A data-driven air separation unit health evaluation method and system

Abstract

The present invention relates to the technical field of mechanical equipment monitoring, and specifically to a data-driven air separation unit health evaluation method and system, comprising obtaining a first data set of key components of the air separation unit during operation; performing data preprocessing on the first data set by arranging intelligent edge measurement units on the key components of the air separation unit, and identifying and marking operation deviation data and abnormal data; constructing a health evaluation model to analyze the first standard data set and generate a first health evaluation index; constructing a hybrid mechanism simulation model to simulate the operation deviation data and abnormal data, analyze the differences between the operation results, correct the health evaluation model and the physical air separation unit, and correct the first health evaluation index to generate a second health evaluation index of the air separation unit; when the second health evaluation index is lower than the health threshold of the air separation unit, simulating and generating an adaptive operation and maintenance strategy through the hybrid mechanism simulation model.

Description

Technical Field

[0001] The present invention relates to the technical field of mechanical equipment monitoring, and in particular to a data-driven air separation unit health evaluation method and system. Background Art

[0002] With the rapid advancement of industrial automation and informatization, air separation units (ASUs), as core equipment in industrial production, play a vital role in multiple key sectors, including chemical engineering, metallurgy, and energy. However, due to their complex structures, variable operating conditions, and high-load operation, the health of ASUs not only directly impacts production efficiency but is also closely related to system safety and stability.

[0003] Traditional health assessment methods rely primarily on offline analysis of operating data. While this approach can reflect equipment status to a certain extent, it exhibits significant lags in the face of sudden failures, often making it difficult to implement effective maintenance measures in a timely manner. Furthermore, existing technical approaches, such as regular inspections, performance testing, and predictive maintenance based on historical data, while able to ensure normal equipment operation to a certain extent, still have significant limitations. First, these methods often detect problems only after a failure has occurred or performance has significantly degraded, lacking the ability to proactively identify potential early failures, leading to delayed maintenance. Second, given the complex and ever-changing operating conditions of air separation units, a single data analysis method is unable to comprehensively and accurately assess the overall health of the equipment, especially when operating under multiple alternating operating conditions, which can lead to misjudgments or missed detections. Finally, when a failure occurs, existing methods often lack efficient intelligent response capabilities, making it difficult to quickly locate the root cause of the problem, let alone provide dynamically optimized maintenance and adjustment plans.

[0004] Therefore, a data-driven health evaluation method and system for air separation units are proposed. Summary of the Invention

[0005] The object of the present invention is to provide a data-driven health evaluation method and system for an air separation unit, which obtains a first data set of key components of the air separation unit during operation; preprocesses the first data set by arranging intelligent edge measurement units on the key components of the air separation unit, and identifies and marks the operation deviation data and abnormal data; constructs a health evaluation model to analyze the first standard data set and generate a first health evaluation index; constructs a hybrid mechanism simulation model to simulate the operation deviation data and abnormal data, analyzes the differences between the operation results, corrects the health evaluation model and the physical air separation unit, and corrects the first health evaluation index to generate a second health evaluation index of the air separation unit; when the second health evaluation index is lower than the health threshold of the air separation unit, simulates and generates an adaptive operation and maintenance strategy through the hybrid mechanism simulation model.

[0006] To achieve the above object, the present invention provides the following technical solutions:

[0007] A data-driven air separation unit health evaluation method includes:

[0008] Acquire a first data set of a key component of an air separation unit during operation; the first data set includes a historical first data set and a real-time first data set; arrange an intelligent edge computing unit on the key component of the air separation unit; perform data preprocessing on the first data set by the intelligent edge computing unit to obtain a first standard data set, operation deviation data, and abnormal data; and identify and mark the obtained operation deviation data and abnormal data;

[0009] Based on the historical first data set, an air separation unit health evaluation model is constructed, and the real-time first standard data set is analyzed to generate a first health evaluation index of the air separation unit;

[0010] The health evaluation model of the air separation unit is coupled with the physical mechanism model of the air separation unit to construct a hybrid mechanism simulation model;

[0011] Using the hybrid mechanism simulation model to simulate the operation deviation data and abnormal data, analyzing the difference between the simulation results and the actual operation results, and revising the air separation unit health evaluation model and the physical mechanism model;

[0012] Based on the revised hybrid mechanism simulation model, the first health evaluation index is revised to generate a second health evaluation index of the air separation unit;

[0013] When the second health evaluation index is lower than the health threshold of the air separation unit, a hybrid mechanism simulation model is used to simulate and generate an adaptive operation and maintenance strategy to obtain a correction strategy for the air separation unit; the adaptive operation and maintenance strategy includes simulating and optimizing the operating parameters of the air separation unit, triggering preventive maintenance operations, and triggering device shutdown protection in a linked manner.

