Dataset Construction and Verification Based on Non-Intrusive Load Monitoring

NILM(Non-intrusive load monitoring) is an efficient and low-cost power monitoring technology widely used to advance the development of intelligent electricity consumption and improve the development of the Internet of things. In order to improve the accuracy of load monitoring, this paper proposes an algorithm for constructing NILM datasets based on high-frequency sampling. The advantage of this algorithm is that it does not need to acquire the multi-load and reduces the workload of dataset construction from exponential level to linear level. Firstly, according to the electromagnetic induction principle of the transformer, the input side is connected to the load to capture the current signal and convert it into the voltage signal. The output side is connected to the oscilloscope for high-frequency sampling to obtain the current signal. Then, all the sampling data are synchronized based on the voltage phase. A standard dataset of 9 loads in 13 states and 8176 multi-load mixed data are constructed. Finally, the correlation coefficient and normalized mean square error are applied to verify some data. The results showed that the correlation coefficients were above 0.95, and the lowest value was 0.9689; The maximum normalized mean square error is 0.0422.