Example 2 - Working with Grouped per Configuration Discrete (Success/Failure) Data
[Download RGA7 Example File (*.rga7)]
Discrete data are also referred to as success/failure data or attribute data. These data sets consist of data from a test where a unit is tested with only two possible outcomes: success or failure.
The Grouped per Configuration data type is used when multiple items, instead of a single item, are tested and the number of units that fail are recorded for each configuration. Each stage represents a unique configuration. Improvements are made to the system after all of the units within a given stage have been tested.
Background
A one-shot system under development underwent reliability growth testing for a total of 68 trials. Corrective actions were incorporated after the 14th, 33rd and 48th trials. The configuration was not changed from trial 49 to trial 68.
- Configuration 1 (trials 1 to 14) experienced 5 failures. (The cumulative number of failures at that point was 5 failures in 14 launches.)
- Configuration 2 (trials 15 to 33) experienced 3 failures. (The cumulative number of failures at that point was 8 failures in 33 launches.)
- Configuration 3 (trials 34 to 48) experienced 4 failures. (The cumulative number of failures at that point was 12 failures in 48 launches.)
- Configuration 4 (trials 49 to 68) experienced 4 failures. (The cumulative number of failures at that point was 16 failures in 68 launches.)
The reliability of the final configuration is calculated in this example using the Crow-AMSAA (NHPP) model. The results showing the reliability achieved by Configuration 4 are displayed using the Reliability vs. Time Plot and the Quick Calculation Pad (QCP).
Analysis and Results
A Standard Folio Data Sheet using the Grouped per Configuration data type is created by selecting the Developmental and Grouped per Configuration options in the New Data Sheet Setup window Expert view, as shown next.
On the Main page of the Standard Folio's Control Panel, the Crow-AMSAA (NHPP) model is selected. The number in each row in the Data Sheet represents a different configuration. The cumulative number of launches for each configuration is entered in the Number of Units column (which has been renamed to Trials (Cumulative) for this example) and the cumulative number of failures for each configuration is entered in the Number of Failures column, as shown next.
On the Analysis tab of the Standard Folio's Control Panel, notice that Maximum Likelihood is selected for the analysis type. This is the only option available for this data type and model. Under Other Options, Input is Cumulative is selected, as shown next.
The data set is analyzed, as shown next.
A plot of the Reliability vs. Time is shown next.
The plot displays a line for each of the four configurations, identified as "Data Intervals" in the plot. You can see from the plot that the reliability for Configuration 4 is about 80%. However, an exact calculation can be obtained from the Quick Calculation Pad, as shown next.
You can see that the reliability of Configuration 4 is .8096 or 80.96%.1
1. You can change the display of calculated values from decimal to percentages by choosing File > User Setup and clicking the Calculations page. In the Other Options area, click the Use Percent Values (0-100%) check box, then click OK.
