A data fusion method for launch pad health status prediction based on digital twin

Abstract

The present invention discloses a data fusion method for predicting the health status of a launch pad based on digital twins, comprising the following steps: constructing a launch pad virtual entity, constructing a launch pad data source, fusing health status data, fault diagnosis and prediction, and launch pad maintenance decision-making: identifying potential hidden dangers in the launch pad, and predicting and maintaining various components of the launch pad according to previous strategies to reduce potential failures and maximize their use value. The present invention improves the reliability and safety of the launch pad. The digital twin can monitor the status of each part of the launch pad in real time and identify potential failure problems in advance. This preventive maintenance can reduce the occurrence of unexpected failures, improve the safety and reliability of the launch pad, and avoid launch failures caused by failures. Through real-time data analysis, the health status of the launch pad is dynamically updated and possible failures are predicted, and the maintenance plan is adjusted according to actual conditions.

Description

Technical Field

[0001] The present invention relates to the field of aerospace launch technology, and in particular to a data fusion method for predicting the health status of a launch pad based on digital twins. Background Art

[0002] At present, domestic launch sites still basically use traditional methods to restore the status of rockets after launch, especially in the maintenance of launch pads. There are still problems to varying degrees such as scattered maintenance areas, numerous maintenance items, lengthy maintenance time, low maintenance efficiency, and high maintenance costs. These problems have seriously restricted the efficiency of high-density space mission support and the efficient use of launch site resources. Summary of the Invention

[0003] The purpose of the present invention is to provide a data fusion method for predicting the health status of a launch station based on digital twins, thereby solving the aforementioned problems existing in the prior art.

[0004] In order to achieve the above object, the technical solution adopted by the present invention is as follows:

[0005] A data fusion method for predicting the health status of a launch pad based on digital twins includes the following steps:

[0006] S100, constructing the launch pad virtual entity: constructing a virtual entity model, an operating parameter model, and a state data model, keeping the model state synchronized with the physical entity in real time, and continuously iterating and optimizing the virtual entity to meet data consistency and integrity requirements;

[0007] S200, constructing a launch station data source: generating metadata by collecting sensor parameters and acquired equipment status history data, acquired equipment failure data, and acquired equipment maintenance record data;

[0008] S300, Fusion of Health Status Data: Data is integrated by target and stage, and the multi-dimensional heterogeneous data sources of the launch station are processed through cleaning, integration and conversion processes. The data format is unified, junk data is eliminated, and digital entities are input to ensure the unity of virtual and real data.

[0009] S400, Fault Diagnosis and Prediction: First, by extracting launch pad fault characteristic values, synchronously simulating digital entities, and comparing and analyzing the operating results with the launch pad historical status data in the fault knowledge base, the cause of the fault can be determined. Second, by selecting launch pad status parameters with strong correlations at multiple times and under different operating conditions, a model of its operating status is established. Through data mining simulation experiments, the possible fault type and location are predicted.

[0010] S500, Launch Pad Maintenance Decision: Identify potential hidden dangers in the launch pad, and maintain the various components of the launch pad according to previous strategies to reduce potential failures and maximize its use value.

[0011] Preferably, step S300 also includes: data layer fusion: used to perform structured integration of heterogeneous data from different sources, convert the data into RDF unified data format using a multi-source data fusion framework, and then perform information fusion through a data fusion algorithm.

[0012] Preferably, data layer fusion is divided into: pixel layer fusion, feature layer fusion and decision layer fusion;

[0013] Pixel-level fusion is used to divide data analysis into three typical stages: data layer, feature layer, and decision layer;

[0014] Feature layer fusion, used to map feature fusion to subspace;

[0015] The decision layer is integrated, using Logistic regression to predict sentiment text and related images, and finally taking the weighted average of the two prediction probabilities.

