Satellite Battery SOC Estimation During Ground-Station Blackouts
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing satellite systems face challenges in accurately determining the state of charge (SOC) of satellite batteries during non-contact durations, as real-time communication with ground stations is not possible, leading to inaccurate battery state estimation.
Innovation Solution
An apparatus and method utilizing an artificial neural network to predict battery state in non-contact durations by using satellite information, including attitude angle, position, eclipse state, current time, and days after launch, and combining this with measurement data from contact durations using Kalman filters and bi-directional long short-term memory (BI-LSTM) networks.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If real-time communication with ground station is used to monitor battery state, then measurement precision of battery state is improved, but availability of battery state information deteriorates during non-contact duration
Solution Approach 1:
The system performs preliminary actions by training artificial neural networks during contact duration when communication is available. The neural networks learn to predict battery states based on historical data and patterns, enabling accurate estimation during non-contact duration without real-time communication.
Solution Approach 2:
The system creates a virtual copy of the battery monitoring function by implementing prediction models (neural networks) that replicate the battery state estimation capability. This allows the satellite to autonomously determine battery states during non-contact duration as if real-time communication were available.
2Loss of information
If artificial neural network prediction is used for non-contact duration, then availability of battery state information is improved, but device complexity increases
Solution Approach 1:
The satellite system performs self-service by implementing autonomous battery state prediction capabilities onboard. The pre-trained neural networks enable the satellite to independently estimate its battery state during non-contact duration without requiring external ground station assistance, thus reducing information loss while avoiding continuous complex communication infrastructure.
3Measurement precision
If both prediction information and measurement information are used for SOC estimation, then SOC estimation accuracy is improved, but device complexity increases
Solution Approach 1:
The system merges two information sources - prediction information from artificial neural networks and measurement information from actual battery sensors - to estimate SOC. By combining these complementary data sources, the system achieves higher accuracy than either method alone, while the integration is managed through coordinated processing during contact and non-contact durations.
Data Source
AI summary
The present disclosure relates to an apparatus and method for estimating a state of charge of a satellite battery in a satellite system. According to the present disclosure, a method of operating an apparatus for estimating a state of charge includes obtaining satellite information indicating the status of the satellite according to a mission plan of the satellite, generating prediction information about a battery state of the satellite in a non-contact duration of the satellite from the satellite information based on a pre-learned artificial neural network, obtaining measurement information about a battery state of the satellite in a contact duration of the satellite; and estimating the state of charge of the battery based on the prediction information and the measurement information.