decision-making criteria.

[0098] In step 340, the equipment status data management system integrates the optimized maintenance plan with the real-time equipment management strategy in a multi-dimensional manner. For centrifuge bearing early warning events, the "immediate replacement" plan recommended by the knowledge graph and the "reduction to 220 rpm" parameter adjustment instruction generated by the prediction model can be integrated to form an enhanced strategy: an emergency maintenance work order is inserted into the equipment management strategy set, specifying that the bearing replacement be completed within 2 hours; at the same time, a production parameter instruction is issued to reduce the centrifuge speed from 300 rpm to 220 rpm in a stepwise manner to alleviate the bearing load.

[0099] Understandably, the equipment status data management system verifies the feasibility of its strategy using digital twins. Simulations show that during the reduced-speed operation, the production line throughput decreased by 7%, but the capacity gap could be compensated by starting backup centrifuges in parallel, ensuring that the overall output met the target. The final enhanced strategy includes 5 parameter adjustment instructions, 3 maintenance work orders, and 2 resource scheduling plans, which are then executed after logical verification.

[0100] In an optional embodiment, the generation of the device state management knowledge graph in step 310 includes:

[0101] Step 311: Extract the equipment component structure tree and functional dependencies between components from the production line equipment database.

[0102] In step 311, the equipment status data management system extracts equipment hierarchical structure data from the production line equipment database and constructs a component structure tree with topological relationships. Taking the extraction tank of the automated astragalus extract production line as an example, the root node of its component structure tree can be defined as the extraction tank body, and the child nodes include secondary components such as tank jacket, stirring paddle, and temperature sensor. The stirring paddle is further decomposed into tertiary parts such as blades, drive shaft, and coupling.

[0103] Furthermore, the equipment status data management system associates the functional dependencies between components through the bill of materials. For example, when a temperature sensor fails, it automatically marks the directly affected upper-level component as the temperature control system and uses a dependency propagation algorithm to determine that the affected process step is the liquid phase extraction step. For critical equipment such as centrifuges, a dynamic dependency relationship between the drum assembly and the power system is established, recording that when the drive motor power drops to 85% of its rated value, the drum speed will deviate by 3.2%.

[0104] Step 312: Parse the fault description text in the historical work order data, and use entity recognition technology to mark the faulty equipment components and the maintenance measures taken for the faulty equipment components.

[0105] In step 312, the equipment status data management system performs natural language processing on historical maintenance work orders to extract structured knowledge elements. For example, it uses a domain-dictionary-enhanced named entity recognition model to accurately locate equipment components and maintenance actions from unstructured text. For another example, when processing work order number VAC-20230417, the system identifies the faulty equipment component as the "vacuum pump exhaust valve spring" from the fault description "vacuum pump outlet pressure fluctuation exceeds the allowable range, and inspection revealed fatigue fracture of the exhaust valve spring," and the maintenance measure as "replacing the high-temperature alloy spring." Furthermore, the system simultaneously establishes a mapping database of fault modes and solutions, cumulatively marking the correspondence between 1235 fault phenomena and 892 maintenance operations, with a 96% coverage rate for vacuum pump-related fault modes.

[0106] Step 313: Generate a heterogeneous relationship network of equipment components, failure modes, and maintenance measures, and add the average repair time and spare parts consumption attributes corresponding to each maintenance measure to the heterogeneous relationship network.

[0107] In step 313, the equipment status data management system generates a heterogeneous relationship network of equipment components, failure modes, and maintenance measures, and adds quantitative attribute dimensions. Specifically, failure modes such as a broken agitator in the extraction tank or a failed centrifuge seal can be used as intermediate nodes, connecting the equipment component nodes on the left to the maintenance measure nodes on the right. For example, the "centrifuge drum seal" component node points to the "sealing surface wear" failure node through an "occurrence" relationship, and this failure node is then connected to the "replace PTFE seal" measure node through a "repair method" relationship. Furthermore, the equipment status data management system adds attributes such as average repair time, spare parts cost, and tool requirements to each maintenance measure node. For instance, the standard operation time for replacing the sight glass of the concentration tank is 45 minutes, requiring the use of specialized tools such as vacuum silicone grease and a torque wrench.

[0108] Step 314: Learn the embedding representation of the heterogeneous relational network through a graph neural network to obtain a set of embedding vectors that reflect the knowledge correlation of device status management.

[0109] In step 314, the equipment status data management system uses a graph neural network to learn features from the heterogeneous relational network, generating a low-dimensional embedded vector representation. For example, the equipment status data management system designs a heterogeneous graph attention network model to map three types of nodes—equipment components, fault modes, and maintenance measures—to a unified vector space.

[0110] During training, the heterogeneous graph attention network model learned that the embedding vector of the fault node "vacuum pump exhaust valve spring breakage" had a vector cosine similarity of 0.89 with its solution node "spring replacement," while the similarity with the irrelevant node "stirring paddle speed adjustment" was only 0.12. After 50 rounds of iterative training, the embedding vector generated by the equipment status data management system accurately reflects the strong correlation between centrifuge spindle wear and bearing replacement solutions, with a correlation strength quantification value of 0.93, significantly higher than the 0.45-0.67 range of other candidate solutions.

[0111] Step 315: Generate a knowledge graph query interface that supports multi-hop reasoning based on the embedded vector set.

