Intelligent Analysis And Decision Of Building Air Conditioning Based On Control Data

With the development of the national economy,many tall buildings have been built in various places in China in recent years.Most buildings use intelligent management to make it easier to regulate energy and meet indoor comfort.Nowadays,many buildings have produced a lot of control data.It is of great significance to analyze and excavate potential rules from mass building air conditioning data to save energy and predict energy.First of all,this paper explores the research of building air conditioning energy conservation at home and abroad,and then clean up and display the historical data of air conditioning control in some buildings.In order to predict the total energy needed for the day,three different methods are used to find the best reference day and its total energy.The first is to directly calculate the distance between the day's weather vector and the historical weather vector,and the minimum selection distance is the best reference day.Second is to cluster the historical control parameters first and then calculate the day's weather vector and each other.The distance of the historical weather vector in a class,the K reference day is obtained,the energy value is calculated synthetically;third is the clustering of the historical weather and the control parameters,then the distance between the day vector and the historical weather vector in each category is calculated,and the N reference days are obtained.Finally,the required energy value is obtained.In order to predict the value of the main control parameters of the air conditioning at the next time,the depth learning algorithm is adopted and the input data of the model are explored.The training and modeling of the historical weather,indoor temperature,machine room parameters,cluster data and label category data are used respectively.A multidimensional weighted vector representation is proposed for non numeric parameters(such as weather phenomenon and wind direction),and it is compared with multidimensional vector and single dimensional vector representation.The results show that the accuracy of multidimensional weighted vector representation can be used to improve the accuracy of the algorithm.Finally,an intelligent analysis and decision system for building air-conditioning control data is developed.