Power quality measurement (IiA 20139)
Invest-in-Arup research project to design and make a convenient way to measure and study the power quality and energy performance of single phase small power and lighting devices, that connect via plug-and-socket.
See IiA page for more project details https://invest.arup.com/?layout=projsheet&projid=20139
The software is designed to run on Raspberry Pi and Raspberry Pi Pico microcontroller, communicating with an analog-to-digital converter (ADC: MCP3912) on a custom PCB to source measurements of power parameters: voltage, current (full range), current (low range) and earth leakage current. Following conversion and scaling to floating point measurement values, the physical measurement samples are further processed to provide a variety of derivative measurements, waveform visualisation and logging. A local touch screen provides user interface.
Partly to enable rapid development cycles, but also to validate calculations, the software will also run on desktop computers using a simulator to take the place of the Pico/ADC to generate sample data.
Raspberry Pi
Minimum hardware requirement is Raspberry Pi 3B+. Non-plus models do not have sufficent thermal dissipation for high CPU loads. Touch screen display with DSI interface, 800x480 pixels. Raspberry Pi OS (32 bit) with desktop. Prepare the image with connectivity to local/required wifi, SSH access, and change the default password. If you have one, pair a Bluetooth keyboard for local access.
Raspberry Pi Pico
Requires Thonny on your laptop computer to install MicroPython on the Pico and verify that you have local REPL access to the microcontroller via a USB cable.
Raspberry PI SD card
Prepare the SD card using Raspberry Pi Imager. As a convenience, a USB stick with Raspberry Pi OS can be prepared first, plugged in the Pi and then that image booted to write to an in-situ SD card -- avoids the need to remove ribbon cable. If using this method, note that a non-initialised Pico must be unplugged from the Pi USB port for this to work. Set a project specific password. Enable the option for remote SSH access. Set the wifi country, SSID and password to be used for initial access.
Pi setup
- Start a terminal session via SSH. (As alternative, you could use a bluetooth keyboard and work locally.)
ssh -X [email protected]
- Clone the repository:
git clone https://github.com/arup-group/pqm-hellebores.git
- Change to the working directory. Create a virtual environment, and activate it:
cd pqm-hellebores
python -m venv .venv
source .venv/bin/activate
- Install the dependencies (the additional mathematics libraries are needed to support the numpy library):
sudo apt install libatlas-base-dev libopenblas-dev
python -m pip install -r requirements.txt
- Set the device identity. In place of PQM-n, use the identity of your specific power quality monitor, eg PQM-1:
echo PQM-n > configuration/identity
- Install bluetooth file transfer client (optional):
sudo apt install blueman
- Enable Raspberry Pi OS features:
In the Preferences | Raspberry Pi Configuration menu, enable the VNC server and the on-screen keyboard: the keyboard is found in the 'Display' tab -- select 'Enable always', and the VNC server is found in the 'Interfaces' tab -- verify that both SSH and VNC are enabled.
- Add shortcuts to the 'Other' desktop menu:
ln -s /home/pi/pqm-hellebores/run/hellebores.desktop /home/pi/.local/share/applications/hellebores.desktop
ln -s /home/pi/pqm-hellebores/run/pqm-launcher.desktop /home/pi/.local/share/applications/pqm-launcher.desktop
- Also add the launcher script to the desktop autostart directory.
ln -s /home/pi/pqm-hellebores/run/pqm-launcher.desktop /home/pi/.config/autostart/pqm-launcher.desktop
Dependencies sometimes change with OS updates. See notes file for potential additional system dependencies and troubleshooting.
Pico setup
Using Thonny:
Open ./pico/main.py and save it to the root folder of the microcontroller.
Open ./pico/stream.py and save it to the root folder of the microcontroller.
To use the interactive user interface, select "START" from the launcher dialog. Alternatively, run "hellebores" from the Raspberry Pi desktop menu.
It's possible to access the running environment on a Pi using SSH. For working on the Pico microcontroller, use ssh -X to connect to the Pi and then start thonny within your SSH session. This will redirect thonny's program window to your local X-server (needs a Linux or WSL2 system, or install an X-server separately).
Ubuntu (incl. WSL) and macOS
On a system with python3 and git installed, proceed as follows:
git clone https://github.com/arup-group/pqm-hellebores.git
cd pqm-hellebores
python3 -m venv .venv
source .venv/bin/activate
python3 -m pip install -r requirements.txt
echo PQM-0 > configuration/identity
run/go.sh
Windows
In a command window with python and git available:
git clone https://github.com/arup-group/pqm-hellebores.git
cd pqm-hellebores
python -m venv .venv-windows
.venv-windows\scripts\activate
python -m pip install -r requirements.txt
echo PQM-0 > configuration\identity
run\go
NB Some code paths are different on Windows to allow the project to run, and the files mswin_pipes.py, go.bat and go.py are exclusive to Windows. Note that some features of the go.sh script (eg software update) are not implemented.
The software is designed with 'Unix Philosophy' in mind, for performance and simplicity benefits. Independent programs are connected in a pipeline with data sent from one program to the next as a text stream. Control settings are stored in a settings.json file: the GUI program modifies this file and sends other programs a Posix signal whenever the settings file is updated, and they then re-load the settings.json. The project is intended to yield to experimentation: the behaviour of the programs at each step of the pipeline can be studied, verified and changed on the command line.
The Pi code will run in WSL, Ubuntu, Mac or similar environments that have a modern python and can support the python dependencies, pipelines and signals. hellebores.py and the rain_chooser.py simulator require SDL for displaying the GUI and to accept touchscreen or mouse input. The GUI can run under X-Windows or in native Mac and MS-Windows if the SDL libraries are installed and are reachable from the python runtime environment. The run script will test whether a serial interface to Pico is available and if not will use the rain_chooser.py simulator as a data source.
When working with git, it's possible to automate minor version increments when making a commit, by editing lines at the end of .git/hooks/pre-commit.sample and then rename the hook file to .git/hooks/pre-commit.
# If there are whitespace errors, print the offending file names and fail.
#exec git diff-index --check --cached $against --
if ! git diff-index --check --cached $against --; then
exit 1
fi
# Increment version number and then add the VERSION file to the commit
./pqm/version.py --increment_sub_version
git add ./VERSION
The four channels of the ADC each have associated calibration factors: DC offset, analogue gain and timing skew. The calibration factors for each device are stored in the calibrations.json file. The identity of the specific device at hand is stored in the configuration/identity file. Hardware scaling and calibration factors are applied to the data stream by the scaler.py program.
The calibrator.py program is used in conjunction with a calibrated true RMS multimeter. The calibration procedure is used to determine device-specific calibration factors to be entered into the calibrations.json file.