Marine bearing residual life prediction method based on transfer learning and multiple time windows

• Main Authors: ZHU KUN, FAN KESEN, CHEN ZHAOXU, ZHOU HONGKUAN, XIAO QI, WAN YIMING, CHEN KAI, KE ZHIWU, LI BANGMING, KE HANBING, GOU JINLAN, WEI ZHIGUO
• Format: Patent
• Language: Chinese; English
• Published: 15.03.2022
• Subjects: CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; ELECTRIC DIGITAL DATA PROCESSING; HANDLING RECORD CARRIERS; PHYSICS; PRESENTATION OF DATA; RECOGNITION OF DATA; RECORD CARRIERS

The invention discloses a marine bearing residual life prediction method based on transfer learning and multiple time windows, and belongs to the technical field of mechanical part loss detection technologies, and the method comprises the steps: training a CNN aging model and a multi-time window prediction model based on an LSTM neural network; inputting a vibration signal of a to-be-predicted bearing into the CNN aging model to judge whether the bearing has an early fault or not; if the early fault occurs, inputting CNN depth features of a plurality of preset length windows corresponding to the vibration signal into a multi-time-window prediction model to obtain life prediction values corresponding to the plurality of preset length windows; fusing the plurality of life prediction values to obtain a target prediction life of the to-be-predicted bearing; according to the method, the problem that a single window is difficult to adapt to multiple degradation modes is solved by adopting a method of fusing multipl