[0014] Preferably, the intelligent edge calculation unit includes a data preprocessing layer, an anomaly detection and marking layer, a data transmission layer and a data update layer;

[0015] The data preprocessing layer obtains a first standard data set through filtering, denoising, standardization and feature extraction;

[0016] The anomaly detection and labeling layer classifies feature vectors using a pre-trained random forest model to identify abnormal data; analyzes unlabeled data using a clustering model to obtain operational deviation data; and labels and stores the abnormal data and operational deviation data.

[0017] The data transmission layer stores the first standard data set, abnormal data and operation deviation data;

[0018] The data update layer applies the adaptive operation and maintenance strategy data generated by the hybrid mechanism simulation model to the key components of the air separation unit.

[0019] Preferably, the air separation unit health assessment model includes a feature extraction layer, a feature selection layer, a feature fusion unit and a health assessment acquisition layer;

[0020] The feature extraction layer extracts features from the first standard data set through a deep learning model to obtain a first feature vector of a key component; the first feature vector includes time domain features, frequency domain features, and data transmission features;

[0021] The feature selection layer analyzes the first feature vector by principal component analysis to obtain a second feature vector related to healthiness;

[0022] The feature fusion unit obtains fusion features of key components in the air separation device by fusing the second feature vectors;

[0023] The health evaluation acquisition layer obtains first health evaluation data of the air separation unit by analyzing the fusion characteristics of the key components in the air separation unit, optimizes the first health evaluation data of the air separation unit through multivariate regression analysis, and integrates the first optimized health evaluation data of the air separation unit to obtain a first health evaluation index.

[0024] Preferably, the process of obtaining the first health evaluation indicator is:

[0025] S10. Analyze the fusion characteristics of the key components of the air separation unit to obtain the first health evaluation data of the air separation unit; the formula for the first health evaluation data is:

[0026] ;

[0027] in, is the first health evaluation data, is the number of key components in the air separation unit, For the The health of key components, For the The health weight of each key component, is the health calculation function, For the Key components operating parameters;

[0028] S20. Optimize the first health evaluation data of the air separation unit through multiple regression analysis to obtain the first optimized health evaluation data. The specific formula is:

[0029] ;

[0030] in, for The first optimized health evaluation data after standardization, and for The minimum and maximum values ​​of is a natural constant, is the regression coefficient, is the error term;

[0031] S30. Obtain a first health evaluation index based on the first health evaluation data and the first optimized health evaluation data. The specific calculation formula is:

[0032] ;

[0033] in, is the first health evaluation indicator, is the fusion weight of the first health evaluation index.

[0034] Preferably, the hybrid mechanism simulation model corrects the air separation unit health evaluation model and the physical air separation unit process as follows:

[0035] S100. Inputting the operation deviation data and abnormal data into the hybrid mechanism simulation model;

[0036] S200. Analyze input data based on the air separation unit health evaluation model to generate simulated key component health; simulate the operating state of the unit under different operating conditions based on the physical characteristics of the physical air separation unit, and simulate the dynamic behavior of the air separation unit through finite element analysis;

[0037] S300. Output of a simulated air separation unit health evaluation model based on the air separation unit health; obtain the actual air separation unit health based on the physical air separation unit;

[0038] S400 obtains the health score difference and the operating parameter difference, and calculates the error; the error includes the health score error and the operating parameter error;

[0039] S500. Adjust the air separation unit health evaluation model and the physical air separation unit through the error;

[0040] S501. Add the error to the training set of the air separation unit health evaluation model, perform incremental learning on the air separation unit health evaluation model, and increase the penalty for abnormal data points;

[0041] S502. Adjust the parameters in the physical mechanism model to minimize the error between the simulation results and the actual results;

[0042] S600. By iterating steps S100 to S500, when the error is lower than a preset threshold, stop the correction process.

