Inference
Inference database items contain rules that are used during the process of inferring missing data values in the network.
Missing network values can be inferred from existing data, allowing gaps in data to be filled with reasonable values. For example, if the width of a pipe is missing but the width of the pipe immediately upstream is known, it is reasonable to assume that the missing width is the same. The Inference object includes a set of inference rules that can be applied to all or part of the network. The use of a common set of rules within the software means that all missing data can be inferred to the same standard, rather than depending on who does the inference or when it is done.
Inference rules are viewed and edited in the Inference Data Editor.
InfoWorks ICM can infer the following missing values:
- Conduit parameters (Width, Height and Shape, Invert Levels, Headloss Type and Coefficient)
- Node Parameters (Coordinates and Ground Level)
Values can be inferred for either currently selected objects or the whole network. You can also choose which missing values are to be inferred and select the user-defined data flag that is to be applied to all inferred values. The use of a data flag is important for auditing purposes, as it allows you to distinguish between inferred data and data created in other ways.
Inferring Missing Values
Before missing values can be inferred, the inference parameters required must be set up in the Inference object. The parameters are viewed and edited in the Inference Data Editor. The Engineering Validation options should also be used to check the direction of pipes in the network before using inference.
To carry out the inference process:
- To apply inference to only part of the network, select the required objects. (See Selecting Network Objects.)
- To start the inference process, drag the Inference object onto the GeoPlan View of the network.
InfoWorks ICM carries out the data inference process for all selected objects. If there is no current selection, an option will be displayed to carry out inference for the entire network.
