Electricity consumption characteristics and prediction model of a large airport terminal in Beijing

The large airport terminal is the city's transportation hub and logistics center, and its building electricity consumption is twice or even higher than that of general large public buildings. Energy-saving and consumption reduction of the terminal building is not only the need to improve the operational efficiency and reduce the operation cost of the airport but also the need to realize the goal of "double carbon" in China. In this study, a large airport terminal in Beijing is taken as the research object. According to the characteristics of its main electrical equipment system, and the measured data of passenger flow from January 2020 to February 2023, combined with Pearson correlation analysis, K-means cluster analysis and multiple regression analysis, the influence and correlation characteristics of building scale, passenger flow and outdoor meteorological parameters on its electricity consumption are analyzed. The results show that: 1) The daily electricity consumption of the terminal can be analyzed by five levels according to the daily passenger flow, namely, 10,000 < N ≤ 30,000,30,000 < N ≤ 50,000, 50,000 < N ≤ 70,000, 70,000 < N ≤ 100,000, and 100,000 < N ≤ 130,000; 2) The electricity consumption of the terminal can be divided into three categories: basic electricity consumption related to the building scale of the terminal, variable electricity consumption I related to passenger flow, and variable electricity consumption II related to outdoor air temperature and passenger flow. Based on this, the terminal's daily electricity consumption prediction model is constructed, and the measured data verify the model's effectiveness. 3) Put forward the evaluation reference value of the daily electricity consumption of different passenger flow in this terminal, When the year passenger flow reaches the design value of 45 million, the normal daily electricity consumption level ranges from 0.44 to 0.48 kW·h/(m2·day). The research results can provide methods and evaluation reference for electricity consumption prediction, and accurate control of electricity consumption of main equipment systems in airport terminal.


