Labelled dataset for Ultra-Low Temperature Freezer to aid dynamic modelling & fault detection and diagnostics

Area of Science:

• Engineering
• Computer Science
• Energy Management

Background:

• Ultra-low temperature (ULT) freezers are vital for storing biological samples but are highly energy-intensive.
• Efficient operation requires data-driven fault detection, diagnostics, and energy optimization.

Purpose of the Study:

• To present a novel, labelled, long-term performance dataset from 53 ULT freezers.
• To facilitate the development of advanced data-driven algorithms for ULT freezer operation.

Main Methods:

• Collected high-resolution historical data (up to 10 years) from 53 ULT freezers with two control strategies.
• Recorded over 10 attributes from critical locations, including chamber and refrigeration systems.
• Labeled data with regular events (e.g., door openings) and fault events from service reports.
• Developed a scalable ETL data pipeline for data preparation.

Main Results:

• A comprehensive, labelled dataset of ULT freezer performance spanning up to a decade.
• The dataset includes detailed operational parameters and event/fault labels.
• Established a robust data pipeline for processing and analyzing ULT freezer data.

Conclusions:

• The presented dataset is the first of its kind, enabling research into data-driven models for ULT freezers.
• It supports advancements in intelligent digital operations, fault detection, and energy optimization.
• The dataset is a valuable resource for improving the efficiency and reliability of critical laboratory equipment.