RELIABILITY ENGINEERING & OPTICS

Research materials

Pseudo datasets and research code for reliability modeling,
lifetime prediction, and optical engineering.

Explore datasets Browse research code

04Datasets

05Code packages

Datasets

Pseudo data based on original measurements. For testing only.

4 datasets

Blue laser · Pseudo datasetCSV

450 nm blue laser pseudo dataset

18trajectories

Based on original measurements.For testing and benchmarking under simulated conditions only.

This pseudo dataset contains 18 generated degradation trajectories for 450 nm blue laser diodes, sampled every 0.5 h. Power and voltage trends are constrained by original measurements. Device-specific aging-current changes after 1500 h are retained, with smooth changes in degradation rate while the recorded aging temperature stays constant.

Attribution: Chizhi Chris Zhang; Changchun Institute of Optics, Fine Mechanics and Physics, Chinese Academy of Sciences; China Electronic Product Reliability and Environmental Testing Research Institute (CEPREI; the Fifth Electronics Research Institute of MIIT), Guangzhou

450 nm35–70 °C50–350 mA aging600 mA characterizationEvery 0.5 h

Download dataset CSV · ZIP · 1.3 MB

Please cite the associated article when using this dataset.

IEEE citation

[1] C. C. Zhang, X. Feng, and J. Jiang, “Segmental Viterbi estimation using transition constraints for covariate-dependent hidden semi-Markov degradation models with right censoring,” Reliab. Eng. Syst. Saf., vol. 278, Art. no. 113615, Feb. 2027, doi: 10.1016/j.ress.2026.113615.

Files & variables

Files

One CSV per device, with cumulative aging time. Each trajectory ends at the device’s last original measurement time (700–2700 h). Apply a chosen failure threshold and treat non-crossing records as right-censored.

BLD_450_Dataset/PBLD_450_Dxx.csv (18 device files)

Variables

Field	Description
time_h	Cumulative aging time (h)
current_mA	Characterization current (mA), fixed at 600 mA
voltage_V	Simulated device voltage at the characterization current (V)
power_mW	Simulated optical power at the characterization current (mW)
aging_current_mA	Aging current (mA), including recorded changes after 1500 h where applicable
aging_temperature_C	Recorded aging temperature (°C), constant within each device

SHA-256 184889883e735b5b2ef41bbde172b5ded74d1195f10934030162e4a443c8b500

Blue laser · Pseudo datasetCSV

455 nm blue laser pseudo dataset

20trajectories

Based on original measurements.For testing and benchmarking under simulated conditions only.

This pseudo dataset contains 20 generated blue-laser degradation trajectories at 3.0, 3.5, 4.0, and 4.5 A. The source tests used a 25 °C water-cooling setpoint. Each trajectory covers 0–500 h at 0.5 h intervals, with optical power, current, voltage, wavelength, and bandwidth. Low-output spectral fluctuations follow the patterns in the original instrument records.

455 nm3.0–4.5 A25 °C cooling setpoint0–500 hEvery 0.5 h

Download dataset CSV · ZIP · 104 KB

Please cite the associated article when using this dataset.

IEEE citation

Files & variables

Files

One CSV per pseudo device. Trajectories cover 0–500 h; records that do not reach the chosen failure threshold by 500 h are right censored. Use the 20 h optical-power value as the reference for normalized-power analyses. Large wavelength and bandwidth fluctuations at very low optical output represent instrument-like tail readings.

BLD_455_Dataset/BLD_455_001.csv … BLD_455_020.csv

Variables

Field	Description
time_h	Elapsed test time (h)
current_A	Drive current (A)
voltage_V	Device voltage (V)
power_W	Optical power (W)
wavelength_nm	Generated wavelength reading (nm); low-output readings may be unstable
bandwidth_nm	Generated spectral bandwidth reading (nm); low-output readings may be unstable

SHA-256 005cc3feaeaeec625e0119503909fda0723abd3a3f83c8f38645c65b84ff8c4f

VCSEL · Pseudo datasetCSV

795 nm VCSEL pseudo dataset

62trajectories

Based on original measurements.For testing and benchmarking under simulated conditions only.

