Automobile engine fault prediction method fusing knowledge graph and multivariate neural network model

• Main Authors: WANG XIAO, YIN YANCHAO, PU HAORAN
• Format: Patent
• Language: Chinese; English
• Published: 07.12.2021
• Subjects: CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; ELECTRIC DIGITAL DATA PROCESSING; MEASURING; PHYSICS; TESTING; TESTING STATIC OR DYNAMIC BALANCE OF MACHINES ORSTRUCTURES; TESTING STRUCTURES OR APPARATUS NOT OTHERWISE PROVIDED FOR

The invention relates to an automobile engine fault prediction method fusing a knowledge graph and a multivariate neural network model, and belongs to the field of automobile engine fault prediction. The method comprises the following steps: taking an automobile engine running state, a fault phenomenon, a fault reason and a maintenance record as input information, forming a representable and reasonable structured knowledge network through knowledge extraction, disambiguation and processing, and performing feature vector conversion; secondly, establishing a multivariate neural network path including a fault record embedding layer, a convolutional layer, a GRU gating layer and an attention mechanism, and forming an engine fault prediction model through feature vector training, and the engine qualitative fault phenomenon can be converted into quantitative fault reasoning and then into mapping transformation of qualitative fault prediction output. According to the invention, the prediction rate of engine faults i