Data from: "Lithium-ion battery degradation: comprehensive cycle ageing data and analysis for commercial 21700 cells"

Intro

Dataset from the publication "Lithium-ion battery degradation: comprehensive cycle ageing data and analysis for commercial 21700 cells", DOI: https://doi.org/10.1016/j.jpowsour.2024.234185

Full details of the study can be found in the publication, including thorough descriptions of the experimental methods and structure. A basic desciption of the experimental procedure and data structure is included here for ease of use.

Commercial 21700 cylindrical cells (LG M50T, LG GBM50T2170) were cycle aged under 3 different temperatures [10, 25, 40] °C and 4 different SoC ranges [0-30, 70-85, 85-100, 0-100]%, as well as a further [0-100]% SoC range experiment which utilised a drive-cycle discharge instead of constant-current. The same C-rates (0.3C / 1 C,  for charge / discharge) were used in all tests; multiple cells were tested under each condition. These are listed in the table below.

Experiment

SOC Window

Cycles per ageing set

Current

Temperature

Number of Cells

1

0-30%

257

0.3C / 1D

10°C

3

25°C

3

40°C

3

2,2

70-85%

515

0.3C / 1D

10°C

2

25°C

2

40°C

2

3

85-100%

515

0.3C / 1D

10°C

3

25°C

3

40°C

3

4

0-100% (drive-cycle)

78

0.3C / noisy D

10°C

3

25°C

2

40°C

3

5

0-100%

78

0.3C / 1D

10°C

3

25°C

2

40°C

3

Cells were base-cooled at set temperatures using bespoke test rigs (see our linked publications for details; the supporting information file contains detailed descriptions and photographs). Cells were subject to break-in cycles prior to beginning of life (BoL) performance tests using the ‘Reference Performance Test’ (RPT) procedures. They were then alternately subject to ageing sets and RPTs until the end of testing. Full details of each of these procedures are described in the linked publication.

The data contained in this repository is then described in the Data section below. This includes a description of the folder structure and naming conventions, file formats, and data analysis methods used for the ‘Processed Data’ which has been calculated from the raw data.

An 'experimental_metadata' .xlsx file is included to aid parsing of data. A jupyter notebook has also been included to demonstate how to access some of the data.

Data

Data are organised according to their parent ‘Experiment’, as defined above, with a folder for each. Within each Experiment folder, there are 3 subfolders: ‘Summary Data’, ‘Processed Timeseries Data’, and ‘Raw Data’.

Summary Data

This folder contains data which has been extracted by processing the raw data in the ‘Degradation Cycling’ and ‘Performance Checks’ folders. In most cases, the data you are looking for will be stored here.

It contains:    

Performance Summary

A summary file for each cell which details key ageing metrics such as number of ageing cycles, charge throughput, cell capacity, resistance, and degradation mode analysis results. Each row of data corresponds to a different SoH.

Degradation Mode Analysis (DMA) was also performed on the C/10 discharge data at each RPT. This analysis uses an optimisation function to determine the capacities and offset of the positive and negative electrodes by calculating a full cell voltage vs capacity curve using 1/2 cell data and comparing against the experimentally measured voltage vs capacity data from the C/10 discharge. See our ACS publication for more details.

Data includes:

·       Ageing Set: numbered 0 (BoL) to x, where x is the number of ageing sets the cell has been subject to.

·       Ageing Cycles: number of ageing cycles the cell has been subject to. *this is not equivalent full cycles.

·       Ageing Set Start Date/ End date: The date that each ageing set began/ ended.

·       Days of degradation: Number of days between the date of the first ageing set beginning and the current ageing set ending.

·       Age set average temperature: average recorded surface temperature of the cell during cycle ageing. Temperature was recorded approximately 1/2 way up the length of the cell (i.e. between positive and negative caps).

·       Charge throughput: total accumulated charge recorded during all cycles during ageing (i.e. sum of charge and discharge). This is the cumulative total since BoL (not including RPTs, and not including break-in cycles).

