Search papers, labs, and topics across Lattice.
LoRM is a self-supervised framework that treats rotating machinery signals as a language, tokenizing local signals and predicting their future evolution using a pre-trained language model. This approach reformulates multi-modal sensor data as a token-based sequence-prediction problem, avoiding hand-crafted features. Experiments on tool condition monitoring demonstrate stable real-time tracking and strong cross-tool generalization, using token-prediction errors as a health indicator.
By reframing machine sensor data as a language, LoRM unlocks the power of pre-trained language models for real-time condition monitoring without relying on hand-engineered features.
We present LoRM (Language of Rotating Machinery), a self-supervised framework for multi-modal rotating-machinery signal understanding and real-time condition monitoring. LoRM is built on the idea that rotating-machinery signals can be viewed as a machine language: local signals can be tokenised into discrete symbolic units, and their future evolution can be predicted from observed multi-sensor context. Unlike conventional signal-processing methods that rely on hand-crafted transforms and features, LoRM reformulates multi-modal sensor data as a token-based sequence-prediction problem. For each data window, the observed context segment is retained in continuous form, while the future target segment of each sensing channel is quantised into a discrete token. Then, efficient knowledge transfer is achieved by partially fine-tuning a general-purpose pre-trained language model on industrial signals, avoiding the need to train a large model from scratch. Finally, condition monitoring is performed by tracking token-prediction errors as a health indicator, where increasing errors indicate degradation. In-situ tool condition monitoring (TCM) experiments demonstrate stable real-time tracking and strong cross-tool generalisation, showing that LoRM provides a practical bridge between language modelling and industrial signal analysis. The source code is publicly available at https://github.com/Q159753258/LormPHM.