[0016] Preferably, step S300 further includes: data association analysis, which involves associating multi-source data of environmental conditions and associated equipment working statuses by designing a data association algorithm, and using association rule mining and machine learning techniques to find potential relationships between the data;

[0017] Preprocess the associated environmental condition data, specifically including: environmental design effect analysis: establish an environmental design effect evaluation index system, and train the environmental design effect through multi-source data analysis;

[0018] Preprocessing of the associated equipment working status data, specifically including: Analysis of the launch pad working efficiency: By designing an equipment working efficiency evaluation model, the equipment performance data is used to analyze the launch pad working efficiency.

[0019] Preferably, the implementation method of the environmental construction effect analysis includes: determining an environmental construction effect evaluation index system, collecting multi-source data with sensors, analyzing and calculating the multi-source data, and then evaluating the environmental construction effect.

[0020] Preferably, the implementation method of the launch station work efficiency analysis includes: designing an equipment work efficiency evaluation model, collecting equipment performance data, analyzing and calculating multi-source data, and then evaluating the equipment work efficiency.

[0021] Preferably, the design equipment work efficiency evaluation model adopts the hierarchical analysis method and the fuzzy comprehensive evaluation method.

[0022] Preferably, the evaluation method of the design equipment work efficiency evaluation model includes:

[0023] The first step is to determine the evaluation index system, including the equipment's normal operating time, operating efficiency and failure rate indicators; then the weight of each indicator is determined through expert scoring and hierarchical analysis method; finally, the comprehensive score of the launch station's working efficiency is calculated based on the indicator value and weight.

[0024] The beneficial effects of the present invention are as follows: the present invention discloses a data fusion method for predicting the health status of a launch pad based on digital twins, comprising the following steps: constructing a virtual entity of the launch pad: constructing a virtual entity model, constructing an operating parameter model, constructing a state data model, keeping the model state synchronized with the physical entity in real time, and continuously iterating and optimizing the virtual entity to meet the requirements of data consistency and integrity; constructing a data source of the launch pad: forming metadata by collecting sensor parameters and acquired equipment status history data, acquired equipment fault data, and acquired equipment maintenance record data; fusing health status data: fusing data by target and stage, processing multiple data of the launch pad through cleaning, integration and conversion processes Dimension heterogeneous data sources, unify data formats, eliminate junk data, input digital entities, and ensure the unity of virtual and real; Fault diagnosis and prediction: First, by extracting the launch pad fault characteristic values, synchronously simulating digital entities, comparing and analyzing the operation results with the launch pad historical status data in the fault knowledge base, and then determining the cause of the fault; Second, by selecting the launch pad status parameters with strong correlation at multiple times and under different working conditions, establish its operating status law model, and predict the possible fault types and fault locations through data mining simulation experiments; Launch pad maintenance decision-making: Find out the potential hidden dangers of the launch pad, and maintain the various components of the launch pad according to the previous strategy prediction, reduce potential faults, and maximize its use value.

[0025] This invention has the following benefits: 1. It improves the reliability and safety of the launch pad. Digital twins can monitor the status of various launch pad components in real time and identify potential faults in advance. This preventative maintenance reduces unexpected failures, improves the safety and reliability of the launch pad, and avoids launch failures caused by faults.

[0026] 2. This invention optimizes maintenance planning and resource allocation. Through real-time data analysis, it dynamically updates the health status of the transmitter and predicts potential failures, allowing maintenance plans to be adjusted based on actual conditions. This changes the traditional fixed-cycle maintenance approach, deploying resources more accurately and efficiently, and reducing unnecessary maintenance interventions.

[0027] 3. This invention reduces maintenance costs. Using digital twin technology, unnecessary maintenance and downtime can be reduced, significantly lowering overall maintenance costs. Furthermore, preventive maintenance can also avoid more expensive major failures or accidents.

[0028] 4. The present invention prolongs the life of the equipment. Real-time monitoring and fault warning help to better manage the workload of the launch station, avoid excessive use or improper operation, extend the service life of the equipment, and reduce the frequency of equipment replacement.