·       Energy throughput: as with "charge throughput", but for energy.

·       C/10 Capacity: the capacity recorded during the C/10 discharge test of each RPT.

·       C/2 Capacity: the capacity recorded during the C/2 discharge test of each even-numbered RPT.

·       0.1s Resistance: The resistance calculated from the 25-pulse GITT test of each even-numbered RPT. This value is taken from the 12th pulse of the procedure (which corresponds to ~52% SoC at BoL). The resistance is calculated by dividing the voltage drop by the current at a timecale of 0.1 seconds after the current pulse is applied (the fastest timescale possible under the 10 Hz recording condition).

·       Fitting parameters: output from the DMA optimisation function; 5 parameters which detail the upper/lower SoCs of each electrode, and the capacity fraction of graphite in the negative electrode.

·       Capacity and offset data: calculated based on the fitting parameters above alongside the measured C/10 discharge capacity.

·       DM data: Quantities of LLI, LAM-PE, LAM-NE, LAM-NE-Gr, and LAM-NE-Si calculated from the change in capacities/offset of each electrode since BoL.

·       RMSE data: the root mean squared error of the optimisation function calculated from the residual between the measured and simulated voltage vs capacity profiles.

Ageing Sets Summary

Data from the ageing cycles, summarised on an average per cycle and an average per ageing set basis. Metrics include mean/ max/ min temperatures, voltages etc.

Processed Timeseries data

Timeseries data (voltage, current, temperature, etc.) from each subtest (pOCV, GITT, etc.) of the RPTs, all grouped by subtest-type and by cell ID.

Contains the same data as in the ‘Performance Checks’ subfolder of the 'Raw Data' folder, but has been processed to slice into relevant subtests from the RPT procedure and includes only limited variables (time, voltage, current, charge, temperature). These are all saved as .csv files. In general this data will be easier to access than the raw data, but perhaps not as rich.

Raw Data

These are the raw data from the performance checks and from the degradation cycles themselves. The data from here has already been processed by me to get values of ‘energy throughput’, ‘charge throughput’, ‘average ageing temperature’, etc., which are all saved in the ‘Summary Data’ folder as described in the relevant section above.

The data in the ‘Degradation Cycling’ folder are organised by ageing set (where an ageing set is a defined number of ageing cycles, as described in the paper). In theory, each cell should have one datafile in each ageing set subfolder. However, due to experimental issues, tests can sometimes be interrupted midway though, requiring the test to be subsequently resumed. In this case, there may be multiple datafiles for each cell in a given ageing set; during analysis, these should be concatenated according to the descriptor in the filename (e.g., ‘cycling7’ + ‘cycling7 (part 2)').

Similarly, the unprocessed raw data from the performance checks (i.e. RPTs) is stored in the 'Performance Checks' folder, and structured in the same way.

The raw data are saved in the .mpr format produced by the Biologic battery cycler. This is a binary format which is storage-efficient but can be more difficult to process for analysis purposes. We have therefore also exported the data into .txt files (called .mpt) for the performance checks (RPTs) which make analysis easier. However, the exported .mpt files could not be included for the degradation cycling files due to their larger size. If you require access these degradation cycle data, the .mpr binary file can be parsed using the Galvani package in python, or you can use Biologic’s (proprietary) BT-Lab software to export the data into .txt files.

File Naming Convention

The raw datafiles are named with a standard format. This is:

            NDK - LG M50 deg - exp 1 - rig 1 - 10degC - cell A - RPT1_01_MB_CB1

            {NDK - LG M50 deg} - {exp 1} – {rig 1} – {10degC} – {cell A} – {RPT1}_{01}_{MB}_{CB1}

{Standard prefix} – {experiment number} – {ID of test rig} – {control temperature} – {Cell ID} – {RPT number or aging cycle number}_{step number for the characterisation procedure (see above)}_{experimental technique name (will always be “MB”)}_{battery cycler channel ID used (always the same for a particular cell/experiment)}