Building energy consumption prediction method

• Main Authors: ZHOU LIANKE, HE MING, LAO ZHENYU, WANG NIANBIN, WANG YONG, WANG HONGBIN
• Format: Patent
• Language: Chinese; English
• Published: 12.04.2022
• Subjects: CONTROL OR REGULATING SYSTEMS IN GENERAL; CONTROLLING; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS ORELEMENTS; PHYSICS; REGULATING

A building energy consumption prediction method belongs to the field of energy consumption prediction. The invention aims to solve the problem that a better prediction effect is difficult to obtain when an existing neural network is used for predicting energy consumption. The method comprises the following steps: firstly, acquiring historical data of a predicted energy consumption index, and constructing a feature tensor X = [Xt, Xt-1,..., Xt-n], wherein Xt-i represents a feature vector at a t-i moment, and t represents a current moment; elements in the feature vectors are predicted energy consumption indexes; on the basis of the feature tensors, feature tensors corresponding to different granularities are constructed according to the time granularities; the building energy consumption prediction model comprises a multi-granularity feature extraction network, an attention mechanism LSTM network and a prediction module, the output of the multi-granularity feature extraction network serves as the input of the a