[0112] In step 315, the equipment status data management system constructs a knowledge graph query interface that supports multi-hop reasoning. When a user queries "centrifuge vibration abnormality," the system performs two-hop reasoning using a graph traversal algorithm: first, it locates the "centrifuge vibration abnormality" fault node and retrieves its directly associated primary fault modes such as "spindle imbalance" and "bearing wear"; then, it extends to secondary associated potential causes such as "coupling misalignment" and "dynamic balance failure." The query interface returns three repair paths: path 1 is "vibration abnormality → bearing wear → bearing replacement + dynamic balance correction," with a historical average repair time of 2.5 hours; path 2 is "vibration abnormality → coupling misalignment → coaxiality adjustment," with an average time of 1.2 hours. Simultaneously, the system provides a semantic retrieval function based on embedded vectors. Inputting the keyword "seal leakage" can expand the query to include cross-equipment category related faults such as "centrifuge shaft seal aging" and "extraction tank flange gasket damage."

[0113] In an optional embodiment, the method further includes:

[0114] Step 410: At least some of the equipment corresponding to the automated Astragalus extract production line is monitored by a multimodal sensor array, and multimodal sensor data corresponding to the at least some equipment is collected in real time. The multimodal sensor data includes equipment vibration spectrum data, infrared thermal imaging data and acoustic signature data.

[0115] In this embodiment, the equipment status data management system deploys a multimodal sensor array to comprehensively monitor the key equipment of the automated astragalus extract production line. For the centrifuge equipment, a triaxial vibration sensor is installed in the drum bearing housing to collect vibration spectrum data of 0-10kHz, while an infrared thermal imaging camera is arranged on the surface of the equipment casing to acquire temperature field distribution images at a rate of 5 frames per second.

[0116] In the extraction tank area, the equipment status data management system deploys a directional acoustic sensor array to capture the acoustic waveforms of the equipment operation. The sampling frequency is set to 48kHz to cover the audible frequency range and ultrasonic components. For example, during the third production cycle, the equipment status data management system simultaneously acquires the X-axis vibration acceleration spectrum of the centrifuge, the infrared thermogram of the extraction tank jacket area, and the acoustic waveforms of the alcohol precipitation tank agitator, forming a multimodal sensor data stream with strictly aligned timestamps.

[0117] Step 420: Perform time-frequency joint analysis on the multimodal sensor data to extract the sensitive features of the equipment's mechanical state.

[0118] In step 420, the equipment status data management system performs time-frequency joint analysis on the multimodal sensor data. For the centrifuge vibration spectrum data, discrete wavelet packet transform is used to decompose it into 16 frequency bands, and the Shannon entropy value of the energy distribution in each frequency band is calculated. When early wear occurs in the drum bearing, the system detects that the energy entropy value in the 3.2-4.8kHz frequency band increases from the baseline of 0.45 to 0.68, which is marked as a mechanical wear characteristic.

[0119] In terms of infrared thermal imaging analysis, the equipment status data management system modeled the temperature gradient field of the flange connection area of ​​the concentration tank. When the seal failed, causing heat leakage, the system measured the abnormal area growth rate at 12 square centimeters per minute, exceeding the normal operating threshold of 4 square centimeters. For the acoustic data of the agitator in the alcohol settling tank, the system extracted 128-Vimel cepstral coefficient feature vectors and identified the blade scraping noise pattern using a pre-trained hidden Markov model, achieving a severity score of 8.7 out of 10.

[0120] Step 430: Align the mechanical state sensitive features with the production line equipment state feature set across modal features to obtain the cross-modal feature difference degree.

[0121] In this step, the equipment status data management system performs cross-modal feature alignment analysis, aligning the mechanical status sensitive features extracted by multimodal sensors with the process parameters in the production line equipment status feature set in a spatiotemporal manner. For example, a Pearson correlation analysis is performed on the centrifuge vibration entropy feature and the vacuum level data of the concentrator tank during the same period. When the correlation coefficient is lower than the historical baseline value of 0.6, the cross-modal feature difference is calculated as D = 1 - |r| = 0.55, exceeding the preset tolerance of 0.4. The equipment status data management system locates the anomaly source through the feature difference matrix. When the difference between the vibration feature and the vacuum level feature continues to exceed the standard, it is determined that there is a mismatch between the centrifuge mechanical status and process parameters.

[0122] Step 440: When the cross-modal feature difference exceeds the preset tolerance, trigger the equipment status review process and correct the input features of the Astragalus extract production line status prediction model.

[0123] In step 440, the equipment status data management system triggers the equipment status verification process: when the acoustic signature difference of the agitator in the extraction tank reaches 0.52 (threshold 0.45), a three-level verification mechanism is initiated: first, a high-definition industrial endoscope is used to visually inspect the agitator; second, a vibration analyst is arranged to conduct on-site spectrum verification; and finally, a multi-physics simulation is performed using a digital twin model. After the verification confirms that there is a 3mm crack in the agitator blade, the equipment status data management system corrects the input features of the Astragalus extract production line status prediction model, adjusting the weight of the acoustic signature cepstral coefficient from 0.3 to 0.5, and adding the crack propagation rate as a new feature dimension.

[0124] Step 450: Based on the corrected output of the Astragalus Extract production line status prediction model, regenerate the equipment management correction strategy set, and enter the feature difference traceability tag corresponding to the cross-modal feature difference degree into the generated equipment status management knowledge graph.

