基于机器学习的建筑能耗检测预警平台构建

打开文本图片集

中图分类号：TP317.4;TU18 文献标识码：A 文章编号：1001-5922（2025）10-0218-04

Construction of building energy consumption detection and early warning platform based on machine learning

XU Jiping （Shandong Urban Construction Vocational College， Jinan 25O1O3，China）

Abstract：In order to solvethe problems oflow prediction acuracy，poorreal-time performanceand weak scalabilityof traditional energy consumption management methods，this paper proposes a distributed building energy consumption monitoring and early warning platform basedon machine learming integrated model.The platformadopts a hierarchical architecture design，and the perception layer collcts multi-sourcedata in realtime through IoTdevices;the data layer completes missing value filling，outliercorrection，discrete coding and standardized preprocessing，and screens six core features of daycategory，time node，dry bulb temperature，relative humidity，solarradiation intensityand energy consumption powerof heating ventilation and air conditioning system based on feature importance evaluation.Application layer integrates visualization and early warning functions.In order to avoid early warning errrs oromissions，this paper proposes a two-layer integrated prediction model：the first layer generates preliminary predictionresults inparalll byartificial neuralnetwork and lightweight gradient hoist，and the second layer fuses thedual modeloutputthrough linear regression weighting，which significantlyimproves the model performance.The modelverification shows thatthe platform canrealize real-time data acquisition，dynamic threshold warning and multi-building cluster paralel management，and provide full-process technical support for energy-saving decision-making.

Key words：building energy consumption prediction;lightweight gradient hoist;artificial neural network;integrated model; monitoring and early warning platform

在全球能源消耗持续增长的背景下，建筑能耗作为能源消耗的重要组成部分，其管理效率直接影响能源节约与环境保护目标的实现。（剩余5071字）