Dark Mode
Industrial IoT Severity and Status Records
Data Science and Analytics
Tags and Keywords
Trusted By
"No reviews yet"
Free
About
Spanning from January 2018 to June 2024, this collection captures detailed industrial alarm events essential for operational oversight and reliability engineering. It facilitates high-resolution time series analysis, machine learning model training for predictive maintenance, and anomaly detection within industrial processes. The records include critical attributes such as severity levels, operational states, and specific error messages like "Device Offline", enabling granular investigation into equipment behaviour and failure patterns.
Columns
- DateTime: The specific timestamp recording when the alarm event occurred.
- ProcessID: A unique identifier for the specific process that generated the alarm.
- AssetID: The identifier for the asset involved in the event (e.g., AAA-BMS-SSIF).
- AlarmSeverityName: The classification of the alarm's severity (High, Medium, or Low).
- State: The transition state of the alarm, such as moving from Abnormal to Normal (A2N).
- TransactionMessage: A detailed text description of the alarm event (e.g., Device Offline, Temperature above Set Point).
- Stage: The status of the alarm within its lifecycle (e.g., Cancelled, Cleared).
- AlarmClassName: The category or class of the alarm (e.g., General-ELV).
- Year: The year in which the event took place.
- Month: The month of the event (1-12).
- Day: The day of the month the event occurred.
- DayOfWeek: The specific day of the week (e.g., Sunday).
- Season: The season during which the alarm occurred (Summer or Winter).
- Hour: The hour of the day the event was logged (0-23).
- ProcessedMessage: A pre-processed version of the transaction message optimised for text analysis.
Distribution
- Format: Tabular CSV (
preprocessed_trendedpointalarm 1.csv). - Size: Approximately 102,000 valid records.
- Structure: 15 columns comprising timestamped, categorical, and textual data.
Usage
- Predictive Maintenance: Developing models to forecast equipment failures before they occur.
- Anomaly Detection: Identifying irregular patterns in asset behaviour or process outliers.
- Root Cause Analysis: Investigating specific failure modes, such as the prevalence of "Device Offline" messages.
- Operational Scheduling: Analysing seasonal or hourly alarm peaks to optimise maintenance shifts.
Coverage
- Time Range: The dataset primarily covers the period from January 2018 to June 2024, with some trace records dating back to 2015.
- Scope: Focuses on industrial alarm monitoring, with a significant concentration of data in the "Summer" season (72%) and "High" severity alarms (63%).
- Demographic/Asset: Includes identifiers for various assets and processes, with the most common asset identifier being "AAA-BMS-SSIF".
License
CC0: Public Domain
Who Can Use It
- Data Scientists: For training machine learning models on time-series data.
- Reliability Engineers: To analyse alarm fatigue and system stability.
- Plant Managers: For exploratory data analysis to improve operational efficiency.
Dataset Name Suggestions
- Industrial Alarm Events Time Series
- Predictive Maintenance Alarm Logs (2018-2024)
- Machine Anomaly and Failure Events
- Industrial IoT Severity and Status Records
Attributes
Original Data Source: Industrial IoT Severity and Status Records
QUALITYLoading...