Input variable acquisition method and system for building energy consumption hybrid model

Abstract

The invention provides an input variable acquisition method and system for a building energy consumption hybrid model. The method comprises the following steps: dividing an initial variable into a load related variable and an electromechanical related variable; the initial variables are sampled and subjected to energy consumption simulation, the influence of the initial variables on output variables is analyzed, and then key variables are extracted for the two related variables; and when the initial data set lacks historical energy consumption data, carrying out association rule mining of the key variables on the actual building data set after the key variables are complemented according to the historical energy consumption data, and speculating the missing key variables from the known key variables by utilizing the mined association rule, so as to obtain all key variable data needing to be input. According to the method, the related variables of the building load and the electromechanical system are comprehensively considered, so that the characteristics of the input variables are richer, and meanwhile, the extraction of the key variables and the speculation of the missing values of the key variables can still be carried out under the condition of lacking historical data.

Description

technical field [0001] The invention belongs to the technical field of building energy consumption prediction, and in particular relates to an input variable acquisition method and system for a building energy consumption hybrid model. Background technique [0002] Fewer input variables in the energy consumption model cannot fully reflect the impact on energy consumption, and too many input variables will lead to long model calculation time and increase the risk of overfitting. In theory, all the variables that need to be input to the energy consumption model have more or less influence on energy consumption, but the increase of input variables often leads to a sharp increase in the amount of computation required by the data-driven model. All relevant variables are used as The input of data-driven models is bound to create a "curse of dimensionality". The impact of each input variable on energy consumption is not the same, and the variable that will have a significant impac...

AI Technical Summary

Problems solved by technology

[0004] In view of this, the present invention aims to solve the problem that the existing input variable extraction method is not ideal for extracting energy consumption models and cannot obtain corresponding results when there is a lack of historical relevant data Or the problem of large error in the result

Images