[0043] Preferably, the process of obtaining the second health evaluation index is:

[0044] Calculate the error based on the difference between the hybrid mechanism simulation model and the actual operation data; the error includes the health score error and the operation parameter error;

[0045] The nonlinear correction function is designed according to the error characteristics, and the calculation formula is:

[0046] ;

[0047] in, is the health score error, For the Key components The operating parameter error, 、 、 and are the correction coefficients, is the number of key components in the air separation unit, is the Sigmoid activation function;

[0048] The second health evaluation index is obtained by combining the first health index and the correction amount and fusing them through the weight coefficient;

[0049] ;

[0050] in, is the second health evaluation index, is the fusion weight of the second health evaluation index.

[0051] A data-driven air separation unit health evaluation system includes:

[0052] a data acquisition unit for acquiring a first data set of key components of the air separation unit during operation; the first data set includes a historical first data set and a real-time first data set; an intelligent edge computing unit is arranged on the key components of the air separation unit; the intelligent edge computing unit performs data preprocessing on the first data set to obtain a first standard data set, operation deviation data, and abnormal data; and the acquired operation deviation data and abnormal data are identified and marked;

[0053] A first model building unit constructs an air separation unit health evaluation model based on a first historical data set; and uses the model to analyze a first real-time standard data set to generate a first health evaluation index for the air separation unit;

[0054] The second model construction unit couples the air separation unit health evaluation model with the physical mechanism model of the air separation unit to construct a hybrid mechanism simulation model;

[0055] a second model correction unit, which simulates the operation deviation data and abnormal data using the hybrid mechanism simulation model, analyzes the difference between the simulation results and the actual operation results, and corrects the health evaluation model and the physical mechanism model of the air separation unit;

[0056] A health acquisition unit amends the first health evaluation index based on the amended hybrid mechanism simulation model to generate a second health evaluation index of the air separation unit;

[0057] The adaptive operation and maintenance unit simulates and generates an adaptive operation and maintenance strategy through a hybrid mechanism simulation model when the second health evaluation index is lower than the health threshold of the air separation unit to obtain a correction strategy for the air separation unit; the adaptive operation and maintenance strategy includes simulating and optimizing the operating parameters of the air separation unit, triggering preventive maintenance operations, and triggering device shutdown protection in a linked manner.

[0058] Preferably, the process of obtaining the second health evaluation index is:

[0059] Calculate the error based on the difference between the hybrid mechanism simulation model and the actual operation data; the error includes the health score error and the operation parameter error;

[0060] The nonlinear correction function is designed according to the error characteristics, and the calculation formula is:

[0061] ;

[0062] in, is the health score error, For the Key components The operating parameter error, 、 、 and are the correction coefficients, is the number of key components in the air separation unit, is the Sigmoid activation function;

[0063] The second health evaluation index is obtained by combining the first health index and the correction amount and fusing them through the weight coefficient;

[0064] ;

[0065] in, is the second health evaluation index, is the fusion weight of the second health evaluation index.

[0066] Compared with the prior art, the present invention has the following beneficial effects:

[0067] 1. This invention combines an air separation unit health assessment model with a physical mechanism model to form a hybrid mechanism simulation model, enabling accurate prediction of the unit's operating status. By utilizing machine learning models to mine and analyze multidimensional sensor data, historical data, and real-time data, and by realistically simulating the unit's dynamic behavior under different operating conditions through physical simulation, the model enhances the ability to identify abnormal operating conditions and potential faults, effectively addressing the shortcomings of traditional technologies in modeling accuracy, fault prediction, and anomaly detection.

[0068] 2. The present invention uses an intelligent edge computing unit to perform real-time filtering, denoising, and anomaly detection on the first data set, obtaining a first standard data set and promptly marking and storing operational deviation and anomaly data. This reduces data lag and network burden, and, through collaboration with physical mechanism models, generates corrected health assessment indicators in real time. This shifts air separation unit operation and maintenance from post-hoc response to predictive and adaptive adjustments, significantly improving the efficiency and accuracy of operation and maintenance decisions.

