Dataset for Intrusion Device Detection in Fieldbus Network

Citation Author(s):
Xi'an Jiaotong University
Submitted by:
Ting Liu
Last updated:
Mon, 06/17/2024 - 00:47
Data Format:
0 ratings - Please login to submit your rating.


The dataset is derived from sampling Modbus commands in RS485 fieldbus networks, where the communication baud rate is 9.6k bit/s and the sampling frequency is set to 100 KS/s. In the dataset, 2000 signals are stored as normal data and abnormal data based on different intrusion devices in .xlsx format, respectively, with a total size of 170 MB. In this dataset, the attack scenario is the unauthorized insertion of a silent intrusion device into the fieldbus, thereby eavesdropping communication data. The data in the DifferentIntrusion_data folder is sampled from a real distribution cabinet under a constant temperature of 21℃. The data in the DifferentTemperature_data folder is sampled from an RS485 testbed inside the temperature-controlled box, where data sampling is conducted every 3℃ from 0℃ to 54℃.


The data in both folders is divided into BV signal and CV signal, where the BV signal is the fail-safe biasing voltage signal and the CV signal is the command voltage signal from the client. The data in the Normal folder is sampled when there are no external intrusion devices in the fieldbus network. In contrast, the data in the Abnormal folder is sampled when intrusion devices made by ZLAN, MOXA, and KonNaD are attached to the fieldbus network, respectively. You can load the '.xlsx' file data using Python and MATLAB according to the file.