Non-invasive electrical appliance analysis and identification device

Abstract:
This device develops a non-intrusive method to monitor the usage of home appliances which is based on STM32F407. It is mainly composed of power supply circuit, sampling circuit, signal conditioning circuit, A/D conversion circuit and microcontroller circuit. The current signal output by the power frequency current transformer is amplified by the INA188 module and sampled by the AD7606 module. Then it uses Fast Fourier Transform (FFT) algorithm to get the current waveform spectrum. The current signal harmonics are used as the optimal characteristic parameters for load identification, and the minimum Euclidean distance is used as the optimization index. By enumerating the combination of all electrical appliances, the similarity coefficient between the electrical appliances and a certain combination is calculated. From this, the type of electrical appliance is derived. The working current of the device is less than 15 mA and the response time is less than 2 s. In general, the recognition accuracy rate is about 95.5% after 600 tests.