Panels are used to help organise models and help to manage visual complexity of larger models. The basic concept of a panel is that it is a collapsible area of modelling space. Real-world systems are often organised into "modules" which might represent analogues to real-world system segmentation. For example, companies are often sub-divided into departments.
Alternatively, there might be more than one representation of a system - for example movement of physical items, and a parallel movement of money, which can be simplified into separate model "regions".
We can use panels to do this. Take a look at the simple example below. In a single-view representation of the system, physical flows (raw materials, processing, and shipped goods) are represented alongside financials flows (cash outflows to purchase raw materials, and cash inflows from sales of shipped goods).
Panels provide an alternative way to represent this system, by subdividing it into two distinct components (Physical Goods and Financial). Panels can be arbitrarily composed (Panels within panels) to manage complex heirarchical or branching systems.
Within a panel, incoming connections (references) from other panels are displayed visually on the left-hand side. Note that a panel may receive connections from more than one panel.
Elements contained in panels can be referred to by using the following syntax:
"Panel Name" -> "Variable Name"
For a simple example of a model using panels, please refer to: https://sheetless.io/@john/panels