CN-121997081-A - Building energy consumption characteristic pattern recognition and diagnosis method and system thereof

CN121997081ACN 121997081 ACN121997081 ACN 121997081ACN-121997081-A

Abstract

The invention provides a building energy consumption characteristic pattern recognition and diagnosis method and a system thereof, comprising the following steps of obtaining building energy consumption historical data; the method comprises the steps of carrying out Fourier transformation on an obtained building energy consumption historical data sequence to obtain corresponding frequency domain data, establishing a building energy consumption characteristic pattern recognition and diagnosis model and index, clustering frequency domain waveform amplitudes by adopting a clustering method to recognize dominant periodic characteristics of building energy consumption, establishing a building energy consumption characteristic pattern recognition and diagnosis classification standard, constructing a corresponding diagnosis quadrant region, obtaining relation distribution of the amplitude and frequency of the building energy consumption dominant waveform based on the frequency domain data and the dominant periodic characteristics, calculating periodic fluctuation characteristic intensity and variation coefficient according to the model and index, and recognizing and diagnosing the building energy consumption pattern according to the classification standard and the quadrant region. The method realizes accurate extraction of the building energy consumption cycle characteristics and scientific diagnosis of the modes, and provides data support and decision basis for building energy conservation optimization.

Inventors

LIU ZHAOHUI
LI ZHENYU
PAN XI
CHEN HAO
SONG JIALING

Assignees

上海建工集团股份有限公司

Dates

Publication Date: 20260508
Application Date: 20251218

Claims (10)

1. The building energy consumption characteristic pattern recognition and diagnosis method is characterized by comprising the following steps of: S1, acquiring historical building energy consumption data; s2, carrying out Fourier transform on the acquired building energy consumption historical data sequence to obtain corresponding frequency domain data; s3, building a building energy consumption characteristic pattern recognition and diagnosis model and indexes based on the frequency domain data, wherein the indexes comprise a first index and a second index; S4, based on the frequency domain data, clustering the frequency domain waveform amplitude by adopting a clustering method, and identifying the dominant periodic characteristics of the building energy consumption; s5, building a building energy consumption characteristic pattern recognition and diagnosis grading standard based on the first index and the second index, and building a corresponding diagnosis quadrant region; And S6, acquiring the relation distribution of the amplitude and the frequency of the building energy consumption dominant waveform based on the frequency domain data of the step S2 and the dominant periodic characteristics of the step S4, calculating the periodic fluctuation characteristic intensity and the variation coefficient according to the model and the index of the step S3, and identifying and diagnosing the building energy consumption mode according to the grading standard and the quadrant area of the step S5.
2. The method for identifying and diagnosing a characteristic pattern of building energy consumption according to claim 1, wherein the historical data of building energy consumption in the step S1 includes at least one of total building energy consumption, building sub-term energy consumption, building system equipment energy consumption or building space energy consumption, and the building sub-term energy consumption includes at least one of heating, ventilation and air conditioning energy consumption, illumination energy consumption, socket energy consumption, power energy consumption or special energy consumption; and/or the number of the groups of groups, The sampling interval of the building energy consumption historical data is 1 hour, and the total sampling duration exceeds 1 year.
3. The method for identifying and diagnosing the characteristic pattern of the energy consumption of the building according to claim 2, wherein the energy consumption of the air conditioning system in the energy consumption of the building system equipment comprises energy consumption data of a water chilling unit/heat pump unit, a chilled water pump/hot water pump, a cooling tower, an air conditioning terminal fan coil/air handling unit or a fresh air system.
4. The method for identifying and diagnosing the characteristic pattern of the energy consumption of the building according to claim 1, wherein the fourier transform in the step S2 is a fast fourier transform method.
5. The method for identifying and diagnosing the characteristic pattern of the energy consumption of the building according to claim 1, wherein the clustering method in the step S4 adopts a hierarchical clustering method.
6. The method for identifying and diagnosing a characteristic pattern of energy consumption of a building according to claim 1, wherein the classification criterion in the step S5 is determined based on the second index and the classification limit of the normalized first index, and/or, The first index is the fluctuation characteristic intensity sigma k,i of the building energy consumption period and is used for representing the fluctuation amplitude of the building energy consumption in a specific period; The second index is a building energy consumption period fluctuation variation coefficient C V,i and is used for representing the discrete degree of building energy consumption characteristic mode difference.
7. The method for identifying and diagnosing a characteristic pattern of energy consumption of a building according to claim 6, wherein the classification limit of the normalized first index includes: A first limit of 0< sigma' k,i < 0.4; a second limit of 0.4< sigma' k,i < 0.8; a third limit of 0.8< sigma' k,i < 1; Wherein sigma' k,i is the normalized building energy consumption period fluctuation characteristic intensity; the grading limit of the second index includes: a first limit of 0< C V,i < 0.1; the second limit is 0.1< C V,i < 0.3; The third limit is 0.3< C V,i < 0.5; a fourth limit of C V,i >0.5; Wherein C V,i is the fluctuation coefficient of variation of the building energy consumption period.
8. The method for identifying and diagnosing a characteristic pattern of energy consumption of building according to claim 2, wherein the period for which the periodic fluctuation characteristic intensity value is calculated in the step S6 includes 1 year period, 1 week period, and 1 day period.
9. The building energy consumption characteristic pattern recognition and diagnosis method according to claim 8, wherein for a plurality of building sub-term energy consumption, the periodic fluctuation characteristic intensity value of each sub-term energy consumption is normalized in the same period, and the periodic fluctuation characteristic intensity normalization result of each sub-term energy consumption is equal to the ratio of the periodic fluctuation characteristic intensity value in the corresponding period to the periodic fluctuation characteristic intensity maximum value in the same period of all sub-term energy consumption.
10. A building energy consumption signature pattern recognition and diagnosis system, comprising: The data acquisition module is used for acquiring building energy consumption historical data; The frequency domain transformation module is used for carrying out Fourier transformation on the acquired building energy consumption historical data sequence to obtain corresponding frequency domain data; The model building module is used for building a building energy consumption characteristic pattern recognition and diagnosis model and indexes, wherein the indexes comprise a first index and a second index; The cluster analysis module is used for clustering the frequency domain waveform amplitude by adopting a clustering method based on the frequency domain data, and identifying the dominant periodic characteristics of the building energy consumption; the standard construction module is used for establishing a building energy consumption characteristic pattern recognition and diagnosis grading standard based on the first index and the second index, and constructing a corresponding diagnosis quadrant region; the identification and diagnosis module is used for acquiring the relation distribution of the amplitude and the frequency of the building energy consumption dominant waveform based on the frequency domain data of the frequency domain transformation module and the dominant periodic characteristics of the cluster analysis module, calculating the periodic fluctuation characteristic intensity and the variation coefficient according to the model and the index of the model building module, and identifying and diagnosing the building energy consumption mode according to the grading standard and the quadrant area of the standard building module.