[0125] In step 450, the equipment status data management system generates an equipment management correction strategy and updates the knowledge graph. Based on the corrected prediction model output, the anomaly probability of the centrifuge bearing is recalculated, increasing from 0.82 to 0.91, and a preventative replacement strategy is immediately generated. Simultaneously, the equipment status data management system inputs the cross-modal feature difference tracing results into the knowledge graph, establishing a correlation path of "vibration entropy anomaly - acoustic signature cepstrum variation - blade crack," and marking the handling plan when the feature difference threshold is exceeded. It is understandable that after the knowledge graph is updated, the efficiency of retrieving maintenance plans for similar operating conditions can be improved, and the mean time to fault location can be shortened.

[0126] In an optional embodiment, step 420, which involves performing time-frequency joint analysis on the multimodal sensor data to extract sensitive features of the equipment's mechanical state, includes:

[0127] Step 421: Perform wavelet packet decomposition on the vibration spectrum data of the equipment, extract the energy distribution entropy of different frequency bands, and use the energy distribution entropy as an indicator of mechanical wear of the equipment.

[0128] In step 421, the equipment status data management system performs wavelet packet decomposition on the vibration spectrum data, decomposing the centrifuge drive end bearing vibration signal into 16 equal-width frequency bands, calculating the normalized energy value of each frequency band, and generating an energy distribution histogram. When pitting defects appear on the bearing raceway, the equipment status data management system detects a sudden increase in the energy proportion of the 6.4-7.2kHz frequency band from 3.8% in the normal state to 9.2%, and the corresponding energy distribution entropy value rises from 0.32 to 0.71, accurately reflecting the trend of intensified mechanical wear. Furthermore, the equipment status data management system establishes a mapping relationship between the energy entropy value of each frequency band and the remaining life of the equipment. For example, when the energy entropy of the high-frequency band (>5kHz) exceeds 0.65 for 3 consecutive hours, the remaining life of the bearing is determined to be less than 72 hours.

[0129] Step 422: Perform regional temperature gradient analysis on the infrared thermal imaging data to determine the area growth rate of the thermal anomaly region on the device surface.

[0130] In step 422, the equipment status data management system processes the thermodynamic characteristics of the infrared thermal imaging data, performing gridded partitioning of the vacuum pump motor housing. The temperature data of each grid cell constitutes a three-dimensional spatiotemporal tensor. When a partial short circuit occurs in the motor winding, the equipment status data management system detects that the temperature gradient value in region C3 reaches 8℃ / cm, which is 4 times higher than the normal operating condition, and the area of ​​the abnormal region expands at a rate of 9.6 square centimeters per minute. Next, the equipment status data management system determines the direction of fault propagation through thermal image sequence analysis. For example, when the high-temperature area migrates from the rear end cover of the motor to the front bearing housing, it is determined to be a frictional heat generation mode caused by lubrication failure.

[0131] Step 423: Extract Mel-Cepstral Coefficients from the acoustic signature signal data and use a Hidden Markov Model to identify the abnormal noise patterns of the equipment.

[0132] In step 423, the equipment status data management system analyzes the pattern characteristics of the acoustic fingerprint signal: for the cavitation noise of the agitator in the extraction tank, 40-dimensional Mel-Negative cepstral coefficients are extracted to characterize the acoustic fingerprint, and the similarity score between it and the standard acoustic fingerprint template is calculated using a Gaussian mixture model. When the agitator blades deform, the equipment status data management system identifies an enhancement of harmonic components in the 200-800Hz frequency band, with its cepstral coefficient variation index reaching 2.35, triggering an abnormal noise mode alarm. Then, based on the state transition probabilities of the hidden Markov model, it determines whether the current acoustic feature belongs to a fault mode such as blade breakage (probability 0.78) or bearing lubrication deficiency (probability 0.21).

[0133] Step 424: Combine the mechanical wear index of the equipment, the area growth rate, and the abnormal noise mode of the equipment operation to generate a joint health assessment function of multimodal sensitive features; wherein, the joint health assessment function integrates at least the frequency band energy distribution entropy, the thermal anomaly growth rate, and the severity score of the abnormal noise mode.

[0134] In step 424, the equipment status data management system constructs a joint health assessment function based on multimodal sensitive features. For example, a health function is defined for a centrifuge:

[0135] H = 0.4 × E_vib + 0.3 × R_thermal + 0.3 × S_audio, where E_vib represents the vibration energy entropy value, R_thermal represents the thermal anomaly area growth rate, and S_audio is the abnormal noise mode score.

[0136] When the health index H of the centrifuge drum bearing dropped from the baseline value of 0.85 to 0.63, the equipment status data management system detected a vibration entropy value E_vib = 0.71 and a thermal anomaly rate R_thermal = 12cm. 2 / min, abnormal noise score S_audio=7.9, comprehensive calculation H=0.4×0.71+0.3×(12 / 4)+0.3×(7.9 / 10)=0.63, which is lower than the preset threshold of 0.70, triggering the mechanical state abnormality judgment.

[0137] Step 425: When the function value of the joint health assessment function is lower than the preset health threshold, the mechanical state of the equipment is determined to be abnormal and a review task priority label is generated.

