Models

In ReliaSoft applications, models can represent probabilities, durations or costs, either fixed or time-dependent. They are used by other resources to represent the reliability of a component, the duration of a task, the expected cost of a repair and many other characteristics.

Like any other resource, you can create or edit models while you’re performing a relevant analysis, and from the Resource Manager. In addition, you can publish models that are based on an existing analysis (e.g., a life data analysis in Weibull++, a diagram in BlockSim, etc.).

This topic describes the properties for both types of models. The interface will vary based on the particular situation.

Naming the Resource

For a new resource, a name will be proposed automatically based on the default naming criteria established for the current database (see Default Name Formats window). You can replace this with your own name, if desired. Remember that the name and identifiers are the primary way in which your team will be able to find the resources you need for your analyses.

Associated Analysis

In published models only, the Associated Analysis area identifies the analysis that the model is based on. If the published model reflects the latest analysis results, the status is "Synchronized." If the analysis has been modified since the model was last published, the status is "Out of Sync."

Tip: If you have the ability to open the original analysis (i.e., if the required ReliaSoft application is activated on your computer and your user account has permissions to access the analysis), the Source will be configured as a link and there will also be a Data Source button at the bottom of the window.

Model Category

The category determines where and how a model can be used. If you are using a resource wizard to create a new model, it will be assigned automatically based on what’s relevant for the current field. If you are publishing a model or creating it from the Resource Manager, you must select the appropriate option. Once a model has been created, you cannot change its category.

Reliability, Probability of Failure and Event Occurrence models represent a likelihood of occurrence.
Duration models represent a length of time.
Cost per Unit Time models are used for costs that accrue over time (e.g., the crew charges $50 per hour).
Cost models are used for costs that don't depend on time (e.g., the part costs $100 or the crew charges $50 per service call in addition to their hourly rate).

Model Type

A distribution model represents behavior that varies based on factors such as time and/or applied stress. For a published model, the inputs and parameters will be specified automatically based on the associated analysis. For manually created models, you can:
- Select a distribution from the drop-down list and then enter the required parameter(s).

What's Changed? Starting in Version 2020, two new distributions are available for use in models: The triangular distribution takes three parameters (lower limit, mode, upper limit) defining the shape of the pdf. The uniform distribution takes two parameters (lower limit, upper limit) defining the bounds of the pdf. For both distributions:

    - Lower limit >= 0
    - Upper limit >= 0
    - Upper limit >= lower limit

For the triangular distribution:

    - Mode >= 0
    - Mode >= lower limit
    - Mode <= upper limit

- Use the Quick Parameter Estimator (QPE) to estimate the parameters of a distribution based on what you know about the behavior.

Tip: A published model includes one additional parameter that is not relevant for manually created models. PNZ stands for percent non-zero. A value of 1 indicates that there are no zero failure times in the data set (which is the most common scenario in life data analysis). A decimal value indicates that the data set does include zero failure times (such as out-of-the-box failures, for example). In such cases, the parameters are calculated based on the non-zero failure times, and then the PNZ value is used as a multiplier when calculating certain metrics (e.g., reliability, unreliability).

A constant model represents a fixed probability (e.g., 0.9), duration (e.g., 2 hours) or cost (e.g., $10 or $10 per hour).
A dynamic model represents a fixed probability, duration or cost, based on a specified variable. The variable can then be programmatically varied during simulation in one of two ways, thereby changing the value of the model for successive simulation runs.
- For event analysis flowcharts, the variable can be varied using the sensitivity analysis/multiple analyses features.
- When simulating a simulation worksheet, the values used by the variable will be specified in the worksheet.

When dynamic models are used outside of these circumstances, they are treated as constant models using the defined initial value of the variable.

A profile model represents a changing duration or cost over time, based on a specified profile. Profile models are available only for the cost and duration model categories. The cost or duration of the model will be programatically varied based on an input time parameter used in the profile. When creating a profile model, you must specify a time unit; this will be used for the profile’s segment times and, if applicable, for the profile’s output duration.

Profile models can be used for all direct costs and acquisition times for spare part pools. The execution time of the task using the spare part pool will be used as the input time parameter for the profile. Profile models are not available for selection anywhere other than in the relevant fields within spare part pools.

Note that models that are in use cannot be changed to profile models.

Additional Tools

Additionally, the following are shown:

Identifiers contains additional identifying information that can be used to search for this resource.

History provides information about when the record was created and last updated. If the history log has been activated at the project level, you can click the View Item History icon to open the Record History Log for the record.