Description

Building energy consumption characteristic pattern recognition and diagnosis method and system thereof Technical Field The invention belongs to the technical field of building energy consumption monitoring and diagnosis, and particularly relates to a building energy consumption characteristic pattern recognition and diagnosis method and a system thereof. Background Along with the increasing severity of global energy crisis and climate change problems, building energy conservation has become a key link for realizing sustainable development strategy. The building energy consumption system is a complex nonlinear system comprehensively influenced by various factors such as outdoor climate, indoor personnel activities, equipment operation strategies and the like, and the operation data contains rich information reflecting the characteristics and the operation state of the system. Therefore, by deep mining and analysis of building energy consumption data, accurate identification of energy consumption characteristics and effective diagnosis of running states are realized, and the method has important significance for improving building energy efficiency and reducing running cost. Currently, the mainstream methods in the building energy consumption analysis field focus on time domain analysis, for example, by analyzing time-by-time and day-by-day energy consumption curves, or calculating month or year statistical indexes (such as energy intensity, load rate and the like) to evaluate the overall energy consumption condition of a building. However, this time domain analysis method has significant limitations: First, it is difficult to effectively strip and quantify the periodicity laws hidden in the energy consumption data, such as the daily period caused by day-night alternation, the weekly period caused by the difference between the weekdays and weekends, and the annual period caused by seasonal variation, etc.; Secondly, for a complex building energy consumption system formed by a plurality of energy consumption systems (such as heating ventilation air conditioning, lighting, sockets and the like), the time domain analysis method is difficult to clearly reveal the cooperative or asynchronous characteristics of each subsystem on different cycle scales, and the fine management and fault diagnosis capability of the sub-energy consumption are limited; And thirdly, a systematic frequency domain feature extraction and diagnosis system is lacking, key dominant periods for representing the operation rules cannot be effectively extracted from the frequency domain angle, and a specialized index system for distinguishing different operation modes from health states is also lacking. In summary, the prior art has the following defects that (1) key leading periods capable of representing the operation rule of the building energy consumption cannot be effectively extracted and quantified from the frequency domain characteristics of the building energy consumption, (2) a diagnosis index system which is specially aimed at the frequency domain characteristics of the building energy consumption and is used for distinguishing different operation modes and health states is lacked, and (3) a judgment standard and a model for comprehensively comparing and visually diagnosing the multi-system and multi-period characteristics are not established. Disclosure of Invention The invention provides a building energy consumption characteristic pattern recognition and diagnosis method and a system thereof, which realize accurate extraction of building energy consumption periodic characteristics and scientific diagnosis of patterns and provide data support and decision basis for building energy conservation optimization. The technical scheme of the invention is as follows: A building energy consumption characteristic pattern recognition and diagnosis method comprises the following steps: S1, acquiring historical building energy consumption data; s2, carrying out Fourier transform on the acquired building energy consumption historical data sequence to obtain corresponding frequency domain data; s3, building a building energy consumption characteristic pattern recognition and diagnosis model and indexes based on the frequency domain data, wherein the indexes comprise a first index and a second index; S4, based on the frequency domain data, clustering the frequency domain waveform amplitude by adopting a clustering method, and identifying the dominant periodic characteristics of the building energy consumption; s5, building a building energy consumption characteristic pattern recognition and diagnosis grading standard based on the first index and the second index, and building a corresponding diagnosis quadrant region; And S6, acquiring the relation distribution of the amplitude and the frequency of the building energy consumption dominant waveform based on the frequency domain data of the step S2 and the dominant periodic characteristics of the step