[0138] In step 425, the equipment status data management system generates review task priority tags based on the health assessment results. When the health function value of the concentrator agitator is below the threshold for two consecutive hours, the equipment status data management system automatically raises the priority level of the equipment in the maintenance queue. For example, if the system detects that the agitator's health H = 0.65, it calculates the probability of shaft breakage as 38% based on historical fault data, generates a red priority tag, and assigns it to the emergency review queue. Similarly, the equipment status data management system synchronously correlates multimodal feature data, storing the evidence chain such as abnormal vibration entropy, excessive thermal gradient, and abnormal noise feature matching into the review task database.

[0139] As an optional but non-limiting embodiment, step 123, which describes generating a production line topology network based on the physical connections between devices and extracting inter-device association features based on the production line topology network, includes:

[0140] Step 1231: Determine the physical connection relationships between the equipment in the automated Astragalus Extract production line. The physical connection relationships include the docking direction of the material transfer interfaces between the equipment, the sensor signal transmission links, and the power supply dependency paths. Generate a set of adjacent equipment pairs based on the connection status of the input and output ports of each equipment in the physical connection relationships. The set of adjacent equipment pairs includes equipment groups with direct physical connections and the connection interface types of the equipment groups. Generate a production line topology network diagram based on the set of adjacent equipment pairs. The nodes of the production line topology network diagram represent equipment entities, and the edges represent the physical connection relationships between the equipment. The attributes of the edges include the interface type and the signal transmission direction.

[0141] Step 1232: Traverse the adjacent device nodes of each device node in the production line topology network diagram, extract the sequence of operating status parameters of adjacent device pairs in a continuous production cycle, align the timestamps of the operating status parameter sequences, and calculate the synchronization index of parameter transmission between adjacent devices. The synchronization index includes the fluctuation phase difference of the output parameters of the upstream device and the input parameters of the downstream device within the same time window and the correlation coefficient of the parameter change trend.

[0142] Step 1233: Divide the equipment into groups based on the degree centrality of the equipment nodes in the production line topology network diagram. Each equipment group includes key central equipment and its directly connected sub-equipment set. Extract the operating status parameters of each equipment in the same equipment group during the same production cycle, and calculate the coupling degree index of the operating status of the equipment in the group. The coupling degree index includes the similarity measure of the fluctuation of equipment parameters in the group and the co-occurrence frequency of abnormal events.

[0143] Step 1234: Classify the synchronization index and the coupling index according to the equipment and equipment group to generate a set of inter-equipment association features that includes the inter-equipment association strength and group operation dependency; assign weights to the set of inter-equipment association features according to the attributes of the edges in the production line topology network diagram so that the association features corresponding to physical connection relationships with different interface types and signal transmission directions have differentiated contributions.

[0144] Step 1235: Update the attributes of the edges in the production line topology network graph based on the set of inter-device association features after weight allocation to form a dynamic production line topology network that integrates the real-time operating status association degree; extract the association feature vector of each device node from the dynamic production line topology network, the association feature vector includes the synchronicity index of adjacent devices, the coupling degree index of the group to which it belongs, and the weight distribution characteristics of the network edges.

[0145] In step 1231, the equipment status data management system parses the physical connection relationships of the equipment in the automated astragalus extract production line and constructs a precise production line topology network. For example, the equipment status data management system traverses the input / output port configuration tables of core equipment such as extraction tanks, centrifuges, and concentration tanks, identifies the material transfer path where the liquid outlet at the bottom of the extraction tank is connected to the feed inlet of the centrifuge through a DN80 stainless steel pipe, and marks the signal transmission link of the electromagnetic flowmeter installed on the pipe as pointing to the central control system.

[0146] Regarding power supply dependencies, the equipment status data management system confirms that the centrifuge drive motor is powered by the third circuit of power cabinet number three and shares the same power supply busbar with the vacuum pump. Based on this physical connection information, the equipment status data management system generates a set of adjacent equipment pairs, including physical connection groups such as "extraction tank-centrifuge" pair (interface type: material pipeline) and "centrifuge-concentrator" pair (interface type: pneumatic valve control signal). The connection attribute field of each equipment pair records the interface size, media type, and signal transmission direction. For example, the pipeline connection direction from the centrifuge outlet to the concentrate tank feed pump is marked as unidirectional transmission.

[0147] In step 1232, the equipment status data management system calculates the synchronization index of operating parameters between equipment. For example, it can extract the timing data of the extraction tank outlet valve opening and the centrifuge feed pump speed sequence within the third production cycle. By aligning the time axis using a dynamic time warping algorithm, it is found that when the extraction tank valve opening increases to 65%, the centrifuge feed pump speed synchronously increases to 285 rpm after a 2.3-second delay, and the Pearson correlation coefficient of the parameter change trend reaches 0.92. For the parameters of the concentration tank vacuum degree and cooling water circulation flow rate, the equipment status data management system uses window sliding cross-correlation analysis, which detects that for every 0.01 MPa decrease in vacuum pressure, the cooling water flow rate increases by 5.2 m³ / s after 1.8 seconds. 3 / h, the fluctuation phase difference stabilizes within ±0.3 seconds. Then, the equipment status data management system generates a synchronization index matrix for each adjacent equipment pair. For example, the trend correlation coefficient of the "centrifuge-concentrator" pair is 0.87, and the standard deviation of the phase difference is 0.41 seconds, indicating that there is a strong linkage between the two devices.

