Abstract: In order to achieve intelligent detection of the ambient voltage of an air purifier, a BP neural network model based on the temperature rise curve to predict the ambient voltage is constructed by using the sigmoid activation function. The results showed that when there were 27~53 neurons in the input layer and 25 neurons in the implicit layer, the prediction ability of the ambient voltage was better. The correlation coefficient between the actual value of the environmental voltage and the predicted value of the training set is 0.996, the model error range is -2.24%~5.26%, the relative error of the evaluation of the new data is small, and it has good predictability, and the results show that the voltage identification model can be used to intelligently predict the ambient voltage.

Key words: Air cleaner, Neural network, Machine learning, Voltage, Forecast