Energy consumption and the science behind the patterns in electricity expenditure have been rapidly growing fields in technology for decades owing to the vital nature of fuel. This dataset allows room for advancement in exploiting consumption data and generating smart results powered by state-of-the-art algorithms and sensors.

The line of experimentation for this problem statement is focused on enabling users to interpret their power consumption efficiently and interacts with live visual trends and limits their consumption to a customized estimate. Prediction models for forecasting consumption, analyzing power signature of high-load appliances, and monitoring of appliance power state can be built based on parameters such as instantaneous and cumulative energy, power, and solar energy.

Although the scope of this dataset is limited to individual accommodations, the purpose is to extrapolate the functionalities to an industrial scale, where the importance of restricted consumption, and protection against casualties is truly evident. This dataset can be used to improve the accuracy of time series forecasting of energy consumption.

Springer Research Paper - Email for copy. Sensor Setup Guide Old Jupyter Notebooks - API keys have been revoked.

You can find 3 files attached:

1. SME-akash-3P-1ms-influxdata_1jan2021-28feb2022.csv This is data taken from an independent house with 3 phase power line, with an on-grid solar power plant connected. The prominent high load appliances in use are ACs, Geysers (Water Heaters), Pumps, Washing Machines and Fridges.

2. SME-divya-3p-1ms-influxdata_1apr2021-31dec2021.csv This data is taken from an apartment with 3 phase power line only. High load appliances include ACs, Geysers, Washing Machines and a fridge.

3. SME-siva-1p-1ms-influxdata_1jan2021-31dec2022.csv This data is taken from an apartment with single phase power line only. High load appliances include ACs, Geysers, Washing Machines, Dryer and a fridge.

Since this data is taken as a CSV dump from InfluxDB, you can see familiar columns like _start, _stop, _time that indicate when the reading was taken. The _field column includes what parameter of power was recorded [i.e., current, voltage, energy, power, frequency, and power factor]. The _measurement column indicates the which phase these values belong to [i.e., Phase1, Phase2, Phase3 or solar]

Units of each _value:

• Energy: kWh
• Voltage: V
• Current: A
• Power: W
• Frequency: Hz
• Power Factor: (No Units)

In the Fig. for Geyser, the green line shows the resistive load being turned on (Water Heater). The same image also shows the Fridge compressor periodically turning on in blue colour.

In the Fig. for Air Conditioner, the orange line shows the load being turned on. Quadratic regression can be used to map the curve.

For cases involving correlation studies with regards to the solar data, the approximate location of the solar panels is East Bangalore, Karnataka, India. Azimuth = 16 degrees Declination = 30 degrees Peak Power of solar panel modules = 3300w Inverter max output = 3kW

Other use cases include Brownout detection (low voltage), trend analysis, predictions, overload detection, and much more.