This pseudo dataset contains 62 generated full-lifetime VCSEL degradation trajectories at 795 nm under different aging temperatures and drive currents. It supports simulated reliability-method research, lifetime prediction, and algorithm comparison.

Attribution: Chizhi Chris Zhang; Changchun Institute of Optics, Fine Mechanics and Physics, Chinese Academy of Sciences; Changchun ACE Photonics Co., Ltd.

795 nm50–120 °C2.0–4.0 mAHourly

Download dataset CSV · ZIP · 24.9 MB

Please cite the associated article when using this dataset.

IEEE citation

[1] C. C. Zhang, X. Feng, and H. Zhang, “Variable-length multiobjective optimization of inspection schedules for condition-based maintenance of vertical-cavity surface-emitting lasers under imperfect detection,” Reliab. Eng. Syst. Saf., vol. 277, Art. no. 113039, Jan. 2027, doi: 10.1016/j.ress.2026.113039.

Files & variables

Files

One CSV per device. Each trajectory starts at 0 h and ends when optical power first reaches zero.

VCSEL_795_Dataset/VCSEL_795_001.csv … VCSEL_795_062.csv

Variables

Field	Description
time_h	Elapsed time (h)
power_mW	Optical power (mW)
current_mA	Drive current (mA)
aging_temp_C	Aging temperature (°C)
wavelength_nm	Nominal wavelength (nm)
power0_mW	Initial optical power (mW)
relative_power	Power / initial power
voltage_V	Initial device voltage (V)

SHA-256 57732839df55c670548c4dcc620402768688325f680465c50381eda832b993e0

VCSEL · Pseudo datasetCSV

894 nm VCSEL pseudo dataset

45trajectories

Based on original measurements.For testing and benchmarking under simulated conditions only.

This pseudo dataset contains 45 generated full-lifetime VCSEL degradation trajectories at 894 nm under different aging temperatures and drive currents. It supports simulated reliability-method research, lifetime prediction, and algorithm comparison.

Attribution: Chizhi Chris Zhang; Changchun Institute of Optics, Fine Mechanics and Physics, Chinese Academy of Sciences; Changchun ACE Photonics Co., Ltd.

894 nm100–120 °C1.5–2.0 mAHourly

Download dataset CSV · ZIP · 4.2 MB

Please cite the associated article when using this dataset.

IEEE citation

Files & variables

Files

One CSV per device. Each trajectory starts at 0 h and ends when optical power first reaches zero.

VCSEL_894_Dataset/VCSEL_894_001.csv … VCSEL_894_045.csv

Variables

Field	Description
time_h	Elapsed time (h)
power_mW	Optical power (mW)
current_mA	Drive current (mA)
aging_temp_C	Aging temperature (°C)
wavelength_nm	Nominal wavelength (nm)
power0_mW	Initial optical power (mW)
relative_power	Power / initial power
voltage_V	Initial device voltage (V)

SHA-256 db1a1b7c628c1062d16e6a327f5bc64f84903eaf99c05302dcafd37e486634be

Research code

Analysis scripts and companion materials

5 code packages

ReliabilityJDMDPublished

Prefix Treatment and Bayesian Partial Pooling for Early-Stage Lifetime Prediction from Degradation Data

C. C. Zhang, X. Feng, and J. Jiang. Journal of Dynamics, Monitoring and Diagnostics, vol. 5, no. 2, pp. 117–130, May 2026.

Bayesian partial pooling and prefix-based lifetime estimation for laser and battery degradation.

Download ZIP Python · 21 KB

Package details

In the package

Laser analysis and figures
Battery analysis and figures

Code only. The original laser and Oxford battery inputs are required for full reruns.

