涡轮发动机高速旋转状态失衡振动故障诊断系统

Fault Diagnosis System for Unbalanced Vibration of Turbo Engine during High Speed Rotation

摘要:

涡轮发动机是一种通过燃烧燃料并利用高速流动气体产生动力的发动机,在高速旋转状态下,其失衡振动会导致涡轮发动机各部件的磨损加剧,减少使用寿命。为了有效解决这一问题,设计一种涡轮发动机高速旋转状态失衡振动故障诊断系统。该设计将系统框架划分为三层结构,包括下位机层、上位机层以及客户端层。该系统采用振动传感器和声音传感器采集高速旋转状态下的涡轮发动机工作过程中的状态信号,并通过变送器将信号转变为可被下位机识别的信号。下位机运行处理程序,对信号实施滤波,提取故障特征,然后通过Zigbee远程通信模块将信号转发给上位机。上位机层运行知识图谱诊断程序,构建发动机失衡振动故障的知识图谱,并结合贝叶斯网络推断故障类型,计算故障发生概率,实现高精度的失衡振动故障诊断。实验结果表明:与三种传统诊断方法相比,所设计系统的ROC曲线在最上方,曲线下方的面积更大,AUC值=0.847,说明所设计系统的故障诊断能力强,能保证诊断结果的准确性。

Abstract:

Turbine engine is a kind of engine that burns fuel and uses high-speed flowing gas to generate power. In the state of high-speed rotation, its unbalanced vibration will lead to increased wear of turbine engine components and reduce service life. In order to effectively solve this problem, a fault diagnosis system for unbalance vibration of turbine engine in high-speed rotation state is designed. This design divides the system framework into three layers, including the lower computer layer, upper computer layer, and client layer. The system uses vibration sensors and sound sensors to collect status signals during the operation of the turbine engine under high-speed rotation, and converts the signals into signals that can be recognized by the lower computer through a transmitter. The lower computer runs the processing program, filters the signal, extracts fault features, and then forwards the signal to the upper computer through the Zigbee remote communication module. The upper computer layer runs a knowledge graph diagnostic program to construct a knowledge graph of engine imbalanced vibration faults, and combines it with Bayesian networks to infer fault types, calculate the probability of fault occurrence, and achieve high-precision imbalanced vibration fault diagnosis. The experimental results show that compared with the three traditional diagnostic methods, the ROC curve of the designed system is at the top, and the area below the curve is larger. The AUC value is 0.847, indicating that the designed system has strong fault diagnosis ability and can ensure the accuracy of diagnostic results.