[0148] In step 1233, the equipment status data management system divides the equipment into groups and calculates the coupling degree index. Based on the degree centrality analysis of the production line topology network nodes, the centrifuge node is determined to have the highest connectivity (degree value = 5), and is designated as a critical centrality device, forming a core equipment group with directly connected extraction tanks, concentration tanks, lubrication systems, etc.

[0149] Furthermore, the equipment status data management system extracted vibration, temperature, and pressure parameters for this group during the fifth production cycle. Using a dynamic time warping algorithm, it calculated the similarity of parameter fluctuations and found that when the centrifuge vibration amplitude increased by 15%, the similarity index of the concentration tank vacuum fluctuation also rose to 0.78. Statistics on the co-occurrence frequency of abnormal events showed that when the centrifuge bearing temperature exceeded 85℃, the probability of abnormal current in the associated extraction tank agitator increased by 62%. Based on this, the equipment status data management system calculated the group coupling index to be 0.69, reflecting the risk level of fault propagation between equipment.

[0150] In steps 1234, the equipment status data management system assigns weights and aggregates the associated features. For example, the equipment status data management system categorizes the synchronization index (correlation coefficient 0.87, phase difference 0.41 seconds) of the "centrifuge-concentrator" pair into the material transport associated feature subset, and assigns a weight coefficient of 0.8 based on the attribute that the interface type is DN100 stainless steel pipe.

[0151] For the "centrifuge-control cabinet" signal transmission connection, its synchronicity index with a trend correlation coefficient of 0.65 is classified into the control signal association subset, with a weighting coefficient set to 0.5. Then, the equipment status data management system establishes a weight allocation rule base, stipulating that the contribution benchmark value for material transmission connections is 0.7, for power supply connections it is 0.6, and for control signal connections it is 0.4, ensuring that the association characteristics of different physical connection types have differentiated influence weights in subsequent analyses.

[0152] In step 1235, the equipment status data management system constructs a dynamic production line topology network and extracts feature vectors. For example, the system updates network edge weights based on real-time operating data. When the centrifuge feed pump speed increases to 300 rpm, the material transfer edge weight between it and the extraction tank dynamically adjusts from 0.8 to 0.9, reflecting the strong correlation under the current operating conditions. The system extracts the associated feature vector of the centrifuge node from the updated network, including a synchronization index of 0.92 with the extraction tank, a coupling degree of 0.78 to its core group, and distribution features such as a material transfer edge weight of 0.9 and a control signal edge weight of 0.6. This feature vector is encoded into a 128-dimensional embedding vector using a graph convolutional neural network and input into the Astragalus extract production line status prediction model. This model accurately predicts that when the centrifuge vibration synchronization index decreases by 10%, the probability of a chain reaction failure in the associated group of equipment increases to 73% within 24 hours.

[0153] It is understandable that steps 1231, through systematic analysis of the physical connection relationships between devices, construct a topological network that accurately reflects the production line structure, achieving full-link visual modeling of material transfer, signal control, and power supply. This step transforms the connection status of device input / output ports into a standardized set of adjacent device pairs, providing a structured data foundation for subsequent correlation analysis. It effectively solves the problem of ambiguous descriptions of connection relationships in traditional device management, enabling precise labeling of key attributes such as physical interface types and signal transmission directions between devices, and establishing a spatial correlation framework for dynamic analysis of device collaborative operation.

[0154] The dynamic feature extraction system formed in steps 1232-1235 enables in-depth mining of device association states from static topology to real-time operational features. Synchronization index calculation reveals the temporal correlation patterns of parameter transmission between devices, and coupling degree index quantifies the fault propagation effect within a group, constructing a multi-dimensional device operational state correlation model. A weight allocation mechanism distinguishes the impact levels of different types of physical connections on the system, and dynamically updated network edge attributes achieve feature fusion of real-time operational data. The final generated association feature vector transforms complex device relationships into a computable feature space, providing intelligent prediction models with multi-dimensional inputs that can characterize both the spatial topology of devices and reflect real-time operational states, significantly improving the spatiotemporal correlation of device state analysis.

[0155] As an optional but not limiting embodiment, step 340, which involves fusing the sorted candidate maintenance schemes with the set of equipment management strategies to generate an enhanced equipment management strategy that includes multi-dimensional decision-making criteria, includes:

[0156] Step 341: Obtain the sorted priority list of candidate maintenance schemes, and extract the maintenance time window and equipment identifier corresponding to the maintenance node planning schemes generated in the equipment management strategy set; establish a mapping relationship between the candidate maintenance schemes and the maintenance time windows in the maintenance node planning schemes based on the maintenance scheme execution efficiency and resource consumption indicators in the maintenance scheme priority list.

[0157] Step 342: Based on the mapping relationship, analyze the matching degree between the resource requirements of each candidate maintenance scheme and the allocated maintenance resources in the maintenance node planning scheme, and generate resource matching degree analysis results; based on the resource matching degree analysis results, detect the maintenance command conflicts of the same equipment in the same time window between the candidate maintenance scheme and the maintenance node planning scheme, and output the conflict detection results.

