Return-type satellite life evaluation method and device, and storage medium

Abstract

The invention belongs to the technical field of spaceflight launching, and particularly relates to a returnable satellite life evaluation method and device and a storage medium, and the method comprises the following steps: obtaining real-time operation data and historical operation data of a satellite, and constructing a data set; new composite features are constructed through the key features and the periodic change features; constructing a life prediction model by taking the gradient boosting tree as a basic model, inputting the composite features corresponding to the historical operation data into the life prediction model for training, and optimizing the performance of the life prediction model by optimizing model parameters; evaluating the prediction accuracy of the model through cross validation to obtain a life prediction model under the optimal parameter combination; and inputting real-time operation data into the trained service life prediction model, adjusting and optimizing the service life prediction model, and predicting the service life through the service life prediction model. According to the prediction method, dependence on expert experience can be reduced, data-driven automatic evaluation is realized, and consistency and reliability of evaluation results are improved.

Description

Technical Field

[0001] The present invention belongs to the field of space launch technology, and in particular relates to a method, a device and a storage medium for assessing the life of a recoverable satellite. Background Art

[0002] The rapid development of modern recoverable satellite technology has made satellite life assessment increasingly important, especially in the context of the increasing number of recoverable satellites. The operating cost of recoverable satellites is high, so evaluating their service life can not only reduce potential operational risks, but also effectively reduce economic losses. At present, the life assessment of recoverable satellites mainly relies on historical data, simulation models and expert experience. However, these traditional methods often fail to fully consider the changing environmental conditions and complex mission requirements, resulting in limitations in the assessment results. At the same time, the impact of environmental factors such as radiation and temperature changes on recoverable satellite materials and electronic equipment is difficult to accurately model. Moreover, the dynamic adaptability of existing assessment models is poor. They are usually based on static assumptions, and the acquisition of real-time monitoring data is often insufficient. It is difficult to dynamically adapt to complex mission requirements and environmental changes, resulting in insufficient flexibility in the assessment, which further affects the accuracy of the assessment. Moreover, traditional methods are usually unable to comprehensively analyze multiple influencing factors, resulting in one-sided assessment results and failure to fully reflect the health status of the satellite. How to implement data-driven satellite life assessment method research is a problem that needs to be solved. Summary of the invention

[0003] In order to solve the above-mentioned technical problem that it is difficult to accurately evaluate the life of a recoverable satellite in the prior art, the present invention proposes a recoverable satellite life assessment method, device and storage medium, in order to develop a data-driven life prediction model based on real-time data and historical data, so as to improve the ability to predict the remaining service life of a recoverable satellite.

[0004] In order to solve the above technical problems, the technical solution adopted by the present invention is: a recoverable satellite life assessment method, comprising the following steps:

[0005] Step 1: Obtain the real-time and historical operation data of the satellite and build a data set;

[0006] Step 2: Perform feature selection on the data in the dataset to identify key features, then extract periodic change features by analyzing time series data, and construct new composite features through key features and periodic change features;

[0007] Step 3: Use the gradient boosting tree as the basic model to build a life prediction model, input the composite features corresponding to the historical operation data into the life prediction model for training, and optimize the performance of the life prediction model by optimizing the model parameters;

[0008] Step 4: Evaluate the prediction accuracy of the model through cross-validation to obtain the life prediction model under the optimal parameter combination;

[0009] Step 5: Input the real-time running data into the trained life prediction model, adjust and optimize it, and perform life prediction through the life prediction model.

[0010] In step 1, the real-time operation data includes power system state parameters, temperature, radiation level and mechanical component state parameters, and the historical operation data includes: fault record data, maintenance record data, and environmental change data.

[0011] In step 2, the key features include battery aging rate and temperature.

[0012] In step 3, the optimized model parameters include: the maximum number of iterations n_estimators, the learning rate learning_rate, and the maximum depth of the decision tree max_depth.

[0013] In step 4, a grid search GridSearchCV is used for cross validation.

[0014] The step 5 also includes the following steps: the life prediction results obtained by the life prediction model provide a decision-making basis for satellite management. At the same time, an early warning system based on the prediction results is established, and the early warning system is used to issue early warnings in a timely manner and formulate maintenance and replacement plans.

[0015] In addition, the present invention also provides a recoverable satellite life assessment device, comprising: a processor and a memory connected to the processor in communication;

[0016] The memory stores computer-executable instructions;

[0017] The processor executes the computer-executable instructions stored in the memory to implement the recoverable satellite life assessment method.

[0018] In addition, the present invention also provides a computer-readable storage medium, in which computer-executable instructions are stored. When the computer-executable instructions are executed by a processor, they are used to implement the recoverable satellite life assessment method.

