| 摘要: |
| 输出功率是衡量固态雷达发射机功放组件是否故障的重要指标，对其预测可为维护中预先更换功放组件提供决策依据，对雷达的预测性维护具有重要意义。本文利用ARIMA模型和LSTM模型对某型雷达功放组件输出功率预测进行了研究，根据实际监测到的输出功率劣化趋势将功放组件的工作阶段分为平稳工作期、缓慢劣化工作期和快速劣化工作期，并分别进行了输出功率预测实验，给出了不同预测时长下的RMSE、MAPE。研究结果表明，平稳工作期的预测效果好，长期预测和短期预测效果差别不大；当预测时长较短且监测值变化平缓时缓慢劣化工作期预测效果好，否则难以预测；快速劣化工作期难以进行满足预测性维护要求的输出功率预测。 |
| 关键词: 雷达 功放组件 功率预测 故障预测 健康管理 |
| DOI：DOI:10.3969/j.issn.1672-2337.2023.05.004 |
| 分类号:TN95 |
| 基金项目: |
|
| Output Power Prediction of Radar Power Amplifier Module |
|
FEI Taiyong， SUN Zhijian， QU Zhiguo， LI Jingming
|
|
Air Force Early Warning Academy， Wuhan 430019， China
|
| Abstract: |
| The output power is an important indicator to measure whether the power amplifier module of the solid?state radar transmitter is faulty. Its prediction can provide a decision basis for replacing the power amplifier module in advance during maintenance, and is of great significance to the predictive maintenance of the radar. In this paper, the ARIMA model and LSTM model are used to study the output power prediction of a certain radar power amplifier module. According to the actual monitored output power degradation trend, the work stages of the power amplifier module are divided into stable work period, slow deterioration work period and fast deterioration work period. The output power prediction experiments are carried out separately, and the RMSE and MAPE under different prediction time are given. The results show that the prediction effect is good in the stable work period, and the difference between the long?term prediction and short?term prediction is not significant. When the prediction duration is short and the monitoring value changes slowly, the prediction effect of the slow deterioration work period is good, otherwise it is difficult to predict. It is difficult to predict the output power that meets the requirements of predictive maintenance during the rapid deterioration period. |
| Key words: radar power amplifier module power prediction fault prediction health management |