Introduction
In recent years, with the rapid growth of China's social economy, the demand for civil airports has shown an extremely fast-growing trend, the volume of civil aviation business has been increasing, and the construction efforts of civil airports have developed strongly.As a typical large public building, large airport terminals are transportation hubs and logistics centers in cities, with a floor area of usually 300,000m 2 ~ 800,000m 2 , and annual passenger flow of up to 30 ~ 50 million with thousands of various types of equipments and complex equipment systems Airport Department of Civil Aviation Administration of China (2017).According to the 2020 China Annual Development Report on Building Energy Efficiency, the actual average energy consumption of domestic airport terminals is about 177 kW•h/m 2 , which is higher than the relevant constraint value (140 kW•h/m 2 ) and twice as much as the standard for existing commercial buildings (China Building Energy Conservation Association.China Building Energy Conservation Report (2016).In addition, the COVID-19 epidemic, which has lasted for three years, has greatly reduced domestic civil aviation routes and flight.The lowest transport capacity of the whole industry is about 5% ~ 10% in the normal period, which is far lower than the design flow of the terminal Jinghang (2022).Most of the equipment systems in the terminal are in a running rate of far below the design value.The system selection of each electrical equipment in the terminal is too large and not fully utilized Kaiguo et al. (2020).Therefore, it is of great significance to have a good command of the electricity consumption characteristics of the terminal to improve the operating efficiency of the main equipment systems and reduce the operating cost.
There have been many studies on airport electricity consumption characteristics at home and abroad using model prediction, computer simulation, on-site questionnaire surveys and on-site tests.Balaras (2003) surveyed the electricity consumption of 29 airport terminals in Greece, and found that the average energy consumption of the terminals was 234 kW•h/m 2 .Yu (2020) surveyed the electricity consumption of 22 terminal buildings in different climatic zones of China.The results showed that the annual electricity consumption intensity of the terminal was about 129 ~ 281 kW•h/(m 2 a); Li et al. (2021) investigated the electricity consumption of air-conditioning system in seven large terminal buildings in China.The electricity consumption of air-conditioning system in terminal buildings accounts for about 30-60% of the total electricity consumption, and the electricity consumption of air handling unit accounts for 40-74% of the electricity consumption of air-conditioning system.Liu Xiaohua Liu et al. (2019) simulated the dynamic distribution of airport passengers by using AnyLogic software, taking the passenger flow data from the field survey as input data.The results show that the time when the passenger flow exceeds the design value only accounts for 3.6% of the total time, far below the design value; Cardonae et al. (2006) put forward that the energy consumption of airport terminals mainly depends on the changes of structure and operation process, and the structural factors mainly include scale, capacity, building orientation, etc. Operation characteristics are mainly affected by factors such as annual passenger volume and climatic conditions; Xu et al. (2015) combined with the analysis results of energy consumption data of China Pudong International Airport, put forward that monthly average temperature, passenger flow and cargo throughput are the three main factors that affect the energy consumption of large airports.
Research on public building energy consumption prediction can be divided into two categories: statistical method (multiple regression model) and artificial intelligence method (artificial neural network model) Ziwei et al. (2018).Among them, the multiple regression model method is widely used in building energy consumption prediction because of its simplicity.Hao and Feng. (2021) established a grey multiple linear regression air conditioning energy consumption prediction model with high accuracy; Xiaowei et al. (2008) established a prediction model of commercial building energy consumption through sample data collection and statistical methods; Bryan et al. (2012) proposed Meta-model based on multiple regression only by using seven parameters; Xianliang et al. (2020) based on the operation data of an airport terminal in South China, from 2016 to 2019, established the air conditioning system energy consumption prediction model through regression analysis; Kim et al. (2020) established the energy consumption prediction model of the airport through multiple regression model according to the survey data of 30 existing models and the simulation results of 90 specific spatial models in North America.
Research on public building energy consumption evaluation can be mainly divided into three categories: building score evaluation method, simulation analysis method and statistical analysis method Youguo et al. (2007).Chung et al. (2006) established the regression model of energy consumption of commercial buildings through multiple regression analysis, and by removing the bias effect of significant factors, the benchmark percentile table of normalized energy consumption intensity of commercial buildings was given.Ding et al. (2018) used K-means clustering and normal distribution test to analyze the energy consumption data of public buildings in China, and determined the energy consumption quota standard and classification of public buildings in Chongqing.Wu et al. (2020) put forward a method to estimate the total energy consumption of buildings based on the energy consumption indexes of different departments, thus obtaining the classified energy consumption indexes of multi-purpose buildings; Wu et al. (2020) studied the energy consumption status of public transport buildings in China based on a large number of investigation results, and obtained the annual average energy consumption benchmarks.ofthree types of buildings.Park (2016) based on the data from 1072 office buildings in South Korea, applied correlation analysis and a decision tree to Equationte six types of office building energy benchmarks.
To sum up, most of the previous studies were on the consumption characteristics, prediction model and evaluation benchmark value of the total electricity consumption of public buildings, or only on the air conditioning system in its main equipment systems, but it was far from enough for the special large public buildings such as airport terminal, with various main equipment systems and different influencing factors.At the same time, the terminal mainly serves passengers, especially at present, the number of passengers affected by the epidemic fluctuates greatly, so adopting the same evaluation benchmark value for different passengers is not conducive to the development of energy-saving work and the exploration of building energy-saving potential.In addition to analyzing the total power consumption of the terminal, it is urgent to analyze its main equipment systems and establish a prediction model based on the characteristic analysis of main equipment systems and a method for evaluating the reference value of different passenger flow, to solve the difficult problem of forecasting and restricting the terminal under different operation modes.
Therefore, this study takes a large airport terminal in Beijing as the research object, aiming at the accurate control of electricity consumption in the large airport terminal, and research the key influencing factors of the electricity consumption characteristics of the terminal according to its building function characteristics and the measured data from January 2020 to February 2023.Pearson correlation analysis and K-means cluster analysis is used to study the electricity consumption characteristics of main equipment systems.Based on the above analysis results, combined with multiple regression analysis and other statistical methods, the prediction method of the daily electricity consumption prediction model is established.In addition, this study the evaluation reference value of different passenger flow of the studied terminal.It is expected to provide methods and evaluation reference for the prediction of electricity consumption and the accurate control of main equipment systems in the terminal.

Building description
The research object is Beijing new airport, the northern region, with 5 floors above ground and 2 floors underground, with a building height of 50 m.
The equipment systems of the terminal can be roughly divided into 10 categories: air conditioning system(ACS), illuminating system(IS), luggage system(LS), elevator system(ES), commercial facility system(CFS), office facility system(OFS), weak electricity system(WES), firefighting system(FS), corridor bridge maintenance system(CBMS), and electric vehicle charging station system(EVCSS).All energy sources are electric energy.

