Prediction Of Cylinder Fatigue Damage Under The Constant Load

Pneumatic machinery have extensive applications in engineering practices and people’s lives, it plays an important role in national economy and social development. As an important component in pneumatic system, cylinder is mainly responsible for the transmission and the output of power. People use air as the transmission medium in the pneumatic systems, however, the cylinder operational speed is instable for the gas’ compressibility. This makes a larger impact in cylinder working process, resulting in the occurrence of faults. If we don’t find these problems timely and take some proper measures, it will result in equipment failure downtime and even cause the safety accidents. In order to timely replace the cylinder, learning more about cylinder fatigue failure process, making the cylinder failure analysis and prediction are particularly important.Among the existing assessment methods of mechanical product failure, the method based on the test data has been widely concerned, because it has a real data source, convenient access and reliable evaluation result etc. In this paper, writer do a series of exploration and research about the cylinder fatigue failure assessment based on the acquisition of the performance indicators and various failure data in the process of the cylinder fatigue test. Concrete research content and main innovation as described below.(1) Studying fatigue failure process and failure analysis method comprehensively, especially the fault tree analysis method. In addition, this paper puts forward to specific steps for cylinder fatigue failure analysis with the understanding of a cylinder’s structure characteristics.(2) Planning cylinder fatigue testing scheme and establishing cylinder fatigue test platform. In order to ensure the reliability of the data source, it is essential to design a rounded test data acquisition method.(3) Finding out the cylinder life distribution from the failure data and using the limited sample data to predict the life of the cylinder. From perspective of statistics, this paper puts forward using nonparametric estimation method, namely the Kernel Density Estimation(KDE). This method can count on the life of the sample distribution, without controlling by parameter estimation method of the model, which can achieve effective use of sample data.(4) Using of performance indicators monitoring data for performance evaluation in the process of cylinder operation. This paper proposes a performance evaluation method based on the Principal Component Analysis(PCA) and the Wavelet Neural Network(WNN). Benefitting from the statistical tool KDE, the proposed performance evaluation method can utilize the statistical information of the samples within the same class, but not a set of independent samples, so that this method can acquire more complete information from the original samples. Then those extracted main information can be used for neural network training and forecasting to get the instability life of the cylinder.