概率神经网络在车辆齿轮箱典型故障诊断中的应用

为提高汽车齿轮箱典型故障的诊断效率和准确性,提出一种基于概率神经网络的齿轮箱故障诊断方法。通过对某型齿轮箱的实验采集齿轮箱在正常状态、齿根裂纹和断齿状态下的振动信号,经过数据处理得到样本数据后输入概率神经网络模型,通过交叉验证并与BP神经网络对比的结果表明:概率神经网络能准确地识别出齿轮箱典型故障,且与BP神经网络相比,诊断准确率更高、诊断速度更快。     

Modelling upgrading maintenance policy in the one-dimensional renewing warranty period

In the one-dimensional renewing warranty period, the quality of the spares for product is likely to be improved during the warranty period. Therefore, upgrading maintenance becomes more and more common. Then the manufacturers (customers) may have to decide whether or not to provide (buy) the warranty considering upgrading maintenance. This paper presents a mathematical model considering upgrading maintenance for products with multiple failure modes. Upgrading maintenance is taken into account with the assumption that the warranted item is upgraded one time during the warranty cycle. The upgrading maintenance is carried out, when the corrective maintenance is taken place. After upgrading maintenance, the high-quality spares are used to replace the failed item. In the numerical example, the results of the models are calculated. Monte Carlo simulation results are compared with the analytical results to demonstrate the correctness and efficiency of the proposed models considering upgrading maintenance.     

强噪声背景下的柴油机失火故障诊断

柴油机失火是常见的故障模式,传统的诊断方法不仅参数获取困难且准确性差。针对此问题,以3缸四冲程柴油机为研究对象,设计了柴油机失火故障的预置试验,采集排气噪声和缸盖振动信号进行故障诊断研究。为提取强噪声背景下的微弱信号,采用二次采样随机共振系统提取柴油机故障特征频率完成柴油机的失火故障诊断。研究结果表明,通过二次采样处理,随机共振系统可以将噪声能量转移到柴油机微弱特征信号上,达到大参数条件下微弱信号特征提取的目的,能有效识别柴油机的早期故障,对其他复杂机械的振动诊断同样具有参考价值。     

Study on simulation modeling and evaluation of equipment maintenance

Aiming at the actual demand for the maintenance support of new general equipments, this paper analyzes the structure and circulation of maintenance support system, builds the simulation model of running process by entity flow chart method, and constructs the basic frame of the simulation model. The generating method of random variables and the advancing process of the simulation clock are described, and the accurate prediction of maintenance equipment is realized. Moreover, the material readiness rate is analyzed in statistics objective to evaluate the simulation result. The research is of significance for improving the efficiency of equipment maintenance and the promotion of material readiness rate.     

Rigid-flex coupling modeling and simulation on multistage gear transmission system

As multistage gear transmission systems are complex and precise, the flexibility of shaft can influence the dynamic response of system. In order to study dynamic response of the system, we build the rigid model of gear system and the finite element model of the gear shaft. The rigid-flex coupling model is established with the virtual prototype technology, and a contrast between rigid model and rigid-flex coupling model is constructed.With these methods, the dynamic responses with different rotation speeds and different loading magnitudes are examined. We also analyze the influence of shaft flexibility, rotation speeds and loading magnitudes on the vibration characteristics of gear transmission systems.