Research On Multi-electricity Working State Identification Method Based On Data Mining

With the popularity of Internet of Things(Io T)technology,more and more households are using smart sockets.The use of smart sockets can promptly detect the power usage status of appliances,not only allowing remote monitoring of appliance safety,but also determining the living status of family members based on appliance usage,ensuring their living and electrical safety.This thesis uses smart sockets as a data collection device for power usage trends and studies the prediction of current data in the event of occasional missing data.Based on user current data,a genetic algorithm-based method for identifying household appliance operating status is proposed.Finally,a multi-appliance state recognition system is designed and implemented.Firstly,in the process of collecting user power usage data,there is a phenomenon of occasional missing data.To address the occasional missing data in the high-frequency current data collected during user power usage,this thesis proposes a method for estimating missing data using a backpropagation neural network.As the collected data allows for tolerance,this thesis proposes a performance evaluation index for the method of estimating missing data when tolerances exist.The experimental results show that this method can be used to estimate missing data in current data.Secondly,to obtain a multi-appliance state recognition model,a stable current sequence of the appliances needs to be obtained first.The method of dividing the current sequence segment is used to obtain the stable current sequence of the appliances during operation.Then,a multi-appliance state recognition model is constructed,and a genetic algorithm is used to estimate whether multiple appliances are currently in a stable state.The experimental results show that the multi-appliance state recognition model based on genetic algorithms has better performance,with a maximum accuracy rate of 97.24%,and can accurately and timely determine whether multiple appliances are in a stable operating state.Finally,based on the proposed method for estimating missing current and multiappliance state recognition model,a multi-appliance state online recognition system is designed and completed using current data from multiple household appliances as experimental data.The system consists of a smart socket,a local information computing node,and a database server,and uses wireless communication for information exchange.The smart socket completes the collection of appliance usage data.Secondly,by analyzing the current data,multi-appliance states can be quickly recognized and data can be visualized.The experimental results show that the multi-appliance state online recognition system can quickly recognize whether multiple appliances are currently in a stable working state.This thesis studies the multi-appliance state recognition method,establishing a correlation between household appliance states and family member behavior.Timely acquisition of the behavior of small-scale household users is important for focusing on the living and electrical safety of small-scale household members and for the construction of smart communities and intelligent community management.Figure [29] Table [15] Reference [54]...