Research On Several Key Technologies Of Household Appliances Condition Monitoring

With the development of network technology, communication technology and computer technology in21st century, network household appliances have been well-known by more and more people. It is the future of household appliances. Now, the main advantage of the network household appliances are the remote communication and remote control, the PHM (Prognostic and Health Management) technology on them hasn’t be considered.In addition, the appliance manufacturers lack the comprehensive function test methods, they test the appliances on many single function items only. To solve the above problems, the research on several key technologies of household appliances condition monitoring was proposed.The main work and achievements of this dissertation are as follows:Firstly, some important theories were introduced. For example, the time domain and frequency domain signal processing, such as signal characters’parameters, FFT(Fast Fourier Transformation), GABOR transform; the pattern recognition, such as the concept of a class, feature extraction and selection, features optimization method, relevant function matching method, etc.Secondly, two data acquisition methods are described in detail, one is based on the general-purposed testing system platform, and the other is based on the TMS320F2812DSP platform. Sampling method, data accuracy problems were discussed. A profound research on high-accuracy measuring technology of frequency in synchronous collection was done based on the study on two acquisition ways of DSP. Further more a large number of voltage and current waves of appliances were collected by general-purposed testing system platform, which contains air conditioners, microwave ovens, refrigerators, TV sets.Thirdly, it focused on the recognition technology of the work state of appliances, and the microwave oven is chosen as the main research object especially, whose voltages and currents signal were selected to extract the feature. The following scientific issues were studied deeply:off-line manual classification and sample data selection, feature extraction and selection, features optimization, class modification and standard sample category library building, on-line automatic stage algorithm, correlation matching, etc. The related principle of key algorithms are expounded, experiment and data analysis process have been elaborated. It also discussed the transplant work by DSP such as FFT and related correlation algorithm. The end, this dissertation is discussing the method of prediction by using ARMA model which basing on time series technology. Through the ARMA model building, we can obtain the expert knowledge experience from home electrical appliance working state and working sequence. And the state of the operation of home appliance and remaining life will be predicted by using the ARMA model.Overall, the dissertation researched the pattern recognition technology of network household appliance health monitoring, and proposed the method building standard sample category library, which is based on off-line manual classification. The sample selection, feature selection and optimization process have been realized. The manual of classification has been proved correct and effective through clustering analysis.The result of experiments shows the ratio of pattern recognition reached97.67%with the feature combination including current harmonic characteristic in the home appliance operation status identification.The work of this dissertation especially on home appliance pattern recognition can offer reference for researchers who are engaged in intelligent detecting and fault diagnosis prediction of home appliances. In addition, the primary data sample library of all kinds of home appliance based on the general-purposed testing system platform can provide basic data for them, and has significantly practical value.