Batch Distillation Data for Developing Machine Learning Anomaly Detection Methods

TL;DR

This paper tackles the data bottleneck for ML-driven anomaly detection in chemical processing by introducing an open, multimodal database created from a laboratory-scale batch distillation plant. It combines time-series from conventional sensors/actuators with online NMR, audio, and video data, all richly annotated with anomaly ontologies and YAML-based metadata to support interpretable learning. The dataset encompasses 119 experiments across fault-free and anomalous conditions, with detailed metadata, uncertainty estimates, and structured storage to enable robust training, validation, and transferability studies. By providing a CC BY 4.0-labeled resource via Zenodo, the work aims to accelerate development of ML-based AD and anomaly mitigation in chemical engineering, while enabling future extensions to broader operating modes and multicomponent systems.

Abstract

Machine learning (ML) holds great potential to advance anomaly detection (AD) in chemical processes. However, the development of ML-based methods is hindered by the lack of openly available experimental data. To address this gap, we have set up a laboratory-scale batch distillation plant and operated it to generate an extensive experimental database, covering fault-free experiments and experiments in which anomalies were intentionally induced, for training advanced ML-based AD methods. In total, 119 experiments were conducted across a wide range of operating conditions and mixtures. Most experiments containing anomalies were paired with a corresponding fault-free one. The database that we provide here includes time-series data from numerous sensors and actuators, along with estimates of measurement uncertainty. In addition, unconventional data sources -- such as concentration profiles obtained via online benchtop NMR spectroscopy and video and audio recordings -- are provided. Extensive metadata and expert annotations of all experiments are included. The anomaly annotations are based on an ontology developed in this work. The data are organized in a structured database and made freely available via doi.org/10.5281/zenodo.17395544. This new database paves the way for the development of advanced ML-based AD methods. As it includes information on the causes of anomalies, it further enables the development of interpretable and explainable ML approaches, as well as methods for anomaly mitigation.

Figures (10)

• Figure 1: P&I diagram of the batch distillation plant. Color code of the lines: product (red), cooling water (dark blue), cooling ethanol (light blue), nitrogen (yellow), and pressure control (green). The equipment list, in which the labels are declared, is given in the Supporting Information.
• Figure 2: Photo of the laboratory batch distillation plant. The reboiler vessel (V001) and the distillation column (C001-C003) are surrounded by insulation. The visible glass apparatuses are the condenser (HE001-HE003), the buffer vessel (V002), and the distillate receiver (V003). The pumps (P701, P702) are blue, the condensation trap (HE004) has a yellow jacket and is connected to the gray vacuum pump (P301). The benchtop NMR spectrometer is on the top left rack.
• Figure 3: Snapshots from the cameras in the batch distillation plant. Left: Glass heat exchangers HE001 and HE003. Middle: Buffer vessel V002 with connections. Right: Distillate receiver V003.
• Figure 4: Temperatures and pressure recorded during experiment batch_dist_ternary_butan-1-ol+propan-2-ol+water/operating_point_001/test_abormal_experiment_001. The three phases, start-up, operation, and shut-down are indicated.
• Figure 5: Temperatures and pressure recorded during experiment batch_dist_ternary_butan-1-ol+propan-2-ol+water/operating_point_003/test_abormal_experiment_002, in which a perturbation was introduced at $t=69$ min and removed at $t=76$ min. The colors indicate the three phases of the response to the perturbation. The perturbation was a setpoint change of the pressure from $p=500$ mbar to $p=700$ mbar.