[0158] Step 343: Based on the conflict detection results, dynamically adjust the candidate maintenance schemes that have resource conflicts or time overlaps. The dynamic adjustment includes reallocating maintenance time windows or splitting maintenance resources to different device identifiers. Perform joint logical verification between the adjusted candidate maintenance schemes and the production parameter adjustment instructions in the device management strategy set to verify whether the adjusted maintenance time windows cause the throughput after the production parameter adjustment to be lower than the preset guarantee value. Based on the verification results, select candidate maintenance schemes that meet the throughput guarantee and have no resource conflicts to generate an optimized set of maintenance schemes.

[0159] Step 344: Extract the equipment structure attributes and historical maintenance records corresponding to each maintenance scheme in the maintenance scheme optimization set, and perform correlation matching with the fault solutions in the equipment status management knowledge graph; based on the correlation matching results, perform a secondary sorting of the schemes in the maintenance scheme optimization set, and prioritize the retention of maintenance schemes that are compatible with the current equipment structure attributes and have a historical repair success rate higher than a preset threshold; perform time series interpolation fusion between the secondary sorted maintenance scheme optimization set and the maintenance node planning schemes in the equipment management strategy set to generate an enhanced equipment management strategy that includes time dimension, resource dimension and equipment compatibility dimension.

[0160] In step 341, the equipment status data management system obtains a sorted priority list of candidate maintenance schemes and extracts the planned maintenance node information from the equipment management strategy set. Taking the centrifuge bearing replacement scheme (priority P0) as an example, the equipment status data management system matches its execution efficiency index (estimated time 2.3 hours) with the equipment idle time period in the maintenance node plan.

[0161] By analyzing resource availability during the equipment cleaning period at the end of the fifth production cycle (May 18, 02:00-04:00), the Equipment Status Data Management System established a mapping relationship between this plan and the time window: allocating 3 maintenance personnel, dedicated hydraulic puller tools, and SKF 6318 bearing inventory. Simultaneously, the extraction tank seal maintenance plan (priority P2) was mapped to the second-best time period, May 19, 10:00-11:30. During this period, the lubricant inventory met the demand, but the vibration analyzer was already occupied by centrifuge maintenance; the Equipment Status Data Management System automatically marked this as requiring resource coordination.

[0162] In step 342, the equipment status data management system performs resource conflict detection and matching degree analysis. For example, the system detects that both the vacuum pump maintenance plan (priority P1) and the concentrate tank overhaul plan require the use of a vacuum detector from 14:00 to 16:00 on May 18th, generating a resource conflict warning code C-087. Through resource matching degree calculation, the vacuum pump maintenance plan has a 92% matching degree with the current spare parts inventory, while the concentrate tank plan only has a 65% matching degree. The system outputs a conflict detection result suggesting that the vacuum pump maintenance should be prioritized. For the centrifuge spindle calibration plan, the system verifies the reservation status of the required dynamic balancing equipment and finds that the equipment is already occupied in the alcohol precipitation section maintenance task, generating a resource mismatch flag R-112 and triggering a dynamic adjustment process.

[0163] In step 343, the equipment status data management system dynamically schedules and optimizes conflicting solutions. For the instrument conflict between the concentrator maintenance and vacuum pump maintenance, the system splits the concentrator solution into two phases: 14:00-15:00 performs mechanical component replacement without a vacuum tester; 15:30-16:30, after the instrument is released, a sealing test is completed. The adjusted solution is verified through a digital twin system, confirming that the parameter adjustment command (increasing the concentrator vacuum setting from -0.092MPa to -0.089MPa) maintains the production line throughput at 133kg / h, higher than the preset guaranteed value of 130kg / h. The system selects six conflict-free solutions to form an optimization set, among which the centrifuge bearing replacement solution is marked as executable due to complete resource matching and successful simulation verification.

[0164] In step 344, the equipment status data management system combines the knowledge graph to optimize solutions and integrate strategies. For example, the equipment status data management system searches the equipment status management knowledge graph and finds that the SKF 6318 centrifuge bearing replacement solution has a success rate of 96% in 23 applications over the past two years and is fully compatible with the current equipment model, so it is ranked first in the optimization set.

[0165] For the extraction tank sealing ring solution, the knowledge graph shows that the average service life of PTFE seals is 38% longer than that of fluororubber seals under high temperature conditions. Based on this, the equipment status data management system generates material upgrade suggestions and updates the solution parameters.

[0166] Ultimately, the equipment status data management system inserted eight optimization plans into the maintenance plan in a time sequence: centrifuge bearing replacement was performed from 02:00 to 04:00 on May 18th, simultaneously adjusting the feed pump speed reduction command; and the extraction tank sealing ring upgrade maintenance was implemented at 10:00 on May 19th, with related adjustments to the stirring parameters of the alcohol precipitation process. The generated enhanced strategy included 14 maintenance instructions with timestamps accurate to the minute, 9 production parameter compensation rules, and 3 cross-shift resource allocation plans, which were distributed to the workshop execution terminal via the OPC UA protocol.

[0167] Therefore, the maintenance strategy optimization mechanism constructed in steps 341-344 achieves intelligent collaboration between equipment maintenance plans and production scheduling. By establishing a dynamic mapping between candidate plans and maintenance time windows, the system can accurately identify the matching relationship between resource demand and supply, and use conflict detection algorithms to avoid resource competition risks in equipment maintenance in advance. The dynamic adjustment mechanism maximizes maintenance efficiency through time window reallocation and resource splitting while ensuring production continuity. Combined with digital twin verification, it ensures stable production capacity after parameter adjustments, forming a set of optimized plans that balance maintenance efficiency and production assurance.

