This dataset comprises 24 complete milling tool wear cycles collected from a SANJI VMC 650 three-axis vertical machining centre operating under active production conditions at a heavy-duty CNC manufacturing facility. Machining was performed on 38CR chromium alloy steel using Mitsubishi APMT1135PDER-M2 VP15TF PVD TiAlN-coated carbide inserts mounted on a 20 mm two-flute indexable end mill, at a constant spindle speed of 1800 RPM and feed rate of 1500 mm/min under flood coolant. A low-cost, self-contained data acquisition system built around an Arduino UNO microcontroller was deployed non-intrusively on the machine, logging six-axis inertial data (three-axis acceleration and three-axis angular rate) from an MPU6050 GY-521 IMU affixed to the spindle head and spindle motor current from an SCT-013-050 split-core current transformer. Each tool wear cycle begins with the installation of fresh carbide inserts and terminates upon operator-declared failure, identified by deteriorated surface finish, abnormal chatter, and elevated vibration. Remaining Useful Life (RUL) labels are assigned using a timestamp-based linear decay, requiring no production stoppage or optical measurement equipment. The dataset is intended for benchmarking machine learning and deep learning models for CNC tool RUL prediction under real industrial operating conditions.
The CLAPE (Contextual Learner Attributes and Pupil Engagement) dataset is a validated multimodal educational dataset designed to support research on physiological student engagement using low-cost, non-invasive RGB webcam technology. The dataset integrates physiological pupil var…
The rapid proliferation of wireless communication devices has resulted in increasing radio-frequency (RF) activity within the 2.4 GHz Industrial, Scientific, and Medical (ISM) band. Wireless technologies such as Wi-Fi, Bluetooth, ZigBee, and numerous Internet of Things (IoT) devi…
A Single-Token Sensor Substrate for Industrial Condition Monitoring: Fault Classification at Classifier Parity, Free Anomaly and Remaining-Useful-Life Signals, and Multi-Task Decoding at 2,000–4,000× Compression Randolph James Ferlic, M.D., and Kimberly Kate Ferlic · Fieldstone A…
Self-supervised deep learning has emerged as a powerful method for image enhancement when a priori ground-truth references are not available. Stemming from Noise2Noise , it was shown that a convolutional neural network (CNN) can be trained from a noisy input and target pair of th…
Replication package for the study "Does Sentiment Transfer? Label Granularity and Cross-Topic Generalisation in YouTube Comment Classification". The study evaluates how much sentiment-classification accuracy survives when a model is applied to a topic it was not trained on, and h…
CleanCam is a benchmark dataset for underwater camera-viewport fouling severity assessment in aquaculture. It distinguishes material attached to the camera viewport from water-column degradation, including turbidity, haze, suspended particles, lighting variation, and low contrast…