RL control + new feedback signal for cooling towers

*.pyc

__pycache__

*.ipynb_checkpoints

*.mp4

 No newline at end of file

### Installation for ESB condenser water control

The controller (`src/controller.py`) imports measurements from BDX (Buildinglogix Data eXchange), and writes a setpoint to a text file. Access to BDX requires login credentials.

To run the controller, (1) review control settings, (2) create/update the production environment, (3) activate it, call the script.

The environment needs to be created only once. Every time after, the environment is activated and the scipt is called.

#### 0. Quick start

Refer to this section if you're aware of how the script is set up, and you just want to restart it.

```

cd CODE_DIRECTORY                             # navigate to code directory

git reset --hard                              # ignore local changes

git pull                                      # update code from remote branch

conda env update -f=./environment.yml         # update virtual environment (assuming already created)

conda activate esb-prod                       # activate environment

python src/controller.py -s src/settings.ini  # run script, giving path to settings file

```

#### 1. Review settings before running the controller

The controller settings can be specified in a settings file. The default settings file is located in `src/settings.ini`. A file can be created at another location but the file path must be specified when running the controller script. To create a custom settings file, just copy `src/settings.ini` to a location of your choice.

1. Fill in the `username` and `password` values for accessing BDX in the settings file, OR:

    1. Optionally, create `src/bdxcredentials.ini` with these contents instead of populating settings fields. This will prevent the `settings.ini` file credentials from being erased when the repository is updated from a remote branch.

    ```

    [DEFAULT]

    username = USERNAME

    password = PASSWORD

    ```

2. Choose which controllers are to be run by adding a comma separated list of sections in the `controllers` field. For e.g. `controllers=CONTROLLER.ESB,CONTROLLER.KISSAM`. Fill their respective fields as well.

3. Change verbosity settings.

4. Override any default settings in any non-`DEFAULT` section. For e.g. if `verbosity` is specified under `DEFAULT` and `CONTROLLER.ESB`, then for the ESB controller, the more specific setting will take effect over default.

#### 2. Create/update the production environment

Using `conda` (Anaconda/Miniconda package manager):

```bash

# Create new environment

conda env create -f=./environment.yml

# OR, update production environment

conda env update -f=./environment.yml

```

#### 3. Activate environment, run scripts

```bash

python src/controller.py --help

usage: controller.py [-h] [-s SETTINGS] [-l LOGS] [-r LOGS_SERVER]

                     [-v {CRITICAL,ERROR,WARNING,INFO,DEBUG}] [-d] [-n]

Condenser set-point optimization script.

optional arguments:

  -h, --help            show this help message and exit

  -s SETTINGS, --settings SETTINGS

                        Location of settings file.

  -l LOGS, --logs LOGS  Location of file to write logs to.

  -r LOGS_SERVER, --logs-server LOGS_SERVER

                        http://host[:port][/path] of remote server to POST

                        logs to.

  -v {CRITICAL,ERROR,WARNING,INFO,DEBUG}, --verbosity {CRITICAL,ERROR,WARNING,INFO,DEBUG}

                        Verbosity level.

  -d, --dry-run         Exit after one action to test script.

  -n, --no-network      For testing code execution: no API calls.

Additional settings can be changed from the specified settings ini file.

```

Example commands:

```bash

# update production environment

conda env update -f environment.yml

# activate production environment

conda activate esb-prod

# This will exit after taking one action as a test run:

# Uses settings from the default location (src/settings.ini)

# Writes to default logs file (./logs.txt)

python src/controller.py --dry-run

# Or, this will read settings from a custom location

# If settings are provided at command line and in settings file, command line

# will take precedence.

python src/controller.py --settings ~/ESB/mysettings.ini

```

### Monitoring controller

Besides logs written to error stream and a local file, the controller can also communicate with a HTTP server or send emails.

The code comes with a `src/monitor.py` script that can listen in as a HTTP server to updates sent by `src/controller.py` running on a different machine. Settings can be specified in the settings ini file. For duplicate settings, the monitor falls back on values in the `DEFAULTS` section in the ini for the controller if it cannot find them in the `MONITOR` section.

```bash

$ python src/monitor.py --help

usage: monitor.py [-h] [-s SETTINGS] [-l LOGS] [-p PORT] [--host HOST]

                  [-v {CRITICAL,ERROR,WARNING,INFO,DEBUG}]

Monitor for condenser set-point optimization script.

optional arguments:

  -h, --help            show this help message and exit

  -s SETTINGS, --settings SETTINGS

                        Location of settings file.

  -l LOGS, --logs LOGS  Location of file to write logs to.

  -p PORT, --port PORT  Port number of server to listen on.

  --host HOST           Host address of server to listen on. e.g. 0.0.0.0

  -v {CRITICAL,ERROR,WARNING,INFO,DEBUG}, --verbosity {CRITICAL,ERROR,WARNING,INFO,DEBUG}

                        Verbosity level.

Additional settings can be changed from the specified settings ini file.

```

Refer to the [controller documentation](docs/8-control.md).

### Development

channels:

  - conda-forge

dependencies:

  - python=3.7

  - python=3.9

  - numpy

  - pandas

  - matplotlib

  - jupyter_contrib_nbextensions

  - tqdm

  - scikit-learn

  - pytorch::pytorch=1.8

  - pytorch::pytorch=1.10

  - cudatoolkit=10.2

  - tslearn

  - optuna=2.8.0

  - dash=1.17.0

  - optuna

  - dash

  - pyfmi

  - pip

  - pip:

    - python-dateutil

    - tensorboardX

    - higher==0.2.1

    - git+https://git.isis.vanderbilt.edu/SmartBuildings/bdx.git@v0.2.6

    - git+https://github.com/CPCLAB-UNIPI/SIPPY@6c233881f32280ee79267cba06b2f56984424ebd#egg=sippy

 No newline at end of file

---

title: 'Kissam Controller Notes'

order: 8

hasequations: false

---

### Installation for ESB condenser water control

The controller (`src/controller.py`) imports measurements from BDX (Buildinglogix Data eXchange), and writes a setpoint to a text file. Access to BDX requires login credentials.