[0069] 3. The adaptive O&M strategy proposed in this invention dynamically corrects deviations between simulation results and actual data. When the second health evaluation indicator falls below a set threshold, it automatically triggers a unit-specific correction strategy. This includes simulation-based optimization of air separation unit operating parameters and coordinated shutdown protection. It also supports preventive maintenance operations and forms a proactive defense against potential faults, thereby reducing the risk of downtime and safety hazards, and comprehensively improving the overall operational efficiency and economic efficiency of the air separation unit. BRIEF DESCRIPTION OF THE DRAWINGS

[0070] Figure 1 A flow chart of a data-driven air separation unit health evaluation method provided by the present invention;

[0071] Figure 2 A schematic diagram of the structure of a data-driven air separation unit health evaluation system provided by the present invention;

[0072] Figure 3 A schematic diagram of obtaining health evaluation provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0073] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0074] See also Figure 1 and Figure 2 The present invention provides a data-driven air separation unit health evaluation method applied to a data-driven air separation unit health evaluation system. The technical solution is as follows:

[0075] A first data set is obtained during the operation of a key component of an air separation unit; the first data set includes a historical first data set and a real-time first data set; an intelligent edge computing unit is arranged on the key component of the air separation unit; the first data set is preprocessed by the intelligent edge computing unit to obtain a first standard data set, operation deviation data, and abnormal data; the obtained operation deviation data and abnormal data are identified and marked; and the data acquisition unit is applied to a data-driven air separation unit health assessment system; the preprocessing includes filtering, denoising, standardization, and generating feature vectors; the first data set consists of operating parameters of the key components; wherein the key components mentioned in the present invention and the first data corresponding to the key components are shown in Table 1;

[0076] Table 1 Key components and corresponding parameters

[0077]

[0078] Furthermore, the intelligent edge calculation unit includes a data preprocessing layer, an anomaly detection and marking layer, a data transmission layer, and a data update layer;

[0079] The data preprocessing layer obtains a first standard data set through filtering, denoising, standardization and feature extraction;

[0080] The anomaly detection and labeling layer classifies feature vectors using a pre-trained random forest model to identify abnormal data; analyzes unlabeled data using a clustering model to obtain operational deviation data; and labels and stores the abnormal data and operational deviation data.

[0081] The data transmission layer stores the first standard data set, abnormal data and operation deviation data;

[0082] The data update layer applies the adaptive operation and maintenance strategy data generated by the hybrid mechanism simulation model to the key components of the air separation unit.

[0083] In the health assessment method of this invention, intelligent edge computing units enable real-time data collection, preprocessing, and anomaly detection, significantly improving data processing efficiency and accuracy. By deploying edge units at key device components, data can be filtered and initially analyzed at the source, reducing bandwidth requirements and latency for data transmission.

[0084] This invention builds intelligent edge computing units on key components to preprocess acquired data and capture abnormal data, ensuring timely identification of abnormal data and providing a reliable data foundation for subsequent health assessment models and operation and maintenance strategies. For specific data, see Table 2.

[0085] Table 2 Data processing efficiency table

[0086]

[0087] Based on a first historical data set, an air separation unit health evaluation model is constructed; the model is used to analyze a first real-time standard data set to generate a first health evaluation index for the air separation unit; and the model is applied to a first model construction unit of a data-driven air separation unit health evaluation system;

[0088] Furthermore, the air separation unit health evaluation model includes a feature extraction layer, a feature selection layer, a feature fusion unit and a health evaluation acquisition layer;

[0089] The feature extraction layer extracts features from the first standard data set through a deep learning model to obtain a first feature vector of a key component; the first feature vector includes time domain features, frequency domain features, and data transmission features;

[0090] The feature selection layer analyzes the first feature vector by principal component analysis to obtain a second feature vector related to health; the second feature vector is obtained by principal component analysis;

[0091] The feature fusion unit obtains fusion features of key components in the air separation device by fusing the second feature vectors;

[0092] The health evaluation acquisition layer obtains first health evaluation data of the air separation unit by analyzing the fusion characteristics of the key components in the air separation unit, optimizes the first health evaluation data of the air separation unit through multivariate regression analysis, and integrates the first optimized health evaluation data of the air separation unit to obtain a first health evaluation index.

[0093] The air separation unit health assessment model uses a combination of machine learning and statistical methods, combined with real-time data and physical mechanism simulation, to accurately assess unit health and predict failures. The model's construction and optimization process encompasses data preprocessing, feature selection, algorithm application, model training and validation, parameter calculation, and ongoing refinement, ensuring its efficiency and reliability in complex industrial environments.

