Worked examples

Three calls from start to finish. Pick a language; every example switches with it.

Fit a distribution

Fit a Weibull to five failure times, one still running (censored 1). Fitting is Pro.

rf.call("life_data_analysis", "fit", distribution="weibull",
        data=[120, 340, 510, 700, 980], censored=[0, 0, 1, 0, 0], unit="hours")

Analyse an RBD

System reliability, MTTF and importance for a saved diagram (list rbds gives the ids). With simulate, a long run answers with a job to poll.

rf.call("rbd_analysis", "analyze", rbd_id="RBD_ID", times=[1000, 8760])
rf.call("rbd_analysis", "analyze", rbd_id="RBD_ID", simulate=True, wait=True)  # waits for the job

Plan a test

How many units to test, and for how long, to show 90% reliability over 1,000 hours with 90% confidence.

rf.call("life_data_analysis", "plan_demonstration_test",
        reliability=0.9, confidence=0.9, mission_time=1000, unit="hours")

The Python lines use rf = Client() from reliafy-client (see the Python client). The same calls from the command line: reliafy life_data_analysis fit --distribution weibull …. A long job, such as a simulation, is polled as in long-running calculations.