Data for: Simulator-assisted deep learning framework for anomaly detection in continuous solvent extraction for reliable metal recovery in lithium-ion battery recycling

This dataset provides simulator-generated pH trajectories used for developing and evaluating deep-learning-based anomaly detection models in a continuous solvent extraction process for lithium-ion battery recycling. The dataset consists of ten normal-operation trajectories generated under volumetric-flow perturbations and twenty fault-operation trajectories generated under predefined process disturbances. The fault trajectories include step and ramp disturbances introduced at specified process sections after 700 min of normal operation. Each trajectory contains time-series pH data from seven settler outlets sampled at 12 sec intervals. The provided data correspond to simulator outputs before artificial pH measurement noise was added. Measurement-noise injection, normalization, model training, and evaluation were performed in Python and are described in the associated code repository.