A deep learning approach for predicting the antenna pointing error caused by transmission faults with simulation data

Area of Science:

• * Aerospace Engineering
• * Mechanical Engineering
• * Signal Processing

Background:

• * Reflector antennas require high pointing accuracy for applications like deep space exploration and radar.
• * Antenna elevation bearings are critical for maintaining pointing accuracy; faults can cause significant performance degradation.
• * A clear understanding of the link between bearing faults and pointing accuracy is lacking due to insufficient experimental data.

Purpose of the Study:

• * To establish a deep learning model to elucidate the relationship between antenna transmission faults and pointing accuracy.
• * To enable real-time fault diagnosis using vibration signals as a proxy for pointing accuracy, facilitating antenna maintenance.
• * To provide a theoretical foundation for developing advanced antenna maintenance strategies.

Main Methods:

• * Developed a dynamic simulation model of an antenna elevation system with pre-defined transmission faults to generate data.
• * Established a mathematical model for antenna axis error analysis to correlate faults with pointing errors.
• * Trained a deep neural network using labeled fault data and pointing errors to create a predictive model.

Main Results:

• * Key transmission component faults were shown to have a significant impact on antenna pointing errors.
• * The developed deep neural network model demonstrated high accuracy in predicting antenna axis errors.
• * Vibration signal analysis can effectively serve as a basis for real-time fault diagnosis.

Conclusions:

• * The study successfully linked antenna transmission faults to pointing accuracy using a deep learning approach.
• * The predictive model offers a viable method for proactive maintenance decisions in antenna systems.
• * Real-time monitoring of vibration signals can enhance the efficiency and reliability of antenna maintenance strategies.