What are the differences between ALTA and Weibull++?
Much of the general functionality between both software packages is very similar, but the underlying theory is different:
ALTA has been designed to analyze accelerated test data, such that each group of units is tested at a different stress level. You are then able to analyze all of the data together to extrapolate down to a specified use stress level. ALTA requires data at a minimum of two stress levels to conduct the analysis. ALTA is available in two versions: Standard and PRO. ALTA is an entry-level tool for basic accelerated life testing analysis. ALTA PRO is an advanced tool for analyzing data from tests with time-dependent stress profiles (e.g. step-stress) using a single stress. In addition, the PRO version has the ability to analyze data with up to eight simultaneous stress types. ALTA includes an accelerated degradation utility, reporting capabilities, as well as a large array of plots to help you conduct your analysis. Some case studies can be seen at http://www.reliasoft.com/alta/a6examples.htm.
Weibull++ has been designed for life data analysis and does not allow you to analyze data collected across multiple stress levels. Weibull++ is for data where all of the units under test are at the same test conditions. Weibull++ includes multiple distributions, the ability to do competing failure modes analysis, warranty analysis and many other capabilities to aid you in your analysis. Any calculated results are also at the same test conditions. You can check out some case studies at http://Weibull.ReliaSoft.com/w6examples.htm.
When conducting a Likelihood Ratio Test in ALTA 6, what does the significance level represent?
The Likelihood Ratio Test in ALTA tests the hypothesis that the shape parameter is not dependent on the stress level (i.e. constant across stress levels). The significance level represents the probability of rejecting this hypothesis. So, as the significance level decreases, the probability of accepting the hypothesis increases. This is not the same as confidence level.
A typical value for the significance level is 0.1, which indicates that there is a 10% probability you will reject the hypothesis that the shape parameter is constant across stress levels.
Using the Function Wizard in ALTA or Weibull++, how can you calculate functions such as R(t=MTBF)?
The obvious method would be to calculate the MTBF and use this value as the input for the reliability calculations. However, this calculation can be performed automatically using the Function Wizard's nested function capability. The following example uses the Function Wizard in Weibull++, but the process is the same for ALTA.
With a calculated data set in your Folio, open the Function Wizard and select Reliability given time from the function list. If the Composite Function Area at the bottom of the window is not visible, click the More/Less button to display it. Leave the Mission End T input set to the default (100) and click the Insert Selected Function icon.
This will insert the reliability function into the Composite Function Area, as shown in Figure 1.
Figure 1: Function Wizard with reliability function inserted into the Composite Function Area.
The next step is to select MTBF from the function list. Then, select the time value inside the reliability function (in this case, the 100 inside the brackets "") and click the Insert Selected Function icon again to replace the time value with the MTBF function. The Function Wizard will now look like the one displayed in Figure 2.
Figure 2: Function Wizard with reliability function and MTBF function inserted into the Composite Function Area.
As you can see, the reliability value is now a function of the MTBF. To calculate R(t=MTBF), click the Calculate icon.
In this example, the result is approximately 45.3%.
The functions inside the Composite Function Area can be customized to fit your needs. Standard math functions (+, -, *, /, ^, etc) can be used to return a wide variety of results. Figure 3 displays a modified version of the previously entered function. Given this new function, 1 - R(t = MTBF), the result is approximately equal to 54.7%.
Figure 3: Function Wizard with 1 - R(t = MTBF) defined in the Composite Function Area.
Copyright 2002 ReliaSoft Corporation, ALL RIGHTS RESERVED