[0029] 5. The present invention improves decision-making efficiency. It uses real-time simulation and analysis of the launch pad's operating status through digital twins, provides intuitive health status displays and fault prediction data, and helps service managers make quick and efficient decisions, thereby continuously improving the scientific and robustness of personnel decision-making.

[0030] 6. This invention supports operations in complex environments. For complex equipment like launch pads, environmental factors (such as climate and vibration) may affect their performance. Digital twin technology can simulate the impact of these complex environmental factors on the launch pad, providing comprehensive data support for equipment maintenance.

[0031] 7. This invention improves technological innovation. The use of digital twin technology for intelligent predictive maintenance is an advanced way of modern equipment management and maintenance. It not only improves the operation and maintenance efficiency of the launch station, but also brings more technological innovations to the industry. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 It is a flow chart of a data fusion method for predicting the health status of a launch station based on digital twins of the present invention;

[0033] Figure 2 is a flow chart of another embodiment of the present invention;

[0034] Figure 3 It is a three-layer data fusion model diagram of the present invention;

[0035] Figure 4 It is the predictive maintenance assurance process based on digital twins of the present invention;

[0036] Figure 5 This is the PHM architecture diagram based on digital twins of the present invention. DETAILED DESCRIPTION

[0037] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention will be further described in detail below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0038] The digital twin-based launch pad health status prediction maps the environmental conditions of the physical world where the launch pad is located into an environmental model in the digital space, and maps the behavioral characteristics and physical and chemical properties of the equipment itself into a multidisciplinary model in the digital space. It integrates the failure physics model at the micro scale and the data update model at the macro scale to achieve a detailed characterization of the launch pad failure mechanism and degradation trend.

[0039] The digital twin-based predictive maintenance of the launch pad's health status is based on digital twin technology and is divided into five main steps: building a virtual entity, building a data source, data fusion, pattern recognition, and maintenance decision-making.

[0040] Reference Figure 1 、 Figure 2 、 Figure 3 、 Figure 4 and Figure 5 The data fusion method for predicting the health status of a launch pad based on digital twins shown in FIG. 1 includes the following steps:

[0041] S100, constructing the launch pad virtual entity: constructing a virtual entity model, an operating parameter model, and a state data model, keeping the model state synchronized with the physical entity in real time, and continuously iterating and optimizing the virtual entity to meet data consistency and integrity requirements;

[0042] S200. Construction of launch pad data source: Metadata is formed by collecting sensor parameters and acquired equipment status historical data, acquired equipment fault data, and acquired equipment maintenance record data. Sensors collect parameters such as the launch pad's mechanical system, electrical system, and external environment. Daily operation and maintenance management provides equipment historical status data and maintenance records. Simulation tools are used to simulate launch pad fault status data in specific scenarios.

[0043] S300, integration of health status data: data integration by target and stage, processing the multi-dimensional heterogeneous data sources of the launch station through cleaning, integration and conversion processes, unifying the data format, eliminating junk data, inputting digital entities, and ensuring the unity of virtual and real; data fusion based on stage division is a fusion strategy based on an overall perspective, which improves the feasibility of the fusion process by decomposing the data fusion work goals.

[0044] Phased data fusion is the fusion of data from different stages in the order of the process. Data from all stages can be fused separately, or data from only a certain stage can be fused.

[0045] S400, Fault Diagnosis and Prediction: First, by extracting launch pad fault characteristic values, synchronously simulating digital entities, and comparing and analyzing the operating results with the launch pad historical status data in the fault knowledge base, the cause of the fault can be determined. Second, by selecting launch pad status parameters with strong correlations at multiple times and under different operating conditions, a model of its operating status is established. Through data mining simulation experiments, the possible fault type and location are predicted.

