基于人工智能的丝杠寿命预测技术

Intelligent Monitoring System for Screw Life Evaluation

为了研究丝杠在不同加工条件下的性能退化趋势,研究了影响丝杠寿命的关键因素:丝杠转速和负载力.利用振动信号和切削力信号实时监测丝杠性能状态,用经验模态分解方法对传感器信号滤波后,通过时域、频域和时频域分析方法提取影响丝杠寿命的关键特征,采用多模型融合技术和B样条模糊神经网络,建立了丝杠寿命预测模型.试验结果表明,寿命预测的最大误差为846 h,能够满足丝杠的主动维护需求.

The effects of rotating speed and load on screw life were investigated to study the performance degradation of screws of NC（numerical control） machine tools under different machining conditions.The real time data of vibration and cutting forces were collected.The key effects on the screw life were identified by analyses on time and frequency domains and wavelet analysis following filtering of the collected data with an empirical mode decomposition.A screw life prediction model was proposed using a multi-model fusion and a B-spline fuzzy neural network.Experimental results show that the maximal error of life prediction is 846 h,and the proposed system meet the need of active maintenance of screw.