Measured data and processing
The sample data analyzed in this study are the daily electricity consumption of the terminal from January 2020 to February 2023 (1155 data) and the daily electricity consumption of 10 main equipment systems from January 2022 to February 2023 (424*10 data).Daily passenger flow information comes from the daily official announcement of the airport; Daily outdoor meteorological parameters comes from the data of the site located at the airport on the website of the China Meteorological Center.
Combined with the operating characteristics of the terminal, about 10 unreasonable values of main equipment systems were eliminated.

Pearson correlation coefficient
In this paper, Pearson correlation analysis is used to study the close relationship between variables Macqueen (1967).The Pearson correlation coefficient between two variables X and Y can be expressed as Eq. 1.
The range of the Pearson correlation coefficient is [-1,1].When X and Y are independent, the Pearson correlation coefficient is 0. When the correlation is negative, the Pearson correlation coefficient is between -1 and 0; When the correlation is positive, the Pearson correlation coefficient is between 0 and 1.The degree of correlation of variables is characterized by the absolute value of the correlation coefficient.The correlation coefficient is classified as follows: 0.5 ~ 1.0 is a strong correlation; 0.2 ~ 0.5 is a weak correlation; 0 ~ 0.2 means no correlation.
where ρ XY is the correlation coefficient of variables X and Y; E is the mathematical expectation of the sample.

K-means clustering
Cluster analysis usually refers to the classification of specified object groups with similar characteristics according to a certain principle.K-means clustering is a widely used method of clustering analysis Li (2017).K-means clustering is an iterative clustering analysis algorithm, which includes the following steps: dividing the pairs into k groups in advance, calculating the distance between each object and each cluster center, assigning each object to the nearest cluster center, recalculating each new cluster center after clustering, and calculating the distance between each object and each new cluster center, and then assigning each object to the nearest new cluster center again, and repeating the above process until the cluster center does not change Yuan and Song (2009).K-means clustering algorithm has the advantages of easy implementation, simple principle and fast clustering.
The most important parameter in the K-means algorithm is the K-value.So the Silhouette Coefficient and Kalinsky-Harabas index are used as indicators to choose the best K-value.
(1) Silhouette Coefficient (SC): Silhouette coefficient takes the value range [-1,1], the more similar the distance of the same category samples the more distant the different category samples, and the higher the score, as shown in Eq. 2.
where, SC is the contour coefficient; a indicates the average distance between the sample and other samples in the same cluster; b indicates the average distance between the sample and all the samples in the nearest cluster.
The Euclidean distance is chosen to determine the similarity between the data, as in Eq. 3.
where, ρ is the linear distance between the point ( x 2 , y 2 ) and the point ( x 1 , y 1 ).
(2) Kalinsky-Harabas index (CH): the tightness within a class is measured by calculating the squared sum of the distances between individual points in the class and the midpoint in the class, and the separation of the data set is measured by calculating the squared sum of the distances between the center of each class and the center of the data set, as shown in Eq. 4. (1) where, CH is the Kalinsky-Harabas index; n denotes the total number of instances, k denotes the number of clusters, Tr(S B ) denotes the trace of the inter-class departure matrix, and Tr(S w ) denotes the trace of the intra-class dissimilarity matrix.

Regression analysis and evaluation index of accuracy
Regression analysis is a statistical analysis method to study the quantitative relationship between a dependent variable and one or more independent variables.When there are multiple independent variables, it is called multiple regression Cheng et al. (2018).At present, CVRMSE (Coefficient of Variation of the Root Mean Squared Error) and NMBE (Normalized Mean Bias Error) are commonly used to evaluate errors Wang and Wu (2015).
In this paper, NMBE, CVRMSE and fitting index (IA) are used to evaluate the error of regression analysis.The above indicators reflect the deviation between the predicted value and the experimental value, NMBE is the average of the absolute values of the deviation between all individual observed values and the arithmetic mean, and CVRMSE is the average of the measured values obtained by normalizing the root mean square error, and IA is the indicator for testing the actual value and the simulated value.Among them, when NMBE is within 5% and CVRMSE is within 15%, it can be considered that the calculated values are in good agreement with the real values, and The value range of IA is [0,1], and the closer to 1, the higher the simulation accuracy.
where NMBE is the normalized mean bias error, %; CVRMSE is the coefficient of variation of the root mean squared error, %; IA is the fitting index; O i is the actual observation value of day i; P i is the calculated value of day i; O pave is the average of the actual observed values of all verification. (4)

