-- Lawrence Berkeley National Laboratory – Berkeley, CA
-- National Renewable Energy Laboratory – Golden, CO
-- Pacific Northwest National Laboratory – Richland, WA
-- Oak Ridge National Laboratory – Oak Ridge, TN
Performance Period: October 1, 2019 – September 30, 2022
Funding Type: Core Sensors & Controls Lab Call
Related Projects: Spawn, Open Building Control, VOLTTRON
The need for advanced control strategies (ACS) in buildings is growing due to emerging objectives to reduce energy consumption, integrate with electric power grid, integrate with district thermal networks, and improve responsiveness and service to occupants. Examples of ACS that are gaining attention are advanced rule-based control methods such as ASHRAE Guideline 36 sequences for VAV systems, various Model Predictive Control (MPC) approaches, and building-to-grid controls for demand response (DR). ACS also include automated fault detection and diagnosis (AFDD). While these ACS and others show promise, there is no framework that allows for comparing approaches across a standard set of buildings and climates. Standard comparison frameworks and benchmarks help guide research and help practitioners choose the most promising approaches to implement.
Building simulation can be used to create standard comparison and benchmarking frameworks through standardized building models, boundary conditions, solvers, and performance metrics.
The BOPTEST (Building Operation TESTing) Framework consists of a set of Modelica models that represent different buildings with different HVAC systems in different climate zones. BOPTEST exposes the "control points" of these models using a standard, familiar API that allows control algorithms to interact with the models as if they are physical buildings. The BOPTEST Framework also includes standardized key performance indicators (KPI) and reports and infrastructure for simulation-based comparison, benchmarking, and debugging of ACS.
BOPTEST is an open and level playing field on which different control algorithms can be quantitatively benchmarked and compared. In addition, it is a virtual environment field that supports meaningful experiments with control algorithms without the need for physical installations in existing buildings.