Run from the extracted project root

python -m pip install -r requirements.txt
python scripts/final_laser_figures.py
python scripts/final_battery_figures.py

SHA-256 7b66b95838c0cfaee120b2f604b9cd156cfaf8002a283c225d62e32db7fd3e6d

ReliabilityRESSPublished

Variable-length multiobjective optimization of inspection schedules for condition-based maintenance of vertical-cavity surface-emitting lasers under imperfect detection

C. C. Zhang, X. Feng, and H. Zhang. Reliability Engineering & System Safety, vol. 277, Art. no. 113039, 2027.

Single and multiple inspection decisions, fixed and variable-length schedules, and imperfect detection.

Download ZIP Python · 37 KB

Package details

In the package

Processed study datasets
Four standalone case-study scripts

The existing public archive is preserved. It includes processed data and analysis scripts.

Run from the extracted project root

python scripts/case1_single_inspection_data.py
python scripts/case2_two_inspection_optimization.py
python scripts/case3_fixed_length_optimization.py
python scripts/case4_variable_length_optimization.py

SHA-256 73551414de7373572854d615a6038891cafbf9edef8f8907abc70e6ea8ce6796

OpticsOptics ExpressPublished

Rapid stress birefringence mapping in mounted curved lenses with a curved-layer Jones approximation

C. C. Zhang, S. Zhang, H. Lou, J. Jiang, L. Sheng, and G. Markou. Optics Express, vol. 34, no. 21, pp. 39011–39027, 2026.

A self-contained chain from elastic finite-element analysis to curved-layer Jones calculations and figure-ready outputs.

Download ZIP Python · 1.8 MB

Package details

In the package

Finite-element and optical solver
Separate figure generator
Computed CSV and JSON tables

Numerical analysis runs without an external mesh or measured stress field. Archived computed tables support figure generation.

Run from the extracted project root

python -m pip install -r requirements.txt
python scripts/main.py --self-test
python scripts/main.py --mode quick
python scripts/main.py --mode paper
python scripts/figures.py

SHA-256 5bb70f6fbd92ad2299b874a7d2ad136e8b3319f9e35c6cb498ffb9b04f08ef86

ReliabilityRESSUnder review

Nonhomogeneous hidden semi-Markov modeling for censored accelerated degradation tests of blue laser diodes

C. C. Zhang, X. Feng, and J. Jiang. Reliability Engineering & System Safety, submitted for publication, 2026.

Covariate-dependent semi-Markov estimation, state-age filtering, and remaining-life prediction with censored laser histories.

Download ZIP Python · 10.2 MB

Package details

In the package

Analysis and figure code
Model-generated pseudo-data
Computed results and plotting inputs

Pseudo-data are included for demonstration. Full measured-laser reruns require the original life-test records. See the archive README for the pseudo-data workflow.

Run from the extracted project root

python -m pip install -r scripts/requirements.txt
python -B scripts/main_analysis.py --quiet
python -B scripts/generate_figures.py

SHA-256 05f56a715a7ce7435db6a391527d3269bc43e2b03e3106439de93e40a72ee4ac

ReliabilityRESSPublished

Segmental Viterbi estimation using transition constraints for covariate-dependent hidden semi-Markov degradation models with right censoring

C. C. Zhang, X. Feng, and J. Jiang. Reliability Engineering & System Safety, vol. 278, Art. no. 113615, 2027.

Operational threshold constraints, exact segmental estimation, and filtering for ordered first-passage prediction.

Download ZIP Python · 4.3 MB

Package details

In the package

Analysis and independent figure code
Reproducible fitted-model pseudo-data
Public HUST battery inputs
Plotting coordinates and native SVGs

The website release includes illustrative pseudo-data, not original laser measurements. See the archive README for the reproducibility workflow.

Run from the extracted project root

python -m pip install -r requirements.txt
python scripts/generate_pseudo_data.py
python scripts/main_analysis.py smoke --output-dir outputs/example
python scripts/generate_figures.py

SHA-256 4a5003abfa592ade8e2d3166fb70c186688144f7b57292d45401a21e9d6cb633

Please cite the associated article when using a research code package.