Developing an artificial neural network model to analyze emission spectra of high-frequency electrodeless lamps

Developing an artificial neural network model to analyze emission spectra of high-frequency electrodeless lamps Version 1 Description The Data Management Plan (DMP) is developed in the framework of the project 1.1.1.9 Research application No 1.1.1.9/LZP/1/24/023 of the Activity "Post-doctoral Research" " Developing an artificial neural network model to analyze emission spectra of highfrequency electrodeless lamps ". Project manager: Ph. D. Natalja Zorina. Project implementation time: 01.03.2025 - 29.02.2028. Total project funding: 184 140 EUR (incl. ERDF 156 519 EUR). Summary of the postdoctoral research grant project: High-frequency electrodeless discharge lamps (HFEDLs) combine compact design, long lifetime, and stable narrowband emission (FWHM 0.03–0.12 cm 1−), making them a competitive alternative to hollow cathode lamps in atomic absorption spectrometry. Operating with heavy metal vapors (e.g., mercury, thallium) and rare gases (e.g., argon, xenon) as buffer gases, HFEDLs generate low-pressure discharges sustained by highfrequency fields. Their high intensity and reduced contamination improve detection limits, but the resulting spectra often contain multiple closely spaced and overlapping lines from different elements. This complicates manual interpretation and underscores the need for accurate and efficient spectral analysis to optimize lamp performance and ensure reproducible results in applications such as atomic absorption spectroscopy, plasma diagnostics, and environmental monitoring. The research aims to develop an artificial neural network (ANN) model for analyzing the emission spectra of high-frequency electrodeless lamps (HFEDL). The model will be trained to identify characteristic patterns in spectral data and explore their connection to relevant lamp parameters and operating conditions within a defined frequency range. LICENSE:CC-BY-4.0 DOI: - 13/10/2025 Funder Grant Latvian Council of Science||LCS Developing an artificial neural network model to analyze emission spectra of highfrequency electrodeless lamps(1.1.1.9 Research application No 1.1.1.9/LZP/1/24/023 of the Activity "Post-doctoral Research") Researchers Natalja Zorina (0000-0001-9065-155X) Organizations University of Latvia Data Management Plan | Developing an artificial neural network model to analyze emission spectra of high-frequency electrodeless lamps LICENSE:CC-BY-4.0 DOI: - 13/10/2025 1. Main Info Title of DMP: Developing an artificial neural network model to analyze emission spectra of high-frequency electrodeless lamps Description: The Data Management Plan (DMP) is developed in the framework of the project 1.1.1.9 Research application No 1.1.1.9/LZP/1/24/023 of the Activity "Post-doctoral Research" " Developing an artificial neural network model to analyze emission spectra of highfrequency electrodeless lamps ". Project manager: Ph. D. Natalja Zorina. Project implementation time: 01.03.2025 - 29.02.2028. Total project funding: 184 140 EUR (incl. ERDF 156 519 EUR). Summary of the postdoctoral research grant project: High-frequency electrodeless discharge lamps (HFEDLs) combine compact design, long lifetime, and stable narrowband emission (FWHM 0.03–0.12 cm 1−), making them a competitive alternative to hollow cathode lamps in atomic absorption spectrometry. Operating with heavy metal vapors (e.g., mercury, thallium) and rare gases (e.g., argon, xenon) as buffer gases, HFEDLs generate low-pressure discharges sustained by highfrequency fields. Their high intensity and reduced contamination improve detection limits, but the resulting spectra often contain multiple closely spaced and overlapping lines from different elements. This complicates manual interpretation and underscores the need for accurate and efficient spectral analysis to optimize lamp performance and ensure reproducible results in applications such as atomic absorption spectroscopy, plasma diagnostics, and environmental monitoring. The research aims to develop an artificial neural network (ANN) model for analyzing the emission spectra of high-frequency electrodeless lamps (HFEDL). The model will be trained to identify characteristic patterns in spectral data and explore their connection to relevant lamp parameters and operating conditions within a defined frequency range. Researchers: Natalja Zorina (0000-0001-9065-155X) Organizations: University of Latvia Contact: Zorina Natalja ([email protected]) Data Management Plan | Developing an artificial neural network model to analyze emission spectra of high-frequency electrodeless lamps LICENSE:CC-BY-4.0 DOI: - 13/10/2025 2. Funding Funding organizations: Latvian Council of Science||LCS Grants: Developing an artificial neural network model to analyze emission