[0019] Compared with the prior art, the present invention has the following beneficial effects:

[0020] The present invention provides a method for assessing the life of a recoverable satellite. Based on real-time data and historical data, a gradient boosting tree data-driven algorithm is selected to develop a data-driven life prediction model to improve the ability to predict the remaining service life of a recoverable satellite. In addition, the prediction method of the present invention can reduce the dependence on expert experience, realize data-driven automated evaluation, and improve the consistency and reliability of evaluation results. The feasibility has been proven through a series of experiments, simulations, and practical applications. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 A schematic flow chart of a recoverable satellite life assessment method is provided for an embodiment of the present invention. DETAILED DESCRIPTION

[0022] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all the embodiments; based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0023] Embodiment 1

[0024] like Figure 1 As shown, the first embodiment of the present invention provides a recoverable satellite life assessment method, comprising the following steps:

[0025] Step 1: Obtain the real-time and historical operation data of the satellite and build a data set.

[0026] Furthermore, in the present embodiment, in step 1, the real-time operation data includes power system state parameters, temperature, radiation level and mechanical component state parameters, and the historical operation data includes: fault record data, maintenance record data, and environmental change data.

[0027] Real-time monitoring of the satellite's operating status, including power systems, temperature, radiation levels, and mechanical component status. At the same time, systematic collection and integration of historical operating data, including fault records, maintenance records, and environmental change information, can build a comprehensive and rich data set, providing a solid foundation for subsequent analysis.

[0028] Step 2: Perform feature selection on the data in the dataset to identify the key features related to satellite lifespan, then extract periodic changing features by analyzing the time series data, and construct new composite features through key features and periodic changing features.

[0029] In this embodiment, on the one hand, feature selection is performed in the acquired data to identify key features that have a significant impact on the life of the satellite, such as battery aging rate and temperature change. On the other hand, on this basis, by analyzing the time series data and extracting the features of periodic changes, new composite features are constructed to improve the expression ability of the model.

[0030] Step 3: Use the gradient boosting tree (GradientBoostingRegressor) as the basic model to build a life prediction model, input the composite features corresponding to the historical operation data into the life prediction model for training, and optimize the performance of the life prediction model by optimizing the model parameters.

[0031] In this embodiment, a gradient boosting tree data-driven algorithm is selected to build a life prediction model based on data characteristics and prediction requirements. Historical data is used for training, and the model performance is optimized by tuning parameters to ensure that it accurately predicts the remaining service life of the satellite and guarantees the number of reuses of the recoverable satellite.

[0032] The optimized model parameters include: maximum number of iterations n_estimators, learning rate learning_rate, and maximum depth of decision tree max_depth.

[0033] Step 4: Evaluate the prediction accuracy of the model through cross-validation and obtain the life prediction model under the optimal parameter combination.

[0034] Specifically, in this embodiment, a grid search GridSearchCV is used for cross validation. The cross validation method is used to evaluate the prediction accuracy of the model to ensure its stability and generalization ability on different data sets. In practical applications, the prediction effect of the model is continuously monitored, and the model is dynamically adjusted and optimized according to the newly obtained real-time data to improve the adaptability of the model.

[0035] Furthermore, in this embodiment, the model performance is evaluated on the test set, MAE (Mean Absolute Error) and MSE (Mean Squared Error) are calculated, and the optimal parameters are output.

[0036] Step 5: Input the real-time running data into the trained life prediction model, adjust and optimize it, and perform life prediction through the life prediction model.

[0037] Furthermore, in this embodiment, the following steps are also included: the life prediction results obtained by the life prediction model provide a decision-making basis for satellite management. At the same time, an early warning system based on the prediction results is established, and the early warning system is used to issue early warnings in a timely manner to remind personnel at the ground measurement and operation control center to pay attention to the health status of the satellite and formulate maintenance and replacement plans to reduce potential risks and economic losses.

[0038] Embodiment 2

[0039] Embodiment 2 of the present invention provides a recoverable satellite life assessment device, including: a processor and a memory communicatively connected to the processor;

[0040] The memory stores computer-executable instructions;

[0041] The processor executes the computer-executable instructions stored in the memory to implement the recoverable satellite life assessment method described in the first embodiment.

[0042] Embodiment 3

[0043] Embodiment 3 of the present invention provides a computer-readable storage medium, in which computer-executable instructions are stored. When the computer-executable instructions are executed by a processor, they are used to implement the recoverable satellite life assessment method described in Embodiment 1.

