Skip to content

Introduction to System Modeling

Broadly speaking, "system modeling" refers to the process of starting with raw, unstructured data acquired from a control system, and building a normalized representation of the data and equipment. This includes creating entities to represent different pieces of equipment, tagging data points with standard labels, and adding attributes and relationships between different pieces of equipment.

flowchart LR
  raw[Raw Data]
    --> equip[Equipment Extraction]
    --> tagging[Point Tagging]
    --> relations[Relationship Modeling]
    --> qa[Validation]

The complexity of this task is determined entirely by which applications are going to consume this data model. For instance, in some applications with simple data requirements, a very simple data modeling process may be sufficient which only models a few types of equipment and points.

Info

Normal Framework primarily supports modeling relationships points and equipment. Future releases will add more support for modeling entities such as locations and other important objects.

When discussing system modeling, we assume that all relevant points have already been discovered or imported into Normal; for instance by performing bacnet discovery.

Normal Framework Approach to Modeling

Because system modeling can be such a complicated and time-consuming process, Normal has adopted a few high-level principles to how system modeling is performed.

  1. Build equipment models first: The starting point for modeling is to create entities which represent actual pieces of equipment; e.g., terminal units, air handlers, pumps, and so on. When we model equipment, we model both the type of equipment and the desired points we expect to label on the equipment.

  2. Everything in batch mode: unlike other control systems where many controllers have custom, built-up logic, building controls tend to be highly repetitive with tens or hundreds of nearly identical pieces of equipment per site. Therefore, all of our modeling flows encourage the use of batch techniques to handle highly repetitive environments.

  3. Batteries included: Normal ships with an included, robust set of templates based on the Project Haystack. These provide a rich starting point for system modelers which include many different types of equipment. Advanced modelers can change the ontology to better fit their use case.