LoRM: Learning the Language of Rotating Machinery for Self-Supervised Condition Monitoring

2026-04-07

AI Summary

This study addresses the challenge of self-supervised condition monitoring for rotating machinery using multimodal signals by proposing a novel paradigm that eliminates the need for manual feature engineering. The approach treats mechanical signals as a form of “machine language,” discretizing future signal segments into symbolic tokens and leveraging pretrained language models to predict contextual sequences. Real-time health monitoring is achieved through prediction error analysis. The key innovation lies in the first-time adaptation of language modeling paradigms to industrial signal analysis, integrating signal tokenization, multimodal sequence modeling, and lightweight fine-tuning strategies to significantly enhance cross-device generalization. Experimental validation on tool condition monitoring demonstrates stable real-time tracking performance and strong generalization capability, confirming the method’s effectiveness and practical utility.

Abstract

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.

Problem

• rotating machinery
• self-supervised learning
• condition monitoring
• multi-modal signals
• token prediction

Innovation

• self-supervised learning
• tokenization
• pre-trained language models
• condition monitoring
• rotating machinery