YANG Xin.Research on Lightweight AI Model for Health Prediction of ATC UPS Batteries[J].Journal of Chengdu University of Information Technology,2025,40(06):753-760.[doi:10.16836/j.cnki.jcuit.2025.06.002]
空管UPS蓄电池健康预测的轻量化AI模型研究
- Title:
- Research on Lightweight AI Model for Health Prediction of ATC UPS Batteries
- 文章编号:
- 2096-1618(2025)06-0753-08
- Keywords:
- UPS battery; XGBoost; multi-dimensional feature engineering; state transition matrix; deterioration trend prediction
- 分类号:
- TP307
- 文献标志码:
- A
- 摘要:
- 针对传统人工运维在空管UPS蓄电池健康监测中存在的标准判断模糊、分析深度不足、人工预测低效等痛点,提出融合多维度特征工程与轻量化XGBoost的预测模型。基于某空管局160只蓄电池连续4年历史数据,构建数据治理、模型构建和本地部署的技术体系。通过民航通信导航监视规范以及通信行业标准实现健康标签精准定义,结合轻量化模型压缩技术解决实时监测与劣化趋势预测难题。实验表明,该模型在数据不均衡场景下健康预测准确率达99.2%,为高可靠性空管工艺配电系统的智能化运维提供了可工程化的技术路径。
- Abstract:
- To address the pain points of vague judgment criteria, insufficient analysis depth, and low manual prediction efficiency in traditional manual maintenance for ATC UPS battery health monitoring, this paper proposes a prediction model integrating multi-dimensional feature engineering with lightweight XGBoost. Based on four years of historical data from 160 batteries of an air traffic management bureau, a technical system encompassing data governance, model construction, and local deployment is built. Health labels are precisely defined per Civil Aviation Communication, Navigation, and Surveillance Specifications and telecom industry standards. With lightweight model compression technology, the model solves real-time monitoring and deterioration trend prediction challenges. Experiments show the model achieves a health prediction accuracy of 99.2% in data imbalance scenarios, providing an engineerable technical route for intelligent maintenance of high-reliability ATC process power distribution systems.
参考文献/References:
[1] 唐娴静,杜金红,林伊婷,等.对可靠数据中心UPS使用的VRLA电池的预测性维护[J].数据中心建设,2020(6):4.
[2] 开放数据中心委员会(ODCC).数据中心蓄电池智能诊断与运维技术白皮书[R].ODCC-2023-06005,2023.
[3] AP-115-TM-2016-01,民用航空通信导航监视运行保障与维护维修规程(4418蓄电池)[S].
[4] YDT7992010,中华人民共和国通信行业标准通信用阀控式密封铅酸蓄电池[S].
[5] DL/T 724-2021,电力系统蓄电池直流电源装置运行与维护技术规程[S].
[6] 珠海九源绿电科技有限公司.一种用于电池健康度评估的大数据计算方法及系统[P].中国:CN116381537A,2023.
[7] 邓明斌.基于多特征融合的用户窃电行为快速识别研究[J].自动化应用,2017,45(12):2399-2401.
[8] 魏晓斌.阀控式铅酸蓄电池提前失效原因分析及对策[J].通信电源技术,2009,26(2):54-56.
[9] 梅大成.边界与密度适应的SMOTE算法研究[J].计算机应用研究,2022,5:1478-1482.
[10] Sun J,Fan C,Yan H.SOH estimation of lithium-ion batteries based on multi-feature deep fusion and XGBoost[J].Energy,2024,306.
[11] 胡晨.基于深度学习的铅酸电池健康状态估计[J].电池,2021,51(1):63-67.
[12] 杨家鑫.基于多源信息的隐含非线性维纳退化过程剩余寿命预测[J].北京航空航天大学学报,2024(1):328-340.
[13] 张开放.一种基于混淆矩阵的多分类任务准确率评估新方法[J].计算机工程与科学,2021,11:1910-1919.
相似文献/References:
[1]毛开银,赵长名,何 嘉.基于XGBoost的10 m风速订正研究[J].成都信息工程大学学报,2020,35(06):604.[doi:10.16836/j.cnki.jcuit.2020.06.004]
MAO Kaiyin,ZHAO Changming,HE Jia.A Research for 10 m Wind Speed Prediction based on XGBoost[J].Journal of Chengdu University of Information Technology,2020,35(06):604.[doi:10.16836/j.cnki.jcuit.2020.06.004]
备注/Memo
收稿日期:2025-06-30
通信作者:杨鑫.E-mail:Yancey1997@163.com