[0046] S500, Launch Pad Maintenance Decision: Identify potential hidden dangers in the launch pad, and maintain the various components of the launch pad according to previous strategies to reduce potential failures and maximize its use value.

[0047] Preferably, step S300 also includes: data layer fusion: used to perform structured integration of heterogeneous data from different sources, convert the data into RDF (Resource Description Framework) unified data format using a multi-source data fusion framework, and then perform information fusion through a data fusion algorithm.

[0048] Preferably, data layer fusion is divided into: pixel layer fusion, feature layer fusion and decision layer fusion;

[0049] Pixel-level fusion is used to divide data analysis into three typical stages: data layer, feature layer, and decision layer;

[0050] Feature layer fusion is used to map feature fusion into subspaces, such as surface material recognition.

[0051] The decision layer is integrated, using Logistic regression to predict sentiment text and related images, and finally taking the weighted average of the two prediction probabilities.

[0052] Preferably, step S300 also includes: data association analysis, by designing a data association algorithm to associate multi-source data of environmental conditions and associated equipment working status, and using association rule mining and machine learning technology to find out the potential relationship between the data; the implementation method of data association analysis here is to select an association rule mining algorithm or a machine learning algorithm for data association, and use algorithms such as April, FP-Growth, decision tree, random forest and support vector machine to analyze frequent item sets in the data set, and find out the potential association between the data by training the model.

[0053] Preprocess the associated environmental condition data, specifically including: analysis of environmental construction effects: establishing an environmental construction effect evaluation index system, and training environmental construction effects through multi-source data analysis; it should be noted that the data association analysis technology stack here includes association rule mining libraries (such as pyfpgrowth), machine learning frameworks (such as sc i kit-l earn, TensorFlow, PyTorch, etc.) and data analysis libraries (such as pandas, numpy, etc.).

[0054] Preprocessing of the associated equipment working status data, specifically including: Analysis of the launch pad working efficiency: By designing an equipment working efficiency evaluation model, the equipment performance data is used to analyze the launch pad working efficiency.

[0055] Preferably, the environmental configuration effect analysis is implemented by determining an environmental configuration effect evaluation index system, using sensors to collect multi-source data, analyzing and calculating the multi-source data, and then evaluating the environmental configuration effect. For example, by analyzing temperature and humidity data, it is possible to assess whether the environmental configuration meets the normal operation requirements of the transmitter. It should be noted that the environmental configuration effect analysis technology stack includes software (such as UE, Visual Studio 2022), data analysis libraries (such as pandas, numpy), statistical analysis software (such as R, SPSS), and sensor data acquisition software.

[0056] Preferably, the launch pad performance analysis is implemented by designing an equipment performance evaluation model, collecting equipment performance data, and analyzing and calculating multi-source data to evaluate equipment performance. For example, the performance and reliability of the equipment can be evaluated by analyzing data such as the launch pad's uptime, operating efficiency, and failure rate. The adaptability and durability of the equipment can be evaluated by analyzing the launch pad's operating conditions under different environmental conditions.

[0057] Preferably, the design equipment work efficiency evaluation model adopts the hierarchical analysis method and the fuzzy comprehensive evaluation method.

[0058] Preferably, the evaluation method of the design equipment work efficiency evaluation model includes:

[0059] The first step is to determine the evaluation index system, including the equipment's normal operating time, operating efficiency and failure rate indicators; then the weight of each indicator is determined through expert scoring and hierarchical analysis method; finally, the comprehensive score of the launch station's working efficiency is calculated based on the indicator value and weight.

[0060] What needs to be explained here is the launch station work efficiency analysis technology stack, which includes data analysis libraries (such as pandas and numpy), mathematical calculation libraries (such as scipy) and evaluation model libraries. The corresponding library is selected according to the evaluation method, such as the ahpy library for hierarchical analysis method.

[0061] Based on the predictive maintenance of the digital twin launch pad, it is necessary to focus on the condition monitoring, fault diagnosis and remaining service life of the launch pad, and integrate them into the maintenance decision-making of the physical entity.

