Non-intrusive load identification

What is it about?

Non-intrusive load identification is important to load monitoring and smart power utilization. We completely obtain the individual load waveform from the mixed signal. Then, the complete load decomposition waveform directly is identified by constructing a current characteristic filter group, which realizes load identification without the complicated load feature extraction. We have used the collected data to prove that the algorithm is effective.

Featured Image

Why is it important?

At present, in the view of signal separation, the research results of non-intrusive load identification algorithm are relatively few, and there is a lack of complete and comprehensive decomposition methods for independent load, which makes that users cannot get the integrity of power information. One of our contributions is through using the sparseness of current frequency domain signal to obtain an optimal solution. By using this property, we can transform the underdetermined problem into an optimization seeking problem, which can completely separate independent load waveform from mixed signals. Another contribution is to identify the load of decomposed complete waveform through the characteristic filter only by constructing the current characteristic filter set but without extracting load features. The identification process is simple and accurate. Meanwhile, the hardware implementation problem of pattern recognition algorithm in practical application is solved to some degree.

Perspectives