[0168] It is worth mentioning that step 344, through knowledge graph-driven secondary optimization, significantly improves the engineering adaptability of the maintenance strategy. The correlation and matching between equipment structural attributes and historical maintenance records ensures that the optimal solution has equipment compatibility and a high success rate. Time-series interpolation fusion technology transforms discrete maintenance instructions into a time-series operational blueprint. The generated enhanced strategy integrates multi-dimensional decision factors, achieving deep coupling between maintenance resource scheduling, production process compensation, and equipment structural characteristics. This forms a dynamic management solution that can accurately guide on-site execution, effectively solving the pain points of frequent resource conflicts and low solution adaptability in traditional maintenance planning, and significantly improving the systematization level of equipment lifecycle management.

[0169] Based on the same inventive concept, embodiments of this application also provide a device status data management system. See also... Figure 2 As shown, this is a schematic diagram of a possible device status data management system provided in an embodiment of this application. Figure 2 In this system, the equipment status data management system 200 includes a processor 210 and a memory 220. The memory 220 stores computer programs that can be executed by the processor 210. By executing the instructions stored in the memory 220, the processor 210 can perform the steps of the equipment status data management method applied to the automated astragalus extract production line described above.

[0170] Based on the same inventive concept, embodiments of this application provide a computer-readable storage medium including a computer program. When the computer program runs on a device status data management system, it causes the system to perform the steps of the aforementioned device status data management method applied to an automated astragalus extract production line. In some possible implementations, various aspects of the device status data management method for an automated astragalus extract production line provided in this application can also be implemented as a program product including a computer program. When the program product runs on a device status data management system, the program causes the system to perform the steps of the aforementioned device status data management method applied to an automated astragalus extract production line. For example, the device status data management system can perform actions such as... Figure 1 The steps are shown in the figure.

Claims

1. A method for managing equipment status data in an automated astragalus extract production line, characterized in that, The method is executed by the device status data management system, and the method includes: Acquire a set of production line equipment status data, which includes production parameters, operating status parameters, and environmental monitoring parameters of each piece of equipment in multiple consecutive production cycles of the automated Astragalus Extract production line. Extract a set of production line equipment status features from the set of production line equipment status data. The set of production line equipment status features includes multi-dimensional temporal features reflecting equipment operation trends and topological features of inter-equipment relationships. The step of extracting the production line equipment status feature set from the production line equipment status data set includes: The production line equipment status data set is segmented to obtain a subset of equipment status data corresponding to each production cycle; For each subset of equipment status data, equipment operation trend features are extracted. These features include the fluctuation range of production parameters within a continuous time window, the rate of parameter change, and the cumulative duration of parameter deviation from the standard threshold. A production line topology network is generated based on the physical connection relationship between the equipment, and the correlation features between the equipment are extracted based on the production line topology network. The correlation features between the equipment include the synchronization index of parameter transmission between adjacent equipment and the coupling index of the operating status of the equipment group. The equipment operation trend characteristics are fused with the inter-equipment correlation characteristics to obtain the production line equipment status characteristic set; The production line equipment status feature set is used to train the Astragalus Extract production line status prediction model. The Astragalus Extract production line status prediction model is used to predict the probability of equipment abnormality and performance degradation trend within a selected time period based on the current equipment status features. The training of the Astragalus extract production line status prediction model based on the set of production line equipment status features includes: The set of production line equipment status features is divided into a training dataset and a validation dataset, wherein the training dataset includes historical equipment status features and corresponding equipment maintenance records; Generate a multi-layer neural network model that includes a temporal attention mechanism, which is used to capture the weight distribution of device state features at different time steps; The inter-device association features are processed using a dynamic graph convolution module, which is used to update the node connection weights of the production line topology network according to the real-time changes in the device operating status. By jointly optimizing the prediction loss function and the correlation constraint function, the model parameters of the multi-layer neural network model are iteratively adjusted. The prediction loss function is used to measure the prediction error of the device anomaly probability, and the correlation constraint function is used to constrain the contribution of the correlation features between devices to the prediction result to meet a preset threshold. When the prediction accuracy of the verification dataset reaches the convergence condition, the training of the multilayer neural network model is stopped and the Astragalus extract production line status prediction model is output. Based on the output of the Astragalus Extract production line status prediction model, a set of equipment management strategies is generated. The set of equipment management strategies includes production parameter adjustment instructions and maintenance node planning schemes for different equipment maintenance priorities.

2. The method as described in claim 1, characterized in that, The step of generating a set of equipment management strategies based on the output of the Astragalus extract production line status prediction model includes: Based on the distribution of equipment anomaly probabilities output by the Astragalus Extract production line status prediction model, the maintenance priority of each piece of equipment is determined. Based on the analysis results of the performance degradation trend, equipment maintenance types are classified, including emergency shutdown maintenance, production gap maintenance, and preventive maintenance. The system matches different equipment maintenance types with a pre-set production parameter adjustment rule library to generate parameter adjustment instructions that are adapted to the current production line operating status. By combining the maintenance priority ranking and production plan data, maintenance resources are allocated within a preset maintenance time window, and a maintenance node planning scheme is generated. The parameter adjustment instructions are logically verified against the maintenance node planning scheme. After eliminating strategies with resource conflicts or parameter contradictions, the set of equipment management strategies is output.

