FreqCondNorm: Towards Cross-domain Predictive Maintenance through a Frequency-Conditioned Transformer Foundation Model
Deep learning predictive maintenance models suffer from poor transferability across machines and operating conditions, especially when labelled data are scarce and signals span five orders of magnitude in sampling frequency (1 Hz to ~100 kHz). We propose FreqCondNorm, a Transformer-based architecture that introduces a FiLM-style frequency-conditioned normalization layer to unify heterogeneous time-series within a single model. The architecture is pretrained on five public predictive maintenance datasets (CWRU, MFPT, UOC18, PRONOSTIA, CMAPSS) using masked auto-encoding and contrastive learning with balanced domain sampling. On fault diagnosis, the model achieves 99.2% accuracy on CWRU (+6.4 pp over CNN) and 82.1% zero-shot accuracy on MFPT, demonstrating strong transfer across sampling freq
提出 FreqCondNorm,以解决跨领域预测性维护中的频率异构问题。
- 核心方法
- 引入基于 Transformer 的 FreqCondNorm 架构,通过 FiLM 风格的频率条件归一化层统一异构时间序列数据。预训练策略包括掩码自编码和对比学习,采用平衡领域采样。
- 适合谁读
- 研究者、工程师
- 要解决的问题
- 现有深度学习预测性维护模型在不同机器和操作条件下迁移性差,特别是在标签数据稀缺和信号采样频率范围大时。
- 关键实验
- 使用五个公共预测性维护数据集(CWRU, MFPT, UOC18, PRONOSTIA, CMAPSS)进行预训练,实验结果展示了模型在故障诊断任务上的性能提升。
- 主要贡献
- 在故障诊断任务上,模型在 CWRU 数据集上实现了 99.2% 的准确率(相比 CNN 提高了 6.4 个百分点),并在 MFPT 数据集上实现了 82.1% 的零样本准确率,展示了强大的跨频率迁移能力。
- 意义与局限
- 该模型解决了跨领域预测性维护中频率异构的问题,提高了模型的迁移性,特别是在标签数据稀缺的情况下。未来可应用于更多不同频率的工业设备维护。局限性在于模型的泛化能力仍需在更多实际场景中验证。