|
|
|
Foundation ::
Reliability Analysis ::
HARP
|
HARP
Hybrid Automated Reliability Predictor, Version 7.0
|
 |
Moderators:
Adopt This Application!
|
SOURCE CODE AVAILABLE
|
|
HARP is the modeling engine at the core of Langley Research Center's HiRel tool system for reliability/availability prediction. This program can accept textual input interactively or in the form of user-created input files, and it produces tabular structured ASCII files as output. Other components of the HiRel suite, which are available separately from COSMIC, provide interactive graphical input/output interfaces for the HARP program. The Graphics Oriented input program, GO, allows users to "draw" reliability models of complex systems using fault tree notation or Markov graphs. Based on these drawings, GO creates files which HARP can accept as input files. The HARP output program, HARPO, allows users to graphically display the tabular HARP output data in a variety of forms. Information about GO and HARPO will be available shortly on the Open Channel Foundatiojn site.
HARP provides a general Markov modeling capability to conveniently model and predict the reliability/availability of a wide variety of systems. Its primary input is in the form of a fault tree that can be entered in textual form. The fault tree is not limited to the traditional combinatorial modeling approach. The addition of four special fault tree gates, called sequence dependency gates, allows the generation of dynamic fault tree models that were not practically solvable using analytic solution techniques until now. The fault tree, a familiar and convenient notation for expressing reliability models, is automatically converted by HARP into a Markov model which is solved to produce a reliability prediction. HARP will also accept system reliability/availability models expressed directly in the form of a Markov model, where the user specifies the model's origin and destination state together with a state transition rate.
For fault-tolerant systems that use redundancy and subsystem reconfiguration to achieve ultra-high predicted reliabilities, even HARP's automatic model generation capabilities may be inadequate to cope with the potentially millions of Markov states that would be necessary to model the system reliability. HARP offers several innovative model reduction techniques to make it possible to computationally solve large models. These include:
- model truncation with bounds,
- state aggregation with fault tree model generation,
- usage of a simple multi-fault model,
- and behavioral decomposition.
The simple multi-fault model is applicable primarily to ultrareliable fault tolerant system models. Significant state reduction can also be achieved by accurately modeling pairs of near-coincident faults and approximating the occurrence of more than two near-coincident faults.
Of all the techniques, behavioral decomposition offers the greatest state reduction when fault/error handling is required. Behavioral decomposition is a mathematical technique that exploits two specific behaviors of fault-tolerant systems:
- the failure of hardware parts and subsystems typically occurs after thousands of hours of operation; and
- the time involved in the handling of faults/errors is usually on the order of milliseconds to seconds.
The wide disparity, typically six orders of magnitude, of the system time constants makes HARP results accurate enough for practical applications. The mathematical basis of behavioral decomposition guarantees that the behavioral decomposition model will be conservative with respect to the model given to HARP.
HARP can handle dynamic fault trees with repeated nodes (i.e., shared basic events), repairable systems (to determine instantaneous availability) specified using a Markovian fault-occurrence/repair model, systems with sequence dependent failures as dynamic fault trees of Markov chains, and systems with cold and warm spares. HARP can utilize the Weibull failure distribution including hot spare repairable systems. It can also provide automatically-generated, guaranteed parametric bounds on system reliability for a large number of applications of practical interest and for all Markov models.
HARP carries the NASA case numbers LAR-15188 (SUN version) and LAR-15194 (IBM PC version). It was originally released as part of the COSMIC collection.
|
|
More software from National Technology Transfer Center
|
|
|
|
|
|
|
Open Channel Software runs entirely on Open Source Software. We return
value to the Software community in the form of
services and original software. Most of our
content is currently available as source code, with the
copyright owned by the original author, All
Rights Reserved. Everything else is Copyright
©2000, 2001, 2002, 2003, 2004, 2005, 2006, 2007, 2008 Open Channel Software.
View our privacy statement.
Contact webmaster at openchannelsoftware dot org with questions.
|
|
|
