BOPTEST (Building Optimization Performance Tests)

The testcases/ directory contains various building optimization models used for testing. Test cases range from simple prototype models to complex multi-zone commercial buildings.

Each test case supports different combinations of:

• Electricity Prices: constant, dynamic, or highly_dynamic.
• Time Periods: Specific scenarios like test_day, peak_heat_day, typical_cool_day, etc., depending on the model.

For detailed information on a specific test case, navigate to /<testcase_dir_name>/docs within the repository.

BOPTEST (Building Optimization Performance Tests) is a platform designed for simulation-based testing and benchmarking of advanced building control strategies. It provides a standardized environment for researchers and industry to evaluate control performance through four main components:

1. Reference building emulation test cases: Standardized building models available for testing.
2. Key Performance Indicators (KPIs): Metrics used to assess the performance of a control strategy.
3. Software platform: A system to manage test cases, facilitate data exchange (control and measurement) between user software and building emulations, calculate KPIs, and generate reports.
4. Benchmarking procedures: Standardized methods and user guides for testing controls development.

The twozone_apartment_hydronic test case directory contains the following components:

• doc/: Documentation specific to this test case.
• models/: Emulation model files.
• config.py: Configuration used for building the Docker image for this test case.

BOPTEST Test Case 1 is a simple RC network with a feedback-controlled heating device. The test case directory is organized as follows:

• doc/: Contains documentation specific to this test case.
• models/: Contains the emulation model files.
• config.py: Defines the configuration used for building the Docker image for this test case.

Beyond the core service, several interfaces are available for different programming environments and protocols:

• OpenAI-Gym: An environment interface for reinforcement learning, available in ibpsa/project1-boptest-gym.
• BACnet: A BACnet interface located in the /bacnet directory of this repository.
• Julia: A Julia package named BOPTestAPI.jl for interacting with BOPTEST.
• Results Dashboard: A proposed web interface for account management and result sharing.

The weather forecast uncertainty emulator adds stochasticity to deterministic BOPTEST forecasts to simulate real-world behavior. This allows developers to evaluate control strategies against realistic forecast errors.

It models uncertainty for two specific variables using a first-order autoregressive model (AR(1)):

• Dry-bulb temperature (ambient temperature): Uses Gaussian-distributed residuals.
• Global Horizontal Irradiation (GHI): Uses Laplace-distributed residuals to match the heavy-tailed nature of GHI errors.

GHI Special Handling:

• Nighttime: If measured GHI is zero, no error is added ($\hat{X}_{\tau|k} = 0$).
• Smoothing: A 3-sample moving average filter is applied to reduce high-frequency noise.

BOPTEST-Service utilizes Redis for two primary functions:

1. Test Metadata Storage: Persisting state and information about specific tests.
2. Message Bus: Acting as a Pub/Sub mechanism to facilitate communication between the service and workers.