Passenger flow of the terminal
The airport was officially opened to traffic on September 25th, 2019.According to the design plan, the recent annual passenger flow is 45 million.According to statistics, the actual passenger flows in 2020, 2021 and 2022 are respectively 36%, 58% and 23% of the recent planning (Fig. 1).The fluctuation of passenger flows is directly related to the impact of the epidemic in Covid-19 in recent three years.During the severe epidemic, the average daily passenger flow is less than 10,000.
Figure 2 reflects the passenger occupancy rate of the terminal for 24 h a day when the daily number of passengers is 70,000.The passenger occupancy rate is estimated by the number of flights per hour.As shown in the figure, the daily passenger flow generally has two peaks in the morning and the afternoon, among which 6:00 in the morning is the departure time of the plane and 19:00 in the afternoon is the arrival time of the plane, so the peak time of the passengers is from 6: 00 to 19: 00.Similarly, the correlation analysis of outdoor parameters in the following article is also selected in this period.

Total electricity consumption
Figure 3 reflects the monthly change in terminal electricity consumption from January 2020 to February 2023.In the three complete years from 2020 to 2022, the monthly average electricity consumption per unit area was 12.65 kW h/(m 2 month), 12.51 kW h/ (m 2 month) and 10.85 kW h/(m 2 month) respectively.Generally speaking, there is little difference in these three years, but the monthly change trend is similar, indicating that the electricity consumption of the terminal changes in annual cycles, which is similar to those of other airports Xianliang et al. (2020) that have operated relatively stable.Among them, the electricity consumption in 2020 is slightly higher than that in 2021, but the annual passenger flow in 2020 is about 16.5 million, and in 2021 is about 26.5 million, indicating that after more than one year of operation, the equipment system began to enter a stable stage in 2021.In 2022, the overall trend of electricity consumption is the same as before, but due to the epidemic situation, the annual passenger flow is only 10.15 million, so the annual electricity consumption is less than that in 2020 and 2021.The electricity consumption of the terminal is closely related to passenger flow, and is also influenced by outdoor meteorological conditions.The peak of electricity consumption appears in summer and winter.

Electricity consumption of main equipment systems
Figure 4 reflects the electricity consumption of the main equipment systems.The points in Fig. 4 show the distribution of box diagram data in detail.Among them, the air conditioning system is the largest and the fluctuation of data is also large, with an average consumption of about 2.45 kW•h/(m 2 •month); The second is the weak electricity system, with an average consumption of about 1.77 kW•h/(m 2 •month); The third is corridor bridge maintenance, illuminating system, and commercial facility system, with an average energy consumption of about 1.44 kW•h/(m 2 •month); The electricity consumption of Fig. 3 Monthly electricity consumption of the terminal luggage system, firefighting system, and electric vehicle charging station system is relatively small, with an average consumption of about 0.65 kW•h/(m 2 •month); The electricity consumption of office facility is the lowest, and the fluctuation of data is also small.The energy consumption is about 0.12 kW•h/(m 2 •month).

Correlation analysis of total electricity consumption
According to the functional characteristics, and electricity consumption of the terminal, combined with the research results on the influencing factors of electricity consumption of public buildings (Zhang et al. 2021;Dong et al. 2021;Pengfei et al. 2019; Civil Aviation Administration of China 2017), it can be considered that the total electricity consumption of the terminal is closely related to passenger flow and outdoor meteorological conditions.Therefore, according to Sect."Pearson correlation coefficient" and Eq. ( 1), the daily electricity consumption of the terminal and the daily passenger flow, daytime air temperature (6: 00 to 19: 00.), wind speed, and relative humidity were analyzed (Table 1).The analysis results show that the daily electricity consumption of the terminal is moderately related to the daily passenger flow, weakly related to the outdoor daytime average air temperature, but not related to wind speed and relative humidity.

