Research On Key Technologies Of Multi-parameter Detection And Feature Extraction For Washing Machine

In recent years,a large number of manufacturing industries have experienced the phenomenon of “labor shortage”,as well as the washing machine industry.Most of the washing machine detection equipment only has few test function,and the automation level is not enough,which requires a lot of manual testing and analysis.It is urgent to design a large number of advanced multi-parameter automatic detection system to improve production efficiency.To solve the above problems,the multi-parameter detection technology of the washing machine was researched in this dissertation,focusing on the data acquisition and feature extraction of the power,sound and vibration parameters of the washing machine.The main contents are as follows:(1)A multi-parameter data acquisition system for washing machine was designed.The system contains sensor selection,pre-processing circuit design,power module,AD acquisition module and data acquisition and storage platform.All of the system modules are debugged to ensure that the system can work normally.Besides,the power parameters and vibration parameters were calibrated by using relatively standard measuring instruments.There are two kinds of system data acquisition platforms,one is general purposed auto-test system platform of our laboratory's,which is convenient for the data collection and feature extraction in the first instance.The data acquisition and storage platform based on ARM single chip is designed in the later stage of the system.Using timer-control data acquisition and DMA to transfer data to SD card.The experiment shows that the system can collect and store data without loss.It can lay the foundation for realizing multi-parameter automatic detection system of portable washing machine.(2)The method of feature extraction of vibration signal in the washing stage of washing machine was researched mainly in the dissertation.The feature extraction mainly includes automatic segmentation and spectrum analysis.The automatic segmentation is based on the washing period of time domain waveform characteristics of vibration signals.The characteristics of the domain waveform are calculated by filtering and framing calculation energy,and obtaining the position of the minimum value of the energy.Finally,the vibration signal of the complete washing segment is divided into small vibration period signals,and then the vibration periodic signal is analyzed by spectrum,the spectrum analysis is divided into wide spectrum analysis and narrow window spectrum analysis.Spectral analysis and narrow window spectrum analysis obtained the spectral main component of the vibration signal of the washing machine in the washing stage.The spectral principal component of the vibration signal of the washing machine at the washing stage is obtained.The experiments show that the main spectral components of the signal are located in the frequency range of 20Hz? 60Hz?110Hz~120Hz and 160Hz~170Hz.At the same time,the DTW algorithm is used to match the three sub-stages of the complete washing stage.The experimental results can identify different washing sub-stages well.(3)This dissertation also analyzes the sound signal of the washing machine during operation,and uses the method of voice processing to refine the original sound signal by means of the Spectrum map for the complete stage,the washing stage and the dehydration stage of the washing machine.Analyzing the main frequency range of the washing machine sound signal in different stages.It will help to recognise the working state of the washing machine through sound signal in the future.The series of work done in this dissertation has certain practical value for the washing machine detection industry.It provides a reference for detection and signal processing methods for automatic detection and fault diagnosis of low-voltage electrical appliances.This subject has a very broad application prospect and is worthy of further research.