[0094] The health evaluation model provided by the present invention not only improves the accuracy and reliability of health evaluation, but also can respond to potential failures in a timely manner through adaptive operation and maintenance strategies, thus solving the hysteresis problem in detection and maintenance.

[0095] Furthermore, the process of obtaining the first health evaluation index is as follows:

[0096] S10. Analyze the fusion characteristics of the key components of the air separation unit to obtain the first health evaluation data of the air separation unit; the formula for the first health evaluation data is:

[0097] ;

[0098] in, is the first health evaluation data, is the number of key components in the air separation unit, For the The health of key components, For the The health weight of each key component, is the health calculation function, For the Key components operating parameters;

[0099] The health calculation function The specific calculation formula is:

[0100] ;

[0101] in, For the Key components The weight of the operating parameters, For the Key components The optimal operating parameters of the operating parameters, For the Key components The standard deviation of the operating parameters.

[0102] When the operating parameter is equal to the optimal operating parameter, it means that the parameter is in the healthiest state;

[0103] When the operating parameters deviate from the optimal operating parameters, the function value approaches 0 smoothly in a nonlinear manner.

[0104] S20. Optimize the first health evaluation data of the air separation unit through multiple regression analysis to obtain the first optimized health evaluation data. The specific formula is:

[0105] ;

[0106] in, for The first optimized health evaluation data after standardization, and for The minimum and maximum values ​​of is a natural constant, is the regression coefficient, is the error term;

[0107] S30. Obtain a first health evaluation index based on the first health evaluation data and the first optimized health evaluation data. The specific calculation formula is:

[0108] ;

[0109] in, is the first health evaluation indicator, is the fusion weight.

[0110] The first health evaluation index proposed in this paper integrates the health scores of key components in an air separation unit and uses weighted average and multivariate regression analysis to provide a comprehensive quantitative assessment of the entire unit's operating status. This formula not only reflects the operating status of each key component but also ensures the accuracy and reliability of the comprehensive index through optimized weighting coefficients and regression models.

[0111] The air separation unit health evaluation model is coupled with the physical mechanism model of the air separation unit to construct a hybrid mechanism simulation model; and the model is applied to a second model construction unit of a data-driven air separation unit health evaluation system;

[0112] The hybrid mechanism simulation model is used to simulate operation deviation data and abnormal data, analyze the differences between the simulation results and the actual operation results, and revise the air separation unit health evaluation model and the physical mechanism model; and is applied to a second model correction unit of a data-driven air separation unit health evaluation system;

[0113] Furthermore, the hybrid mechanism simulation model modifies the air separation unit health evaluation model and the physical air separation unit process as follows:

[0114] S100. Inputting the operation deviation data and abnormal data into the hybrid mechanism simulation model;

[0115] S200. Analyze input data based on the air separation unit health evaluation model to generate simulated key component health; simulate the operating state of the unit under different operating conditions based on the physical characteristics of the physical air separation unit, and simulate the dynamic behavior of the air separation unit through finite element analysis;

[0116] S300. Output of a simulated air separation unit health evaluation model based on the air separation unit health; obtain the actual air separation unit health based on the physical air separation unit;

[0117] S400 obtains the health score difference and the operating parameter difference, and calculates the error; the error includes the health score error and the operating parameter error;

[0118] S500. Adjust the air separation unit health evaluation model and the physical air separation unit through the error;

[0119] S501. Add the error to the training set of the air separation unit health evaluation model, perform incremental learning on the air separation unit health evaluation model, and increase the penalty for abnormal data points;

[0120] S502. Adjust the parameters in the physical mechanism model to minimize the error between the simulation results and the actual results;

[0121] S600. By iterating steps S100 to S500, when the error is lower than a preset threshold, stop the correction process.

[0122] By combining data-driven models with physical mechanism models, the health model not only captures potential patterns and anomalies in the data but also provides a deeper understanding of the operating mechanisms through physical simulation. This approach effectively addresses the lag in inspection and maintenance of traditional air separation units, significantly improving their operational efficiency and reliability, and providing strong technical support for industrial automation and intelligent maintenance.

[0123] Based on the modified hybrid mechanism simulation model, the first health evaluation index is modified to generate a second health evaluation index for the air separation unit; and the second health evaluation index is applied to a health acquisition unit of a data-driven air separation unit health evaluation system.

