PMU-Based Residential Appliance Dataset (USPCASE, NUST)
Description
The USPCASE–NUST Smart Grid and Power Lab Appliance Energy Dataset is a real-time electrical load monitoring dataset collected at the Smart Grid and Power Lab, USPCASE, National University of Sciences and Technology (NUST), Pakistan. The dataset was developed for research in Non-Intrusive Load Monitoring (NILM), smart grid analytics, appliance classification, load forecasting, and machine learning-based energy monitoring applications. The data was acquired using a PMU/PDC-based monitoring infrastructure integrated with NI LabVIEW software. It contains timestamped measurements of appliance-level electrical parameters recorded under realistic operating conditions. Dataset Summary Total Records: 50,577 Total Features: 6 Time Span: August 2024 – March 2025 Data Type: Time-series electrical measurements Environment: Real-time smart grid laboratory setup Features Date and Time – Timestamp of each measurement Aggregated Power – Total observed power Voltage – Operating voltage measurement Current – Current consumption Power – Real power consumed Appliance – Appliance label/category Appliances Included TV Microwave Oven Refrigerator Vacuum Cleaner 6 Ton AC Iron Electric Kettle Hair Dryer Toaster Washing Machine Coffee Machine Desktop Computer Research Applications This dataset is suitable for: NILM research Appliance load classification Smart home energy analytics Deep learning and machine learning applications Energy consumption prediction Time-series forecasting Load signature extraction Real-time anomaly detection Dataset Characteristics Real-time measurements Timestamped sequential data Multiple appliance categories Suitable for deep learning workflows Realistic appliance operational behavior Keywords Smart Grid, NILM, PMU, PDC, Appliance Classification, Energy Consumption, Deep Learning, Load Monitoring, Time-Series Data, USPCASE, NUST
Files
Steps to reproduce
The dataset was collected at the Smart Grid and Power Lab, US Pakistan Center for Advanced Studies in Energy, National University of Sciences and Technology. Data acquisition was performed using a PMU/PDC-based monitoring infrastructure integrated with NI LabVIEW software for real-time electrical measurements and appliance monitoring. The workflow used to generate the dataset is summarized below: 1. Multiple household and laboratory appliances were connected individually to the monitoring setup under controlled laboratory conditions. 2. Electrical parameters including voltage, current, real power, and aggregated power were continuously monitored and recorded in real time. 3. A Phasor Measurement Unit (PMU) was used to capture synchronized electrical measurements, while a Phasor Data Concentrator (PDC) collected and managed the streaming data. 4. NI LabVIEW software was used for real-time visualization, acquisition, synchronization, and storage of measurements. 5. Data collection was conducted from August 2024 to March 2025 under realistic appliance operating conditions. 6. The monitored appliances included TV, Microwave Oven, Refrigerator, Vacuum Cleaner, 6 Ton AC, Iron, Electric Kettle, Hair Dryer, Toaster, Washing Machine, Coffee Machine, and Desktop Computer. 7. The acquired measurements were exported into CSV format for preprocessing and analysis. 8. Timestamp synchronization, data cleaning, and preprocessing were performed to organize the dataset into structured time-series records suitable for machine learning and NILM applications. 9. The final dataset was prepared for applications including Non-Intrusive Load Monitoring (NILM), appliance classification, smart grid analytics, load forecasting, and machine learning/deep learning research. Software and Tools Used: NI LabVIEW PMU (Phasor Measurement Unit) PDC (Phasor Data Concentrator) CSV-based data acquisition and preprocessing workflows Python/TensorFlow-compatible time-series formatting for machine learning applications
Institutions
- National University of Sciences and TechnologyIslamabad, Islamabad