基于随机森林和长短期记忆网络的高速剑杆织机故障诊断技术OA

High-speed rapier looms often encounter faults such as abnormal tension and high motor operation temperatures,hindering their efficient and stable operation.Benefiting from the progress of data-driven methods,the accuracy of equipment fault diagnosis is improved.However,due to the temporal nature of heterogeneous data from multiple sources,traditional fault diagnosis methods fail to provide reliable predictive performance.Therefore,a fault diagnosis method based on random forest(RF)and long short-term memory(LSTM)networks is proposed in this study,and then applied to the high-speed rapier loom.First,an STM32-based core controller is employed to capture data from various types of sensors,which are then transmitted via a serial port to a diagnostic module implemented on a Raspberry Pi.Notably,both the RF and LSTM models are deployed on the diagnostic module,where the RF model is used to determine the type of equipment failure and the LSTM model is used to estimate the remaining useful life.Practical results confirm that the RF model achieves a 96.4%fault diagnosis accuracy,while the LSTM model achieves a life prediction accuracy with R2=0.947,demonstrating excellent robustness and generalization ability.