Correlation analysis of the main equipment systems
Similarly, Pearson correlation analysis is made for main equipment systems according to Eq. (1) in Sect."Pearson correlation coefficient" (Fig. 5).The results show that, except for illuminating, office facility, weak electricity system, and firefighting system, the remaining equipment systems are strongly correlated with the daily passenger flow.Besides, the electricity consumption of the air conditioning system is also related to the outdoor daytime average air temperature.

Electricity consumption characteristics of the terminal
According to the analysis results in Sects."Electricity consumption of the terminal" and "Correlation analysis", the electricity consumption characteristics of the terminal have a clear annual cycle, and its daily electricity consumption is related to daily passenger flow and outdoor daytime air temperature.Based on this, combined with Fig. 5, the equipment systems can be divided into three categories.The first category mainly includes illuminating, office facility, weak electricity, firefighting.The electricity consumption of this equipment system is relatively stable, which is defined as basic electricity consumption.The second category includes luggage, elevators, commercial facility, corridor bridge maintenance system, and electric vehicle charging station.The electricity consumption of this category fluctuates with the change in passenger flow, which is defined as variable electricity consumption I.The third category is the remaining air conditioning system.The electricity consumption of the air conditioning system fluctuates with the change in passenger flow and outdoor meteorological conditions at the same time, which is defined as variable electricity consumption II.

Model building
Based on the analysis results in Sect."Electricity consumption characteristics of the terminal", the electricity consumption of 10 main equipment systems in the terminal can be classified into basic electricity consumption, variable electricity consumption I and variable electricity consumption II, and the daily electricity consumption prediction model can be constructed by regression analysis.

Daily basic electricity consumption
According to the definition of basic electricity consumption in Sect."Electricity consumption characteristics of the terminal", the basic electricity consumption of the terminal consists of illuminating, office facility, weak electricity, and firefighting system.
Figure 7 shows the variation of daily basic electricity consumption in 2022, and the variation trend of daily electricity consumption is observed in descending order of passenger flow.It can be seen that with the increase in the number of passengers, this part of daily electricity consumption has little variation and fluctuates around 100,000 kW•h, which further proves that this part has little correlation with passenger flow and seasons, so this part of electricity consumption Y1 can be taken as a fixed value (Eq.10).
where Y1 is the daily electricity consumption of the basic electricity consumption, kW•h/ day.

Daily variable electricity consumption
According to the definition of variable electricity consumption I in Sect."Electricity consumption characteristics of the terminal", this part of electricity consumption is mainly caused by the consumption of luggage, elevator, commercial facility, corridor bridge maintenance, and electric vehicle charging station system.Figure 8 shows the daily variation of electricity consumption I in 2022, and the daily variation trend is observed in descending order of passenger flow.As shown in the figure, with the increase in the number of passengers, this part of electricity consumption gradually increases and tends to be stable, which further shows that this part is closely related to the passenger flow.
According to the analysis result of Fig. 9 and the regression analysis method, the fitting formula (Eq.11), and the regression curve are shown in Fig. 11.
where, Y2 is the daily electricity consumption per passenger of variable electricity consumption I, kW•h/(people•day); N is the daily passenger flow, people/day.

Daily variable electricity consumption
Variable electricity consumption II is the air conditioning system.Figure 10 shows the variation of the daily variable electricity consumption II with time in 2022.It can be found that the seasonal characteristics of this part is obvious, and there are also obvious fluctuations between China Lunar 24 solar terms.At the same time, the obvious fluctuations in each solar term are mostly affected by the fluctuation of the number of people in the solar term.From these two aspects, it is further verified that the variable electricity ( 10)   Combining the results of the Pearson correlation analysis in Sect."Correlation analysis", and according to the analysis results in Fig. 12 and the multiple regression analysis methods, the fitting formula (Eq.12) about daily variation electricity consumption II can be obtained.
where Y3 is the daily electricity consumption of variable electricity consumption II, kW•h/day; T is the outdoor daytime average air temperature,℃, N is the daily passenger flow, people.

Daily electricity consumption prediction model
Combining Eq. ( 10) to Eq. ( 12), the daily electricity consumption prediction model (Eq.13) of the terminal can be established.When the daily passenger flow and outdoor daytime average air temperature of the terminal are known, the daily electricity consumption Y of the terminal can be calculated according to Eq. ( 13).When the electricity consumption in the next few days is known, the equipment operation in the terminal can be adjusted in advance, which is more conducive to reducing the operating cost of the airport.
Where Y is the daily electricity consumption of the terminal, kW•h/day, and N is the daily passenger flow, people/day.

