6. Software Architecture¶
This section describes the software architecture of the controls design tool and the functional verification tool. In the text below, we mean by plant the HVAC and building system, and by control the controls other than product integrated controllers (PIC). Thus, the HVAC or building system model may, and likely will, contain product integrated controllers, which will be out of scope for CDL apart from reading measured values from PICs and sending setpoints to PICs.
Fig. 6.1 shows the overall system with the Controls Design Tool and the Functional Verification Tool. Both use a CDL Parser which parses the CDL language. This parser is currently in development at https://github.com/lbl-srg/modelica-json. 1 The CDL parser reads a CDL-compliant Control Sequence, which may be provided by the user or taken from https://simulationresearch.lbl.gov/modelica/releases/v10.0.0/help/Buildings_Controls_OBC_ASHRAE.html and the CDL Library, see https://simulationresearch.lbl.gov/modelica/releases/v10.0.0/help/Buildings_Controls_OBC_CDL.html All these components will be made available through OpenStudio. This allows using the OpenStudio model authoring and simulation capability that is being developed for the Spawn of EnergyPlus (SOEP). See also https://www.energy.gov/eere/buildings/articles/spawn-energyplus-spawn and its development site https://lbl-srg.github.io/soep/softwareArchitecture.html.
6.1. Controls Design Tool¶
Fig. 6.2 shows the overall software architecture of the controls design tool. The OpenStudio invokes a Modelica to json parser which parses the Modelica libraries to json, and it invokes the HVAC/controls tool. The HVAC/controls tool reads the json representation of the Modelica libraries that are used. The HVAC/controls tool updates the json reprensentation of the model, and these changes will be merged into the Modelica model or Modelica package that has been edited. For exporting the sequence for simulation or for operation, OpenStudio invokes OPTIMICA which generates an FMU of the sequence, or multiple FMUs if the sequence is to be distributed to different field devices. The Building Operating System then imports these FMUs.
If a Building Automation System prefers not to run FMUs to compute the control signals, then it could convert the json format to a native implementation of the control sequence.
Optionally, to aid the user in customizing sequences, a Sequence Generator could be generated. This is currently not shown in Fig. 6.2. The Sequence Generator will guide the user through a series of questions about the plant and control, and then generates a Control Model that contains the open-loop control sequence. This Control Model uses the CDL language, and can be stored in the Custom or Manufacturer Modelica Library. Using the HVAC/controls tool, the user will then connect it to a plant model (which consist of the HVAC and building model with exposed control inputs and sensor outputs). This connection will allow testing and modification of the Control Model as needed. Hence, using the Schematic editor, the user can manipulate the sequence to adapt it to the actual project.
How sequences can be exported to control systems is described in Section 10.
6.2. Functional Verification Tool¶
The Functional Verification Tool consists of three modules:
An I/O Configuration module that adds I/O information to the point list,
a Engine that is used to conduct the actual verification, and
a Viewer that displays the results of the verification.
The Functional Verification Tool uses that same CDL Parser as is used
for the Controls Design Tool.
The I/O Configuration module will allow users (such as a
commissioning agent) to update the point list.
This is needed as not all
point mappings may be known during the design phase.
The Engine invokes OPTIMICA to export an FMU-ME of the control
blocks. As OPTIMICA does not parse CDL information
that is stored in vendor annotations (such as the point mapping),
the Engine will insert point lists into the
of the FMU-ME.
To conduct the verification, the Engine will connect to a
HIL Module, such as Volttron or the BCVTB, and set up a
closed loop model, using the point list from the FMU’s
During the verification, the Engine will write reports
that are displayed by the Viewer.
Using a parser that only requires Java has the advantage that it can be used in other applications that may not have access to a OPTIMICA installation.