Cycler management, data pipeline, and data visualisation for Empa's robotic battery lab.
- Track samples, experiments and results with a database.
- Sample data is imported from the Aurora battery assembly robot.
- Automatically harvest and analyse cycling data.
- Results in consistent, open format including metadata with provenance tracking and sample information.
- Reads data from
tomatoservers, Biologic's EC-lab, and Neware's BTS software running on different machines. - Control experiments on
tomatoservers with a graphical interface. - Convenient, in-depth data exploration using
Dash-based webapp.
The Aurora cycler manager can be used to do all tomato functions (load, submit, eject, ready, cancel, snapshot) from one place to multiple tomato servers. Jobs can be submitted using C-rates, and can automatically calculate the current required based on measured electrode masses from the robot.
Functions are available to snapshot and download data from tomato servers. Harvesters are available to download new data from Biologic's EC-lab, converting from the closed .mpr filetype using yadg, and from Neware's BTS .xlsx reports or closed .ndax files using NewareNDA.
Full cycling data (voltage and current vs time) is converted to fast, efficient .h5 files with provenance tracked metadata. This data is analysed to extract per-cycle summary data such as charge and discharge capacities, stored alongside metadata in a .json file.
A web-app based on Plotly Dash allows rapid, interactive viewing of data, as well as the ability to control experiments on tomato cyclers through the graphical interface.
In a Python environment:
pip install git+https://github.com/EmpaEconversion/aurora-cycler-manager.git
After successfully installing, run and follow the instructions:
aurora-setup
To view data from an existing set up:
- Say yes to 'Connect to an existing configuration and database', then give the path to this folder.
To interact with servers on an existing set up:
- Interacting with servers (submitting jobs, harvesting data etc.) works with OpenSSH
- Generate a public/private key pair on your system with
ssh-keygen - Ensure your public key is authorized on the system running the cycler
- In config.json fill in 'SSH private key path' and 'Snapshots folder path'
- Snapshots folder path stores the raw data downloaded from cyclers which is processed. This data can be deleted any time.
- To connect and control
tomatoservers,tomato v0.2.3must be configured on the remote PC - To harvest from EC-lab or Neware cyclers, set data to save/backup to some location and specify this location in the shared configuration file
To create a new set up:
- Use
aurora-setupto create a configuration and database - it is currently designed with network storage in mind, so other users can access data. - Fill in the configuration file with details about e.g. tomato, Neware and EC-lab servers. Examples are left in the default config file.
Uninstall and reinstall with pip. From versions 0.5.x you do not have to re-do any of the setup steps, otherwise follow instructions in Installation.
pip uninstall aurora-cycler-manager
pip install git+https://github.com/EmpaEconversion/aurora-cycler-manager.git
A web app allows users to view analysed data and see the status of samples, jobs, and cyclers, and submit jobs to cyclers if they have access. Run with:
aurora-app
To upload sample information to the database, place output .csv files from the Aurora robot into the samples folder defined in the configuration.
Hand made cells can also be added, a .csv must be created with the headers defined in the shared configuration.
Loading samples, submitting jobs etc. can be performed on tomato directly, or using the aurora-app GUI, or by writing a Python script to use the functions in server_manager.py.
With SSH access, automatic data harvesting and analysis is run using:
aurora-daemon
This software was developed at the Materials for Energy Conversion Lab at the Swiss Federal Laboratories for Materials Science and Technology (Empa).