To run the controller, (1) review control settings, (2) create/update the production environment, (3) activate it, call the script.

The environment needs to be created only once. Every time after, the environment is activated and the scipt is called.

#### 0. Quick start

Refer to this section if you're aware of how the script is set up, and you just want to restart it.

```

cd CODE_DIRECTORY                             # navigate to code directory

git reset --hard                              # ignore local changes

git pull                                      # update code from remote branch

conda env update -f=./environment.yml         # update virtual environment (assuming already created)

conda activate esb-prod                       # activate environment

python src/controller.py -s src/settings.ini  # run script, giving path to settings file

```

#### 1. Review settings before running the controller

The controller settings can be specified in a settings file. The default settings file is located in `src/settings.ini`. A file can be created at another location but the file path must be specified when running the controller script. To create a custom settings file, just copy `src/settings.ini` to a location of your choice.

1. Fill in the `username` and `password` values for accessing BDX in the settings file, OR:

    1. Optionally, create `src/bdxcredentials.ini` with these contents instead of populating settings fields. This will prevent the `settings.ini` file credentials from being erased when the repository is updated from a remote branch.

    ```

    [DEFAULT]

    username = USERNAME

    password = PASSWORD

    ```

2. Choose which controllers are to be run by adding a comma separated list of sections in the `controllers` field. For e.g. `controllers=CONTROLLER.ESB,CONTROLLER.KISSAM`. Fill their respective fields as well.

3. Change verbosity settings.

4. Override any default settings in any non-`DEFAULT` section. For e.g. if `verbosity` is specified under `DEFAULT` and `CONTROLLER.ESB`, then for the ESB controller, the more specific setting will take effect over default.

#### 2. Create/update the production environment

Using `conda` (Anaconda/Miniconda package manager):

```bash

# Create new environment

conda env create -f=./environment.yml

# OR, update production environment

conda env update -f=./environment.yml

```

#### 3. Activate environment, run scripts

```bash

python src/controller.py --help

usage: controller.py [-h] [-s SETTINGS] [-l LOGS] [-r LOGS_SERVER]

                     [-v {CRITICAL,ERROR,WARNING,INFO,DEBUG}] [-d] [-n]

Condenser set-point optimization script.

optional arguments:

  -h, --help            show this help message and exit

  -s SETTINGS, --settings SETTINGS

                        Location of settings file.

  -l LOGS, --logs LOGS  Location of file to write logs to.

  -r LOGS_SERVER, --logs-server LOGS_SERVER

                        http://host[:port][/path] of remote server to POST

                        logs to.

  -v {CRITICAL,ERROR,WARNING,INFO,DEBUG}, --verbosity {CRITICAL,ERROR,WARNING,INFO,DEBUG}

                        Verbosity level.

  -d, --dry-run         Exit after one action to test script.

  -n, --no-network      For testing code execution: no API calls.

Additional settings can be changed from the specified settings ini file.

```

Example commands:

```bash

# update production environment

conda env update -f environment.yml

# activate production environment

conda activate esb-prod

# This will exit after taking one action as a test run:

# Uses settings from the default location (src/settings.ini)

# Writes to default logs file (./logs.txt)

python src/controller.py --dry-run

# Or, this will read settings from a custom location

# If settings are provided at command line and in settings file, command line

# will take precedence.

python src/controller.py --settings ~/ESB/mysettings.ini

```

### Monitoring controller

Besides logs written to error stream and a local file, the controller can also communicate with a HTTP server or send emails.

The code comes with a `src/monitor.py` script that can listen in as a HTTP server to updates sent by `src/controller.py` running on a different machine. Settings can be specified in the settings ini file. For duplicate settings, the monitor falls back on values in the `DEFAULTS` section in the ini for the controller if it cannot find them in the `MONITOR` section.

```bash

$ python src/monitor.py --help

usage: monitor.py [-h] [-s SETTINGS] [-l LOGS] [-p PORT] [--host HOST]

                  [-v {CRITICAL,ERROR,WARNING,INFO,DEBUG}]

Monitor for condenser set-point optimization script.

optional arguments:

  -h, --help            show this help message and exit

  -s SETTINGS, --settings SETTINGS

                        Location of settings file.

  -l LOGS, --logs LOGS  Location of file to write logs to.

  -p PORT, --port PORT  Port number of server to listen on.

  --host HOST           Host address of server to listen on. e.g. 0.0.0.0

  -v {CRITICAL,ERROR,WARNING,INFO,DEBUG}, --verbosity {CRITICAL,ERROR,WARNING,INFO,DEBUG}

                        Verbosity level.

Additional settings can be changed from the specified settings ini file.

```

 No newline at end of file

hasequations: false

---

Data for setpoint is recorded in BDX trend 3481. Cooling tower 1 and 2 data is in trends 2422, 2841.

## Periods

Timestamps for various system changes.

    2021-06-15 1403-5   Script errored, couldn't read BDX due to faulty TLS certificates

    2021-06-16 1240-5   Script restarted with correct paths

    2021-06-17 0505-5   Chiller 2 online (check exact time)

    2021-08-15 0610-5   Controller offline. Server was restarted.

    2021-08-17 1214-5   Controller restarted.

    2021-08-31 1100-5   Chiller 1 online

## Timespans to investigate