PSI - Issue 19

A. Halfpenny et al. / Procedia Structural Integrity 19 (2019) 150–167 Author name / Structural Integrity Procedia 00 (2019) 000–000

160

11

Care should be taken when comparing Weibull parameters based on different distributions. Measured data are modelled by a 2-parameters Weibull distribution, which is significantly different of a 3-parameters Weibull distribution since under the location parameter, the probability of phenomenon does not exist. This is discussed in the case study. The Weibull location parameter η is inconsistent between the 2 and 3-parameter Weibull plots. In this case a comparison between θ=η+γ is recommended. Similarly, the values of β are not directly comparable, and a comparison of the Reliability vs. Time plot and the Probability Density Functions are preferred. 4. Case study This case study considers the reliability of an air-cooled intercooler. It compares physical reliability test data with simulation data. Both deterministic and stochastic fatigue simulations are considered. The aim is to understand the impact of parameter uncertainty and use simulation as a means of extrapolating physical tests. 4.1. Reliability tests Reliability tests were performed on 5 intercoolers. The intercoolers were subjected to continual pressure cycles at elevated temperature as illustrated in Fig. 9. (The test levels are omitted for commercial sensitivity). Fatigue failures were observed in the header rail as highlighted. The number of pressure cycles to failure were recorded and are listed in Table 1.

Fig. 9. Accelerated intercooler reliability tests

Table 1. Observed pressure cycles to failure

351,630 445,000 465,000 615,000 649,800 505,286 1.848

Test

Life in cycles

1 minimum

2 3 4

5 maximum

Mean

Ratio (max:min)