[0044] The following is part of the implementation source code:

[0045] import pandas as pd

[0046] import numpy as np

[0047] from sklearn.model_selection import train_test_split,GridSearchCV

[0048] from sklearn.ensemble import GradientBoostingRegressor

[0049] from sklearn.metrics import mean_absolute_error,mean_squared_error

[0050] #Data Acquisition and Integration

[0051] data=pd.read_csv('real_time_data.csv')

[0052] historical_data=pd.read_csv('historical_data.csv')

[0053] combined_data=pd.merge(data,historical_data,on='satellite_id')

[0054] #Feature Engineering

[0055] features=combined_data[['battery_age','temperature','radiation','mechanical_status']]target=combined_data['remaining_life']

[0056] #Model building

[0057] X_train,X_test,y_train,y_test=train_test_split(features,target,test_size=0.2,

[0058] random_state = 42)

[0059] #Select the gradient boosting tree algorithm

[0060] model=GradientBoostingRegressor()

[0061] #Parameter tuning

[0062] param_grid = {

[0063] 'n_estimators':[100,200],

[0064] 'learning_rate':[0.01,0.1,0.2],

[0065] 'max_depth':[3,4,5],

[0066] 'min_samples_split':[2,5,10],

[0067] 'min_samples_leaf':[1,2,4]

[0068] }

[0069] grid_search = GridSearchCV(estimator=model, param_grid=param_grid, cv=5,

[0070] scoring='neg_mean_squared_error')

[0071] grid_search.fit(X_train, y_train)

[0072] # Obtain the best model

[0073] best_model = grid_search.best_estimator_

[0074] # Model evaluation

[0075] predictions = best_model.predict(X_test)

[0076] mae = mean_absolute_error(y_test, predictions)

[0077] mse = mean_squared_error(y_test, predictions)

[0078] print(f'Best Parameters: {grid_search.best_params_}')

[0079] print(f'Mean Absolute Error: {mae}')

[0080] print(f'Mean Squared Error: {mse}')

[0081] # Prediction and decision support

[0082] new_data = pd.DataFrame({

[0083] 'battery_age': [2.0],

[0084] 'temperature': [30.0],

[0085] 'radiation': [0.5],

[0086] 'mechanical_status': [1]

[0087] })

[0088] remaining_life_prediction=best_model.predict(new_data)

[0089] print(f'Predicted Remaining Life:{remaining_life_prediction}')

[0090] In summary, the present invention provides a method for assessing the life of a recoverable satellite. Based on real-time data and historical data, a gradient boosting tree data-driven algorithm is selected to develop a data-driven life prediction model to improve the ability to predict the remaining useful life of a recoverable satellite. In addition, the prediction method of the present invention can reduce dependence on expert experience, realize data-driven automated evaluation, and improve the consistency and reliability of evaluation results. The feasibility has been proven through a series of experiments, simulations, and practical applications.

[0091] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for assessing the life of a recoverable satellite, characterized in that: The following steps are involved: Step 1: Obtain the real-time and historical operation data of the satellite and build a data set; Step 2: Perform feature selection on the data in the dataset to identify key features, then extract periodic change features by analyzing time series data, and construct new composite features through key features and periodic change features; Step 3: Use the gradient boosting tree as the basic model to build a life prediction model, input the composite features corresponding to the historical operation data into the life prediction model for training, and optimize the performance of the life prediction model by optimizing the model parameters; Step 4: Evaluate the prediction accuracy of the model through cross-validation to obtain the life prediction model under the optimal parameter combination; Step 5: Input the real-time running data into the trained life prediction model, adjust and optimize it, and perform life prediction through the life prediction model.

2. A recoverable satellite life assessment method according to claim 1, characterized in that: In step 1, the real-time operation data includes power system state parameters, temperature, radiation level and mechanical component state parameters, and the historical operation data includes: fault record data, maintenance record data, and environmental change data.

3. A recoverable satellite life assessment method according to claim 1, characterized in that: In step 2, the key features include battery aging rate and temperature.

4. A recoverable satellite life assessment method according to claim 1, characterized in that: In step 3, the optimized model parameters include: the maximum number of iterations n_estimators, the learning rate learning_rate, and the maximum depth of the decision tree max_depth.

5. A recoverable satellite life assessment method according to claim 1, characterized in that: In step 4, a grid search GridSearchCV is used for cross validation.

6. A recoverable satellite life assessment method according to claim 1, characterized in that: The step 5 also includes the following steps: the life prediction results obtained by the life prediction model provide a decision-making basis for satellite management. At the same time, an early warning system based on the prediction results is established, and the early warning system is used to issue early warnings in a timely manner and formulate maintenance and replacement plans.

7. A recoverable satellite life assessment device, characterized in that: include: A processor and a memory communicatively connected to the processor; The memory stores computer-executable instructions; The processor executes the computer-executable instructions stored in the memory to implement the recoverable satellite life assessment method according to any one of claims 1 to 6.

8. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer-executable instructions, and when the computer-executable instructions are executed by a processor, they are used to implement the recoverable satellite life assessment method according to any one of claims 1 to 6.

Images