A Vibration Signal Dataset for Lithology Identification

Fuente: Zenodo
Salvato in:
Dettagli Bibliografici
Autore principale: Liu, JunSong
Natura: Recurso digital
Pubblicazione: Zenodo 2026
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866901391137570816
author Liu, JunSong
author_facet Liu, JunSong
contents <p><strong>Description</strong></p> <p>This dataset is designed for lithology identification and is constructed based on vibration signals collected from a laboratory rock drilling test platform. The experiments were conducted using an ASK-101 drilling machine, and three-axis vibration acceleration signals generated during the drilling process were acquired using a CT1000SLEP vibration sensor.</p> <p>The dataset contains vibration signals from <strong>seven typical lithology classes</strong>. Under <strong>three different drilling conditions</strong>, <strong>15 groups of raw continuous vibration signals</strong> were collected for each lithology class. Each signal has a duration of approximately <strong>20 seconds</strong> and is sampled at <strong>51.2 kHz</strong>. All signals are three-axis measurements, capturing multi-directional dynamic characteristics during the drilling process.</p> <p>The dataset consists of two parts:</p> <ul> <li><strong>rawdata</strong>: raw sampled data stored in Excel format, with a clear structure for ease of understanding and analysis</li> <li><strong>dataset</strong>: a preprocessed subset derived from the raw data, which can be directly used for model training and experiments</li> </ul> <p>In addition, example code is provided to demonstrate basic data usage and processing procedures, enabling users to quickly get started.</p> <p><strong>Dataset Characteristics</strong></p> <ul> <li>Includes multiple lithology classes (7 classes), suitable for classification and identification tasks</li> <li>High sampling rate (51.2 kHz), preserving rich temporal dynamics</li> <li>Collected from real drilling experiments, with strong engineering relevance</li> <li>Suitable for research on measurement-while-drilling (MWD) lithology identification and cross-condition generalization</li> </ul> <p>Under varying drilling parameters (e.g., rotational speed and weight on bit), the distribution of vibration signals exhibits noticeable shifts. Therefore, this dataset can also be used to study <strong>distribution shift</strong> and the generalization capability of models under different operating conditions.</p> <p> </p>
format Recurso digital
id zenodo_https___doi_org_10_5281_zenodo_19702433
institution Zenodo
language
publishDate 2026
publisher Zenodo
record_format zenodo
spellingShingle A Vibration Signal Dataset for Lithology Identification
Liu, JunSong
<p><strong>Description</strong></p> <p>This dataset is designed for lithology identification and is constructed based on vibration signals collected from a laboratory rock drilling test platform. The experiments were conducted using an ASK-101 drilling machine, and three-axis vibration acceleration signals generated during the drilling process were acquired using a CT1000SLEP vibration sensor.</p> <p>The dataset contains vibration signals from <strong>seven typical lithology classes</strong>. Under <strong>three different drilling conditions</strong>, <strong>15 groups of raw continuous vibration signals</strong> were collected for each lithology class. Each signal has a duration of approximately <strong>20 seconds</strong> and is sampled at <strong>51.2 kHz</strong>. All signals are three-axis measurements, capturing multi-directional dynamic characteristics during the drilling process.</p> <p>The dataset consists of two parts:</p> <ul> <li><strong>rawdata</strong>: raw sampled data stored in Excel format, with a clear structure for ease of understanding and analysis</li> <li><strong>dataset</strong>: a preprocessed subset derived from the raw data, which can be directly used for model training and experiments</li> </ul> <p>In addition, example code is provided to demonstrate basic data usage and processing procedures, enabling users to quickly get started.</p> <p><strong>Dataset Characteristics</strong></p> <ul> <li>Includes multiple lithology classes (7 classes), suitable for classification and identification tasks</li> <li>High sampling rate (51.2 kHz), preserving rich temporal dynamics</li> <li>Collected from real drilling experiments, with strong engineering relevance</li> <li>Suitable for research on measurement-while-drilling (MWD) lithology identification and cross-condition generalization</li> </ul> <p>Under varying drilling parameters (e.g., rotational speed and weight on bit), the distribution of vibration signals exhibits noticeable shifts. Therefore, this dataset can also be used to study <strong>distribution shift</strong> and the generalization capability of models under different operating conditions.</p> <p> </p>
title A Vibration Signal Dataset for Lithology Identification
url https://doi.org/10.5281/zenodo.19702433