Abstract: Aiming at the problems of fragmented data analysis, low efficiency of anomaly diagnosis and poor low-temperature adaptability in power battery sealing inspection, this paper takes leakage rate as the core quality characteristic and constructs an integrated statistical analysis framework including data distribution analysis, comprehensive process capability evaluation, and rule-intelligence dual anomaly diagnosis. A dynamic quality monitoring system is established synchronously. Based on verification from 600 sets of measured data, the optimized short-term process capability index (adjusted) Cpk increases from 1.05 to 1.62, the reject rate is reduced from 0.8% to 0.3%, the low-temperature qualification rate is raised from 89.2% to 99.1%, and the anomaly tracing time is shortened from 45 min to 12 min. The research realizes accurate data analysis, process stability improvement and early defect warning, providing a data-driven scheme for power battery incoming quality control and a reference for similar sealing inspection.

Key words: battery sealing; statistical analysis; process capability; anomaly diagnosis; quality control

摘要： 针对动力电池密封性检测存在数据分析碎片化、异常诊断效率低、低温适配性差等问 题，文章以泄漏率为核心质量特征，构建“数据分布分析-全维度过程能力评估-规则+智能双 重异常诊断”一体化统计分析框架，并建立动态质量监控体系。通过 600 台套实测数据验证， 优化后的短期过程能力指数（修正）Cpk 由 1.05 提升至 1.62，不合格品率由 0.8%降至 0.3%， 低温（T<0 ℃）检测合格率由 89.2%提升至 99.1%，异常追溯时间由 45 min 缩短至 12 min。 研究实现了检测数据精准解析、过程稳定性提升与缺陷早期预警，为动力电池入库质量管控 提供数据驱动方案，可为同类密封检测质量控制提供参考。

关键词: 电池密封性；统计分析；过程能力；异常诊断；质量控制