spectra of highfrequency electrodeless lamps(1.1.1.9 Research application No 1.1.1.9/LZP/1/24/023 of the Activity "Post-doctoral Research") Project: 3. License License: CC-BY-4.0 Access Rights: Public Publication Date: 2025-10-17 4. Templates Descriptions Neural network model, processed and synthetic spectral datasets, and educational application for emission spectra analysis of highfrequency electrodeless lamps (HFEDL) This dataset includes synthetic and processed emission spectra, preprocessing scripts, trained neural network models, and an interactive educational program developed within the project “Developing an artificial neural network model to analyze emission spectra of high-frequency electrodeless lamps.” It provides training scripts, evaluation results, model files, and comprehensive data descriptions together with technological instructions explaining data preparation, preprocessing workflow, neural network training procedures, and usage guidelines for the developed tools and software. The educational application visualizes atomic emission spectra and allows users to explore and identify elements based on their characteristic lines. Template: LCS FARP Data Management Plan | Developing an artificial neural network model to analyze emission spectra of high-frequency electrodeless lamps LICENSE:CC-BY-4.0 DOI: - 13/10/2025 Type: Dataset 1 Data Summary 1.1 Data Summary 1.1.1 What is the purpose of the data collection/generation? The purpose of the data collection and generation is to obtain experimental emission spectra of high-frequency electrodeless lamps (HFEDLs) under different operating conditions and to create synthetic spectral datasets. These data will be used to train and validate artificial neural network (ANN) models for automated spectral analysis. The documentation and metadata will ensure that the datasets are well-documented, accessible, and suitable for future use. 1.1.2 Type of data generated/collected • Programme source code • Machine-readable text • Experimental data • Textual data • Process produced data 1.1.3 Format of data generated/collected • TXT • PDF • CSV • XLS • JSON • Others zip, exe, py Data Management Plan | Developing an artificial neural network model to analyze emission spectra of high-frequency electrodeless lamps LICENSE:CC-BY-4.0 DOI: - 13/10/2025 1.1.4 Expected size of the data – give expected size and choose unit of measurement 10 GB 1.1.5 Are you re-using this data set? No 1.1.6 If No, please describe, if you have considered re-using any of existing data? Existing available spectral datasets were considered, but they do not cover the specific experimental conditions and configurations of HFEDLs relevant to this project. Therefore, new experimental and synthetic data will be generated to ensure consistency and suitability for training and validating the machine learning models. 1.1.8 To whom the data might be useful ("data utility")? Other relevant remarks) The data will be useful for researchers in plasma diagnostics and spectroscopy, particularly those working with electrodeless discharge lamps, atomic absorption spectrometry, and spectral analysis. It may also benefit the wider scientific community interested in machine learning applications in physics, as well as educators and students seeking reference datasets for training and demonstration purposes. 2 Research Data Management 2.1 Metadata and documentation 2.1.1 Will data be attributed with standard identification mechanism (e.g. persistent and unique identifiers such as Digital Object Identifiers – DOI)? Yes DOI 2.1.2 Do you plan to provide metadata for your data? Yes 2.1.3 Will you use any metadata standard? Data Management Plan | Developing an artificial neural network model to analyze emission spectra of high-frequency electrodeless lamps LICENSE:CC-BY-4.0 DOI: - 13/10/2025 Yes Dublin Core 2.1.4 Will search keywords be provided that optimize possibilities for re-use? Yes 2.1.5 Will you follow any naming conventions for keywords? No 2.1.6 Will you provide clear version numbers? Yes 2.2 Making data openly accessible 2.2.1 What type of access the data generated/produced will have (choose one)? open (anyone is able to access data without restrictions) 2.2.2 Will you apply embargo period to access for your data? Yes Comment: An embargo period will be applied until the end of the project or acceptance of related publications. After this period, selected datasets will be made openly accessible. 