3. The method as described in claim 2, characterized in that, The method further includes: The latest equipment status data of the automated astragalus extract production line is obtained in real time, and the current equipment status features corresponding to the latest equipment status data are extracted and input into the astragalus extract production line status prediction model. When the probability of an abnormality of the target equipment is predicted to exceed the dynamic threshold by the Astragalus Extract production line status prediction model, a real-time alarm is triggered and the maintenance priority ranking in the equipment management strategy set is updated. The ongoing maintenance node planning scheme is dynamically adjusted based on the updated maintenance priority ranking, and inactive maintenance resources are reallocated to target devices. The adjusted maintenance node planning scheme will be synchronized and calibrated with the current production parameter adjustment instructions to ensure that the throughput of the production line after parameter adjustment is not lower than the preset guarantee value.

4. The method as described in claim 3, characterized in that, The method for determining the dynamic threshold includes: Statistical analysis of the distribution of equipment status characteristics during the target time period prior to the occurrence of historical equipment anomalies; Calculate the feature value offset of each equipment status feature in the equipment status feature distribution before the occurrence of the historical equipment abnormal event and the maximum allowable offset of the normal production cycle; The critical values ​​of anomaly probability at different confidence levels are determined based on the cumulative distribution function of the eigenvalue offsets. Based on the key indicator requirements of the current production stage, a confidence level that matches the stability requirements of the production line is dynamically selected, and the anomaly probability threshold corresponding to the matched confidence level is set as the dynamic threshold.

5. The method as described in claim 1, characterized in that, The method further includes: Generate a knowledge graph for equipment status management, which includes the relationships between equipment structural attributes, historical maintenance records, and fault solutions; The output of the Astragalus extract production line status prediction model is matched and retrieved with the equipment status management knowledge graph to obtain a set of candidate maintenance schemes related to the current predicted equipment status. Based on historical evaluation data of the performance of maintenance plans, candidate maintenance plans are ranked, and maintenance plans with higher execution efficiency than the preset benchmark and lower resource consumption than the preset upper limit are selected first. The ranked candidate maintenance schemes are integrated with the set of equipment management strategies to generate an enhanced equipment management strategy that includes multi-dimensional decision-making criteria.

6. The method as described in claim 5, characterized in that, The generated device status management knowledge graph includes: Extract the equipment component structure tree and functional dependencies between components from the production line equipment database; The fault description text in historical work order data is parsed, and entity recognition technology is used to annotate the faulty equipment components and the maintenance measures taken for the faulty equipment components; Generate a heterogeneous relationship network of equipment components, failure modes, and maintenance measures, and add the average repair time and spare parts consumption attributes corresponding to each maintenance measure to the heterogeneous relationship network; By performing embedding representation learning on the heterogeneous relational network through a graph neural network, a set of embedding vectors reflecting the knowledge correlation of device status management is obtained; A knowledge graph query interface supporting multi-hop reasoning is generated based on the set of embedded vectors.

7. The method as described in claim 1, characterized in that, The method further includes: The automated Astragalus extract production line is monitored by a multimodal sensor array, and multimodal sensor data corresponding to the at least some equipment is collected in real time. The multimodal sensor data includes equipment vibration spectrum data, infrared thermal imaging data and acoustic signature data. Time-frequency joint analysis is performed on the multimodal sensor data to extract the sensitive features of the equipment's mechanical state; The mechanical state sensitive features are aligned with the production line equipment state feature set across modal features to obtain the cross-modal feature difference degree; When the cross-modal feature difference exceeds the preset tolerance, the equipment status review process is triggered and the input features of the Astragalus extract production line status prediction model are corrected. Based on the corrected output of the Astragalus Extract production line status prediction model, a new set of equipment management correction strategies is generated, and the feature difference traceability tags corresponding to the cross-modal feature difference degree are entered into the generated equipment status management knowledge graph. The step of performing time-frequency joint analysis on the multimodal sensor data to extract the sensitive features of the equipment's mechanical state includes: Wavelet packet decomposition is performed on the vibration spectrum data of the equipment to extract the energy distribution entropy of different frequency bands, and the energy distribution entropy is used as an indicator of mechanical wear of the equipment. Regional temperature gradient analysis was performed on the infrared thermal imaging data to determine the area growth rate of the thermal anomaly region on the equipment surface. Mel-frequency cepstral coefficients were extracted from the acoustic signature data, and the abnormal noise patterns of the equipment were identified by combining them with a hidden Markov model. By combining the equipment's mechanical wear index, area growth rate, and abnormal noise mode, a joint health assessment function of multimodal sensitivity features is generated; wherein, the joint health assessment function integrates at least the frequency band energy distribution entropy, thermal anomaly growth rate, and abnormal noise mode severity score; When the function value of the joint health assessment function is lower than the preset health threshold, the equipment is determined to be in abnormal mechanical condition and a review task priority label is generated.

8. A device status data management system, characterized in that, It includes a processor and a memory, wherein the memory stores a computer program that, when executed by the processor, causes the processor to perform the steps of any one of the methods described in claims 1 to 7.

Images