| Energy consumption has been a hot topic of enduring popularity,and the annual energy consumption of buildings around the world is gradually increasing.Today,domestic urbanization has significantly accelerated building energy consumption.One of the main energy-consuming systems in buildings is the central air-conditioning system.The cloud storage platform that appeared in the early 21 st century enabled central airconditioning to accumulate a large amount of energy consumption data in the operation,and stored it in the Oracle database.The amount of data has reached tens of millions,laying a foundation for data mining.Traditional data analysis techniques include query,screening,induction,statistics,etc.,often analyzing the mathematical characteristics of the data itself,and it is difficult to find hidden information.Aiming at the problem that the central air-conditioning data has a large amount of data and it is difficult to find hidden information,the use of data mining technology is one of the important ways to study the optimization problem of central air-conditioning system energy-saving strategy.This topic takes the central air-conditioning system of a shopping mall as the research object,collects the system data through the sensor and stores it in the Oracle database,pre-processes the collected data to obtain the operating data under stable conditions,and analyzes the energy consumption.And energy consumption feature selection.The Boruta feature selection algorithm is used to select the appropriate subset of energy consumption characteristics from the original energy consumption variables,namely load rate,host water temperature,chilled water pump frequency and cooling water pump frequency.The energy consumption feature subset is used as the post-data mining model and energy saving.The object of strategy optimization.According to the stable operation data of the pre-processed central air-conditioning system,two data mining models of the central air-conditioning system are established,which are the central air-conditioning BP neural network model and the central airconditioning association rule model,which can effectively solve the central airconditioning system.The problem of difficult mechanism modeling and strong system coupling.Based on two data mining models of central air conditioning system,an improved genetic algorithm optimization strategy is proposed.This paper is called hybrid genetic association algorithm.The hybrid genetic association algorithm optimizes the variables to select the subset of energy consumption features,optimizes the target to select the energy consumption ratio EER of the central air conditioning system,and finally uses the original data to verify the feasibility of the hybrid genetic association algorithm.In this paper,two representative operating conditions of load rate of 75% and 90% are selected.Genetic algorithm and hybrid genetic association algorithm are used to optimize the subset of energy consumption characteristics under two operating conditions to find the best under this condition.Run the strategy.Comparing the optimization strategies of the two algorithms,the average genetic algebra of the genetic algorithm is 1.6~2 times that of the hybrid genetic association algorithm,and the energy consumption ratio obtained by the hybrid genetic association algorithm is increased by 4.15%~8.40%.Finally,the polynomial fitting of each working condition optimization value is carried out.After the analysis,the trend of the fitting curve can be consistent with the central airconditioning energy-saving operation.The fitting curve analysis result can provide reference for the central air-conditioning operation strategy.In summary,the data mining technology and genetic algorithm adopted in this paper provide an effective path for air conditioning operation strategy. |