[0062] Remaining service life can accurately reflect the evolution trend of equipment health status and improve the robustness of maintenance strategies.

[0063] Predictive maintenance of launch pads based on digital twins is essentially a type of sequential decision-making problem. It collects and analyzes launch pad monitoring data in real time, updates its health status, predicts potential failure risks, and dynamically adjusts maintenance plans.

[0064] The beneficial effects of the present invention are as follows: the present invention discloses a data fusion method for predicting the health status of a launch pad based on digital twins, comprising the following steps: constructing a virtual entity of the launch pad: constructing a virtual entity model, constructing an operating parameter model, constructing a state data model, keeping the model state synchronized with the physical entity in real time, and continuously iterating and optimizing the virtual entity to meet the requirements of data consistency and integrity; constructing a data source of the launch pad: forming metadata by collecting sensor parameters and acquired equipment status history data, acquired equipment fault data, and acquired equipment maintenance record data; fusing health status data: fusing data by target and stage, processing multiple data of the launch pad through cleaning, integration and conversion processes Dimension heterogeneous data sources, unify data formats, eliminate junk data, input digital entities, and ensure the unity of virtual and real; Fault diagnosis and prediction: First, by extracting the launch pad fault characteristic values, synchronously simulating digital entities, comparing and analyzing the operation results with the launch pad historical status data in the fault knowledge base, and then determining the cause of the fault; Second, by selecting the launch pad status parameters with strong correlation at multiple times and under different working conditions, establish its operating status law model, and predict the possible fault types and fault locations through data mining simulation experiments; Launch pad maintenance decision-making: Find out the potential hidden dangers of the launch pad, and maintain the various components of the launch pad according to the previous strategy prediction, reduce potential faults, and maximize its use value.

[0065] This invention has the following benefits: 1. It improves the reliability and safety of the launch pad. Digital twins can monitor the status of various launch pad components in real time and identify potential faults in advance. This preventative maintenance reduces unexpected failures, improves the safety and reliability of the launch pad, and avoids launch failures caused by faults.

[0066] 2. This invention optimizes maintenance planning and resource allocation. Through real-time data analysis, it dynamically updates the health status of the transmitter and predicts potential failures, allowing maintenance plans to be adjusted based on actual conditions. This changes the traditional fixed-cycle maintenance approach, deploying resources more accurately and efficiently, and reducing unnecessary maintenance interventions.

[0067] 3. This invention reduces maintenance costs. Using digital twin technology, unnecessary maintenance and downtime can be reduced, significantly lowering overall maintenance costs. Furthermore, preventive maintenance can also avoid more expensive major failures or accidents.

[0068] 4. The present invention prolongs the life of the equipment. Real-time monitoring and fault warning help to better manage the workload of the launch station, avoid excessive use or improper operation, extend the service life of the equipment, and reduce the frequency of equipment replacement.

[0069] 5. The present invention improves decision-making efficiency. It uses real-time simulation and analysis of the launch pad's operating status through digital twins, provides intuitive health status displays and fault prediction data, and helps service managers make quick and efficient decisions, thereby continuously improving the scientific and robustness of personnel decision-making.

[0070] 6. This invention supports operations in complex environments. For complex equipment like launch pads, environmental factors (such as climate and vibration) may affect their performance. Digital twin technology can simulate the impact of these complex environmental factors on the launch pad, providing comprehensive data support for equipment maintenance.

[0071] 7. This invention improves technological innovation. The use of digital twin technology for intelligent predictive maintenance is an advanced way of modern equipment management and maintenance. It not only improves the operation and maintenance efficiency of the launch station, but also brings more technological innovations to the industry.

