Research On Technology Of Bath Time Of Water Heater Forecasting Based On SVM And Deep Learning

The realization of intelligent control is the main development direction of the storage electric water heater.It is the key to realize the true intelligence of the storage electric water heater based on the research of the user behavior.The intelligent control technology based on user behavior overcomes the long heating cycle,large energy consumption,and the inconvenience of frequently setting.Therefore,it has certain practical significance for improving the quality of life of the user,and has important reference significance for energy saving and comfort Type intelligent home appliances.Most of the work style of water storage type electric water heater uses a circular heating and keeping warm mode.The water heating process has a great lag,and the insulation process also means a lot of energy consumption.Therefore,it is gradually attracted attention to pay attention to the intelligent control of storage electric water heater.The user’s behavior data as a time series to the user to use the historical behavior of electric water heater data based on the use of machine learning algorithm for mining historical data into the user’s future behavior prediction,which can predict the value of electricity water heater for heating in advance,and finally has a very good realization of the storage-type electric water heater intelligent control.The main purpose of this paper is to predict the behaviors of different behavioral habits and find a relatively good forecasting method.Firstly,this paper summarized the research status of the machine learning algorithm,and analyzed the behavior characteristics of the user using the storage electric water heater,and compared the advantages and disadvantages of the different machine learning prediction models,and studied the user behavior law.Moreover,according to the behavior of users of the strength of the law,the behavior of using water was randomly classified,and generates the corresponding level of user behavior data.Secondly,the method of user behavior prediction is studied by using time series prediction based on ARIMA model,model prediction based on support vector machine(SVM),time series prediction based on depth learning.We also compared and analyzed the prediction results of these three methods.The genetic algorithm and particle swarm optimization(PSO)are used to optimize the parameters of the SVM model when using the SVM to forecast the user’s bathing time.When predicting with using depth learning,the depth of learning has been classified with water-free binary at first,and then the water sequence was predicted by depth learning,and a better prediction result was obtained.Finally,the water heater data forecasting system is designed and implemented on the basis of WINFORM technology.The system includes data management,data forecasting,data analysis,image display and system setting module,and can completely realize the technical route of the remaining water bath time prediction.