Cycler control, data pipeline, and data visualisation from Empa's robotic battery lab.

• Tracks samples, experiments and results.
• Control Neware and Biologic cyclers on multiple machines from one place.
• Automatically collect and analyse cycling data.
• Results in consistent, open format including metadata with provenance tracking and sample information.
• Convenient cycler control and in-depth data exploration using Dash-based webapp.

Jobs

Aurora cycler manager can be used to control and submit experiments to Biologic and Neware cyclers.

Jobs can be submitted with a cycler-specific file (e.g. .xml or .mps).

Alternatively, a unicycler universal .json protocol can be used, which is converted to the appropriate format on submission.

Experiments can use C-rates and the program will automatically calculate the current required based on the sample information in the database.

Data harvesting

Data is automatically gathered from cyclers, all incoming files are converted to one open standard - accepts Biologic .mpr, Neware .ndax, Neware .xlsx, and tomato .json. Raw time-series data is converted to a hdf5 file including provenance tracked metadata.

Data is converted using NewareNDA and yadg, processing the raw binary data directly. This is much faster and more space efficient than exporting to text or Excel formats from these cyclers.

Analysis

The time-series hdf5 data is analysed to extract per-cycle summary data such as charge and discharge capacities, stored alongside metadata in a .json file.

Visualisation

A web-app based on Plotly Dash allows rapid, interactive viewing of time-series and per-cycle data, as well as the ability to control experiments on tomato cyclers through the graphical interface.

Installation

In a Python environment:

pip install aurora-cycler-manager

To view data from an existing set up:

aurora-setup connect --project-dir="path\to\your\setup"

To interact with servers on an existing set up:

• Interacting with servers (submitting jobs, harvesting data etc.) works with OpenSSH, servers must have OpenSSH installed and running
• Generate a public/private key pair on your system with ssh-keygen
• Copy your public key (usually in %USERPROFILE%\.ssh\id_rsa.pub) to the cycler server
• Add it to the server's authorized_keys file (usually in C:\Users\username\.ssh\authorized_keys)
• (optional) You can make changes to your user config, this is stored in your user folder e.g. /users/yourname/appdata/local/aurora_cycler_manager/
  • "SSH private key path" can be changed, if your key is not in a standard location
  • "Snapshots folder path" is where raw data downloaded from cyclers is stored, this can become very large

To create a new set up:

aurora-setup init --project-dir="path\to\your\setup"

• This generates subfolders within the directory, a database, and a configuation file
• Fill in the configuration file with details about e.g. Neware and EC-Lab servers, examples are left in the default config
• In Servers, the server_type must be neware, biologic, or tomato
• The server_label should be short and only letters and numbers, no special characters like -_/\
• The shell_type is the default shell when SSH-ing into the machine, it must be cmd or powershell
• To set up a neware server, follow the instructions from aurora-neware
• To set up a biologic server, follow the instructions from aurora-biologic
• To set up a tomato server, follow instructions from tomato-0.2.3
• If you change database columns in the shared configuration file, you can update the database with aurora-setup update

aurora-setup update

• Use the option --force if you want to permanetly delete columns and their data.

Updating

Upgrade with pip, you do not have to redo any setup steps:

pip install aurora-cycler-manager --upgrade

If upgrading from earlier than 0.5.0, first pip uninstall aurora-cycler-manager then follow the installation steps.

Usage

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

• There are three tabs, samples plotting, batch plotting, and database.
• To upload sample information to the database, use the 'Add samples' button in the database tab, and select a .json file defining the cells.
• Hand-made cells can also be added, a .json must be created with the keys defined in the configuration file.
• Protocols can be created in database -> protocols.
• In database -> pipelines, load samples onto the pipelines, then submit a protocol.
• Loading samples, submitting jobs, analysing data etc. can also be run in Python scripts directly - see the example in server_manager.py.

With SSH access, automatic data harvesting and analysis is run using:

aurora-daemon

Contributors

• Graham Kimbell

Acknowledgements

This software was developed at the Laboratory of Materials for Energy Conversion at Empa, the Swiss Federal Laboratories for Materials Science and Technology, and supported by funding from the IntelLiGent project from the European Union’s research and innovation program under grant agreement No. 101069765, and from the Swiss State Secretariat for Education, Research, and Innovation (SERI) under contract No. 22.001422.