[0124] Furthermore, the second health evaluation index acquisition process is:

[0125] Calculate the error based on the difference between the hybrid mechanism simulation model and the actual operation data; the error includes the health score error and the operation parameter error;

[0126] The nonlinear correction function is designed according to the error characteristics, and the calculation formula is:

[0127] ;

[0128] in, is the health score error, For the Key components The operating parameter error, 、 、 and are the correction coefficients, is the number of key components in the air separation unit, is the Sigmoid activation function;

[0129] The second health evaluation index is obtained by combining the first health index and the correction amount and fusing them through the weight coefficient;

[0130] ;

[0131] in, is the second health evaluation index, is the fusion weight of the second health evaluation index.

[0132] The second health evaluation index provided by the present invention achieves an accurate assessment of the health of the air separation unit by combining the first health evaluation index with a correction value calculated based on the difference between the simulation model and the actual operating data. The design of the nonlinear correction function enables the correction process to flexibly respond to errors of different types and degrees, and the integration of weight coefficients ensures the stability and accuracy of the comprehensive index. This method improves the real-time and reliability of health evaluation by dynamically adjusting the health score, providing strong support for the efficient operation and preventive maintenance of the air separation unit; for the specific health evaluation acquisition process, please refer to Figure 3 .

[0133] When the second health evaluation indicator falls below the ASU health threshold, a hybrid mechanism simulation model is used to simulate and generate an adaptive operation and maintenance strategy to obtain a corrective ASU strategy. This adaptive operation and maintenance strategy includes simulating and optimizing the ASU operating parameters, triggering preventive maintenance operations, and triggering a shutdown protection mechanism. This system is applied to an adaptive operation and maintenance unit in a data-driven ASU health evaluation system. When the health indicator is above the ASU health threshold, the ASU health evaluation is output.

[0134] The present invention first realizes the real-time collection and preprocessing of historical data and real-time data during operation by arranging intelligent edge measurement units on key components; at the same time, it uses pre-trained random forest models and cluster analysis methods to identify and mark operation deviation data and abnormal data, thereby improving the accuracy and response speed of abnormality detection. Secondly, based on the health evaluation model constructed by deep learning and multivariate regression analysis, combined with the hybrid mechanism simulation model coupled with the physical mechanism model, the operation data is analyzed to generate a first health evaluation index. By analyzing the differences between the simulation results and the actual operation results, the health evaluation model and the physical model are dynamically corrected to improve the accuracy and reliability of the health assessment. In addition, based on the corrected hybrid mechanism simulation model, the first health evaluation index is nonlinearly corrected to generate a more accurate second health evaluation index. The automated operation and maintenance strategy effectively reduces the risk of failure and maintenance costs, and significantly improves the operating efficiency and reliability of the air separation unit, see Table 3.

[0135] Table 3 Operation and maintenance strategy optimization table

[0136]

[0137] Table 3 Based on the revised health evaluation index, the adaptive operation and maintenance strategy generated by the present invention effectively reduces equipment downtime, increases the number of preventive maintenance, reduces maintenance costs, and improves the operating efficiency and economy of the air separation unit.

[0138] Example 2

[0139] The present invention provides a data-driven air separation unit health assessment method and system, based on the first embodiment, further comprising coupling the air separation unit health assessment model with a physical mechanism model of the air separation unit to construct a hybrid mechanism simulation model; wherein the coupling process includes:

[0140] 1) Data synchronization and integration

[0141] In this embodiment, real-time data exchange between the two models is achieved through a unified data interface. Key operating parameters (such as temperature, pressure, and flow) output by the health assessment model are passed as input to the physical mechanism model. Simultaneously, simulation results (such as system response and energy consumption) from the physical mechanism model are fed back to the health assessment model.

[0142] 2) Model interface design

[0143] To achieve seamless coupling between the two models, a standardized interface protocol was designed. Data exchange is performed in JSON or XML format, ensuring compatibility and integrity during transmission. Furthermore, the interface defines the data transmission format, frequency, and error handling mechanisms to ensure the stability and reliability of data exchange.

[0144] 3) Parameter coordination and calibration

[0145] Establish a correspondence between the health assessment model parameters and the physical mechanism model parameters. Map the pressure coefficient in the health assessment model to the pressure transfer parameter in the physical mechanism model. Perform initial parameter calibration on the coupled model based on historical operating data. Using the least squares method, adjust key parameters to minimize the error between simulation results and actual operating data.

