GRIF | Reseda - Assessing the Reliability/Availability of networks

​​​​​​With Reseda’s intuitive graphical interface, creating a network is easy..

• Once the nodes and links have been created, users define their failure distribution, choosing from over 20 probability distributions. Links can be directional or bi-directional according to the system’s specifications, and as with all the modules in the Boolean package, Common Cause Failures can be taken into account. 
• With the “Attribute” feature (system with custom properties), you can add any information you wish to each object of the document, either to provide a more precise description or for traceability purposes.
• “AND” and “M out of N” logic nodes are also available for modeling specific systems in which the output of one node can be obtained provided there are sufficient inputs.

Reseda uses the ALBIZIA calculation engine to produce Binary Decision Diagrams (BDD) which give accurate results. This engine provides all the available results for the other Boolean modules: Unavailability (PFD), Availability, Reliability, Unreliability, Frequency (PFH) and equivalent failure rate. The importance factors are used to identify the network’s weak points and select the nodes/links that need to be improved.

The computed minimal cut sets identify the critical paths so that a redundancy path can be added where necessary.

Factoring in uncertainties:

• To simulate real-life conditions as closely as possible, parameter-related uncertainties can be factored in. For example, you can specify whether a failure rate follows a Uniform, Normal or Log-Normal distribution.
• In addition to the BDD calculation, a Monte-Carlo simulation is performed to obtain mean values.
• Finally, a quantile calculation is performed to provide a confidence interval for each result. The interval can be 60, 70, 80, 85, 90, 95 or 99% (centered or not).

N.B. in accordance with IEC standard 61511.

User-friendly design:

• An image of the system can easily be added to the background to facilitate the creation of a network.
• In addition to the editing window for each object, GRIF produces data tables that make it easier to control the quality of input data and modify information (find/replace, suffixes, prefixes).
• GRIF features plug-ins for entering input data, either based on internal feedback from the field, standards or commercial sources. Having accurate input data saves time for users.

• Option to automate calculations (batch runs) and draw variations for sensitivity analysis.
• Results are stored in the document and can be exported in a variety of formats (csv, XML, Excel, etc.).
• Results can be viewed as line graphs, pie charts or histograms.
• Printing in PDF vector format produces high-quality images and the files are small enough to be sent by email, even if the document contains hundreds of pages.
• External files (PDF certificates, system images, etc.) can be added to the document and incorporated into the full report.
• Interaction with the operating system: option to copy/paste either to or from word processing software, spreadsheets, or presentation tools.

The Reseda module is part of the Boolean package of the GRIF software suite. Any designed reliability network can be integrated into the Bool module, which combines the characteristics of all the modules in this package. Calculations can be performed in the same document using event trees (ETree), fault trees (Tree), reliability block diagrams (BFiab), safety instrumented systems (SIL), reliability networks (Reseda), or Bow-Tie / LOPA tables (Risk).

Models can be linked together, and a link can define any object (Event, Block, Barrier, Network node, etc.) using any model: Fault Tree, Block Diagram, SIS, Event Tree; etc.

Users in different departments of a company can link their respective models and work together on a shared network. For example, an engineer can create a network comprising 30 nodes, with each node representing a piece of equipment whose failure can be defined using a specific fault tree. The fault trees can also share gates and basic events so that dependencies between nodes can be taken into account.