Claims

1. A data fusion method for predicting the health status of a launch platform based on digital twins, characterized in that: The following steps are involved: S100, constructing the launch pad virtual entity: constructing a virtual entity model, an operating parameter model, and a state data model, keeping the model state synchronized with the physical entity in real time, and continuously iterating and optimizing the virtual entity to meet data consistency and integrity requirements; S200, constructing a launch station data source: generating metadata by collecting sensor parameters and acquired equipment status history data, acquired equipment failure data, and acquired equipment maintenance record data; S300, Fusion of Health Status Data: Data is integrated by target and stage, and the multi-dimensional heterogeneous data sources of the launch station are processed through cleaning, integration and conversion processes. The data format is unified, junk data is eliminated, and digital entities are input to ensure the unity of virtual and real data. S400, Fault Diagnosis and Prediction: First, by extracting launch pad fault characteristic values, synchronously simulating digital entities, and comparing and analyzing the operating results with the launch pad historical status data in the fault knowledge base, the cause of the fault can be determined. Second, by selecting launch pad status parameters with strong correlations at multiple times and under different operating conditions, a model of its operating status is established. Through data mining simulation experiments, the possible fault type and location are predicted. S500, Launch Pad Maintenance Decision: Identify potential hidden dangers in the launch pad, and maintain the various components of the launch pad according to previous strategies to reduce potential failures and maximize its use value.

2. The data fusion method for launch platform health status prediction based on digital twin according to claim 1 is characterized in that: The step S300 also includes: data layer fusion: for structurally integrating heterogeneous data from different sources, converting the data into RDF unified data format using a multi-source data fusion framework, and then performing information fusion through a data fusion algorithm.

3. The data fusion method for predicting the health status of a launch station based on digital twins according to claim 2 is characterized in that: The data layer fusion is divided into: pixel layer fusion, feature layer fusion and decision layer fusion; The pixel-level fusion is used to divide the data analysis work into three typical stages: data layer, feature layer and decision layer; The feature layer fusion is used to map the feature fusion to the subspace; The decision layer fusion uses Logistic regression to predict the sentiment of text and related images, and finally weighted averages the two prediction probabilities.

4. The data fusion method for predicting the health status of a launch station based on digital twins according to claim 3 is characterized in that: The step S300 also includes: data association analysis, which involves designing a data association algorithm to associate multi-source data on environmental conditions and associated equipment working status, and using association rule mining and machine learning techniques to find potential relationships between the data; Preprocess the associated environmental condition data, specifically including: environmental design effect analysis: establish an environmental design effect evaluation index system, and train the environmental design effect through multi-source data analysis; Preprocessing of the associated equipment working status data, specifically including: Analysis of the launch pad working efficiency: By designing an equipment working efficiency evaluation model, the equipment performance data is used to analyze the launch pad working efficiency.

5. The data fusion method for predicting the health status of a launch station based on digital twins according to claim 4 is characterized in that: The implementation method of the environmental construction effect analysis includes: determining an environmental construction effect evaluation index system, collecting multi-source data with sensors, analyzing and calculating the multi-source data, and then evaluating the environmental construction effect.

6. The data fusion method for predicting the health status of a launch station based on digital twins according to claim 4 is characterized in that: The implementation method of the launch station work efficiency analysis includes: designing an equipment work efficiency evaluation model, collecting equipment performance data, analyzing and calculating multi-source data, and then evaluating the equipment work efficiency.

7. The data fusion method for predicting the health status of a launch station based on digital twins according to claim 6 is characterized in that: The design equipment working efficiency evaluation model adopts the hierarchical analysis method and the fuzzy comprehensive evaluation method.

8. The data fusion method for predicting the health status of a launch station based on digital twins according to claim 7 is characterized in that: The evaluation method of the design equipment working efficiency evaluation model includes: The first step is to determine the evaluation index system, including the equipment's normal operating time, operating efficiency and failure rate indicators; then the weight of each indicator is determined through expert scoring and hierarchical analysis method; finally, the comprehensive score of the launch station's working efficiency is calculated based on the indicator value and weight.

Images