[0146] 4) Comparison and analysis of simulation results

[0147] During the simulation process, the outputs of the health assessment model and the physical mechanism model are compared in real time. For example, the temperature and pressure changes of the two models under the same operating conditions are monitored and the differences are calculated. Statistical indicators such as mean square error and coefficient of determination are used to evaluate the model fit and identify systematic deviations and outliers.

[0148] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. A data-driven air separation unit health evaluation method, characterized in that: include: Acquire a first data set of key components of an air separation unit during operation; The first data set includes a historical first data set and a real-time first data set; Arrange intelligent edge computing units on key components of the air separation unit; Performing data preprocessing on the first data set by the intelligent edge computing unit to obtain a first standard data set, operation deviation data, and abnormal data; Identify and mark the acquired operational deviation data and abnormal data; Based on the historical first data set, an air separation unit health evaluation model is constructed, and the real-time first standard data set is analyzed to generate a first health evaluation index of the air separation unit; The health evaluation model of the air separation unit is coupled with the physical mechanism model of the air separation unit to construct a hybrid mechanism simulation model; Using the hybrid mechanism simulation model to simulate the operation deviation data and abnormal data, analyzing the difference between the simulation results and the actual operation results, and revising the air separation unit health evaluation model and the physical mechanism model; Based on the revised hybrid mechanism simulation model, the first health evaluation index is revised to generate a second health evaluation index of the air separation unit; When the second health evaluation index is lower than the health threshold of the air separation unit, simulating and generating an adaptive operation and maintenance strategy through a hybrid mechanism simulation model to obtain a correction strategy for the air separation unit; The adaptive operation and maintenance strategy includes simulating and optimizing the operating parameters of the air separation unit, triggering preventive maintenance operations, and linking and triggering unit shutdown protection.

2. The data-driven air separation unit health evaluation method according to claim 1, characterized in that: The intelligent edge calculation unit includes a data preprocessing layer, an anomaly detection and marking layer, a data transmission layer and a data update layer; The data preprocessing layer obtains a first standard data set through filtering, denoising, standardization and feature extraction; The anomaly detection and labeling layer classifies feature vectors using a pre-trained random forest model to identify abnormal data; analyzes unlabeled data using a clustering model to obtain operational deviation data; and labels and stores the abnormal data and operational deviation data. The data transmission layer stores the first standard data set, abnormal data and operation deviation data; The data update layer applies the adaptive operation and maintenance strategy data generated by the hybrid mechanism simulation model to the key components of the air separation unit.

3. The data-driven air separation unit health evaluation method according to claim 1, characterized in that: The air separation unit health evaluation model includes a feature extraction layer, a feature selection layer, a feature fusion unit and a health evaluation acquisition layer; The feature extraction layer extracts features from the first standard data set through a deep learning model to obtain a first feature vector of a key component; the first feature vector includes time domain features, frequency domain features, and data transmission features; The feature selection layer analyzes the first feature vector by principal component analysis to obtain a second feature vector related to healthiness; The feature fusion unit obtains fusion features of key components in the air separation device by fusing the second feature vectors; The health evaluation acquisition layer obtains first health evaluation data of the air separation unit by analyzing the fusion characteristics of the key components in the air separation unit, optimizes the first health evaluation data of the air separation unit through multivariate regression analysis, and integrates the first optimized health evaluation data of the air separation unit to obtain a first health evaluation index.

4. The data-driven air separation unit health assessment method according to claim 3, characterized in that: The process of obtaining the first health evaluation indicator is as follows: S10. Analyze the fusion characteristics of the key components of the air separation unit to obtain the first health evaluation data of the air separation unit; the formula for the first health evaluation data is: ； in, is the first health evaluation data, is the number of key components in the air separation unit, For the The health of key components, For the The health weight of each key component, is the health calculation function, For the Key components operating parameters; S20. Optimize the first health evaluation data of the air separation unit through multiple regression analysis to obtain the first optimized health evaluation data. The specific formula is: ； in, for The first optimized health evaluation data after standardization, and for The minimum and maximum values ​​of is a natural constant, is the regression coefficient, is the error term; S30. Obtain a first health evaluation index based on the first health evaluation and the first optimized health evaluation data. The specific calculation formula is: ； in, is the first health evaluation indicator, is the fusion weight of the first health evaluation index.