2.2.3 Will data, associated metadata, documentation and code be made accessible with means of a repository? Yes DataverseLV https://dv.dataverse.lv/dataverse/lu 2.2.4 What methods or software tools are needed to access the data? Data will be provided in standard open formats (TXT, CSV, XLS, JSON, PDF) that can be accessed with commonly available tools such as spreadsheet programs (e.g., Microsoft Excel, LibreOffice Calc), text editors, and standard data analysis Data Management Plan | Developing an artificial neural network model to analyze emission spectra of high-frequency electrodeless lamps LICENSE:CC-BY-4.0 DOI: - 13/10/2025 environments (e.g., Python, R). Additional materials, such as compressed archives (ZIP) and executable files (EXE), may be provided for demonstration purposes. Python scripts (.py) can be accessed and run using any standard Python distribution. No specialized or proprietary software will be required beyond these widely available tools. 2.2.5 Will documentation about the software needed to access the data included? No 2.2.6 Is it possible to include the relevant software (e.g. in open source code)? No 2.3 Making data interoperable 2.3.1 Are the formats open to software applications and to recombination with different data sets? Yes Data will be provided in widely used, machine-readable formats such as TXT, CSV, JSON, and PDF. These formats are compatible with common open-source and commercial software (e.g., Python, MATLAB, Excel, and standard text editors), which ensures accessibility and facilitates reuse or integration with other datasets. 2.3.2 Will you use data standard vocabulary/taxonomy to make your data interoperable? No 2.4 Increase data re-use 2.4.1 How will the data be licensed to permit the widest re-use possible? Indicate the license planned to use Creative Commons Attribution 4.0 Processed spectral data, synthetic datasets, and documentation will be released under the Creative Commons Attribution 4.0 International (CC BY 4.0) license. This license allows users to copy, redistribute, and adapt the materials provided that proper credit is given to the original authors. Data Management Plan | Developing an artificial neural network model to analyze emission spectra of high-frequency electrodeless lamps LICENSE:CC-BY-4.0 DOI: - 13/10/2025 2.4.2 Are the data usable by third parties, in particular after the end of the project? Yes 2.4.3 When will the data be made available for re-use? 2028-02-29 2.4.5 Are data quality assurance processes provided? Yes 2.4.6 Other relevant remarks Quality assurance procedures will include validation of data consistency, verification of file integrity, and cross-checking of metadata accuracy. Experimental and synthetic datasets will be reviewed before publication to ensure completeness and compliance with repository standards. 3 Resources and Security 3.1 Allocation of resources 3.1.1 How will costs be covered for making data FAIR during project realization? Euro The costs associated with data management, documentation, and repository submission will be covered within the project budget under the dissemination and open science activities. 3.1.2 Who will be responsible for data management in your project? Natalja Zorina (0000-0001-9065-155X) The Project Leader, Natalja Zorina, will be responsible for data management, ensuring proper documentation, version control, and compliance with FAIR and institutional data management policies. The Project Leader will also supervise data preparation for repository submission and coordinate data sharing and access procedures. 3.1.3 How will costs be covered for making data FAIR after project realization? Euro Data Management Plan | Developing an artificial neural network model to analyze emission spectra of high-frequency electrodeless lamps LICENSE:CC-BY-4.0 DOI: - 13/10/2025 preparation for repository submission and coordinate data sharing and access procedures. 3.1.3 How will costs be covered for making data FAIR after project realization? Euro 3.2 Data security 3.2.1 What security measures will be used for data security? • Firewall • Passwords • Physical access control Data security will be ensured through password-protected access, firewall protection, and institutional backup systems. Version control will be implemented to monitor data integrity. Physical access to servers is restricted to authorized staff. 3.3 Ethical aspects 3.3.1 Are there any ethical or legal issues that could have an impact on data collection and sharing? No 3.3.2 Have you got/will you get permission from ethics committee to collect and process data (if applicable)? No 3.3.3 Is informed consent for data sharing and long term preservation included in questionnaires dealing with personal data, if applicable? No 3.3.4 Will data collected/generated include personal data/ sensitive information? No Powered by Data Management Plan | Developing an artificial neural network model to analyze emission spectra of high-frequency electrodeless lamps LICENSE:CC-BY-4.0 DOI: - 13/10/2025 Data Management Plan | Developing an artificial neural network model to analyze emission spectra of high-frequency electrodeless lamps LICENSE:CC-BY-4.0 DOI: - 13/10/2025