5. The data-driven air separation unit health assessment method according to claim 1, characterized in that: The hybrid mechanism simulation model modifies the air separation unit health evaluation model and the physical air separation unit process as follows: S100. Inputting the operation deviation data and abnormal data into the hybrid mechanism simulation model; S200. Analyze input data based on the air separation unit health evaluation model to generate simulated health of key components; simulate the operating state of the unit under different operating conditions based on the physical characteristics of the physical air separation unit, and simulate the dynamic behavior of the air separation unit through finite element analysis; S300. Output of a simulated air separation unit health evaluation model based on the air separation unit health; obtain the actual air separation unit health based on the physical air separation unit; S400 obtains the health score difference and the operating parameter difference, and calculates the error; the error includes the health score error and the operating parameter error; S500. Adjust the air separation unit health evaluation model and the physical air separation unit through the error; S501. Add the error to the training set of the air separation unit health evaluation model, perform incremental learning on the air separation unit health evaluation model, and increase the penalty for abnormal data points; S502. Adjust the parameters in the physical mechanism model to minimize the error between the simulation results and the actual results; S600. By iterating steps S100 to S500, when the error is lower than a preset threshold, stop the correction process.

6. The data-driven air separation unit health assessment method according to claim 1, characterized in that: The process of obtaining the second health evaluation indicator is as follows: Calculate the error based on the difference between the hybrid mechanism simulation model and the actual operation data; the error includes the health score error and the operation parameter error; The nonlinear correction function is designed according to the error characteristics, and the calculation formula is: ； in, is the health score error, For the Key components The operating parameter error, 、 、 and are the correction coefficients, is the number of key components in the air separation unit, is the Sigmoid activation function; The second health evaluation index is obtained by combining the first health index and the correction amount and fusing them through the weight coefficient; ； in, is the second health evaluation index, is the fusion weight of the second health evaluation index, It is the first health evaluation indicator.

7. A data-driven air separation unit health evaluation system, characterized in that: include: A data acquisition unit, which acquires a first data set of key components of the air separation unit during operation; The first data set includes a historical first data set and a real-time first data set; Arrange intelligent edge computing units on key components of the air separation unit; Preprocessing the first data set by the intelligent edge computing unit to obtain a first standard data set, operation deviation data, and abnormal data; identifying and marking the obtained operation deviation data and abnormal data; a first model building unit, which builds an air separation unit health evaluation model based on a first historical data set, analyzes a first real-time standard data set, and generates a first health evaluation index of the air separation unit; The second model construction unit couples the air separation unit health evaluation model with the physical mechanism model of the air separation unit to construct a hybrid mechanism simulation model; a second model correction unit, which simulates the operation deviation data and abnormal data using the hybrid mechanism simulation model, analyzes the difference between the simulation results and the actual operation results, and corrects the health evaluation model and the physical mechanism model of the air separation unit; A health acquisition unit amends the first health evaluation index based on the amended hybrid mechanism simulation model to generate a second health evaluation index of the air separation unit; The adaptive operation and maintenance unit simulates and generates an adaptive operation and maintenance strategy through a hybrid mechanism simulation model to obtain a correction strategy for the air separation unit when the second health evaluation indicator is lower than a health threshold of the air separation unit; The adaptive operation and maintenance strategy includes simulating and optimizing the operating parameters of the air separation unit, triggering preventive maintenance operations, and linking and triggering unit shutdown protection.

8. The data-driven air separation unit health evaluation system according to claim 7, characterized in that: The process of obtaining the second health evaluation indicator is as follows: Calculate the error based on the difference between the hybrid mechanism simulation model and the actual operation data; the error includes the health score error and the operation parameter error; The nonlinear correction function is designed according to the error characteristics, and the calculation formula is: ； in, is the health score error, For the Key components The operating parameter error, 、 、 and are the correction coefficients, is the number of key components in the air separation unit, is the Sigmoid activation function; The second health evaluation index is obtained by combining the first health index and the correction amount and fusing them through the weight coefficient; ； in, is the second health evaluation index, is the fusion weight of the second health evaluation index, It is the first health evaluation indicator.

Images