5Model verification
Figure 11 shows the comparison result between the calculated value and the measured value of the daily electricity consumption of the terminal from January to February 2023.According to Eq. ( 7) to Eq. ( 9), the NMBE value is -2.74%, the CVRMSE value is 5%, which the errors are both within the control threshold (5% and 15%), and the IA is 0.994, which is close to 1.All three evaluation indexes show the effectiveness of the prediction model.

Annual passenger flow and electricity consumption
According to the daily electricity consumption prediction model proposed in Sect."Model building", the daily electricity consumption of the terminal under different passenger flows, as well as the basic electricity consumption, variable electricity consumption I and variable electricity consumption II.The calculation conditions are shown in Table 2, in which the daily passenger flow n is taken as the average of annual passenger flow for 365 days, and the daytime average value T of outdoor daytime air temperature refers to the parameters in 2021.
The results are shown in Fig. 12, with the increase in passenger flow, the annual electricity consumption shows an upward trend.The main factors affecting this trend are the (12) Y 3 = 0.75(118.1T 2 − 2631.8T+ 53790) + 0.39N + 3232, R 2 = 0.72 variable electricity consumption I and the variable electricity consumption II.The basic electricity consumption is stable, which is not affected by the increase of passenger flow.Among the three categories of electricity consumption, basic electricity consumption, variable electricity consumption I and variable electricity consumption II account for about 28%, 49% and 23% of the total electricity consumption respectively.

Evaluation reference of electricity consumption
Applying Eqs. ( 2) to (4), the cluster analysis of this terminal from January 2020 to February 2023 is carried out.Cluster analysis is made on the daily passengers and daily the electricity consumption., and the result shows in Table 3.
The results of cluster analysis show that according to the difference of k value (K = 3 ~ 10), the passenger flow can be classified into eight categories.Generally, the larger SC and CH of the corresponding K value, the classification is more representative.According to this principle and the analysis results in Table 3, it can be considered that the classification method of K = 8 is the best, that is, the daily passenger flow can be divided into 8 levels, namely, N ≤ 2500, 2,500 < N ≤ 5,000, 5,000 < N ≤ 10,000, 10,000 < N ≤ 30,000, 30,000 < N ≤ 50,000,50,000 < N ≤ 70,000,70,000 < N ≤ 100,000, and  Considering that the daily passenger flow less than 10,000 people is mainly affected by the epidemic, which is atypical and does not belong to the normal operation of the terminal, this study focuses on the analysis of the daily passenger flow of more than 10,000.So the daily electricity consumption of the terminal can be divided into 5 levels, namely, 10,000 < N ≤ 30,000,30,000 < N ≤ 50,000,50,000 < N ≤ 70,000,70,000 < N ≤ 100,000, and 100,000 < N ≤ 130,000.
The classification results of 5 levels are shown in Fig. 14.According to the structural characteristics of the box diagram, the upper quarter value, median value and lower quarter value can be used as reference values to evaluate the electricity consumption of the terminal in each classification.The median value as the "admission value".The lower quarter value reflects the energy-saving characteristics of the terminal, which can be defined as "advanced value".The upper quarter value reflects that the electricity consumption level of the terminal deviates from the normal range, which can be defined as "limited value".It should be noted that, the daily passenger flow of 100,000 < N < 130,000 is relatively small, so the sample size of this type of daily passenger flow is supplemented by the daily electricity consumption prediction model established in Sect."Model building".The specific electricity consumption reference values are shown in Table 4.
For the actual operation of the terminal, on the one hand, according to the prediction model in Sect."Model building", the daily electricity consumption in the future can be calculated, then the equipment operation in the terminal can be adjusted in advance, which is more conducive to reducing the operating cost of the airport.On the other hand, the actual operation electricity consumption level can be judged by comparing with the given evaluation reference values in Table 4.If the electricity consumption is close to the upper quarter value in Table 4, it means that the electricity

Conclusion
Based on the investigation and analysis results of electricity consumption characteristics of a large airport terminal in Beijing from January 2020 to February 2023, the following conclusions can be obtained: (1) The daily passenger flow of the terminal is mainly concentrated in the period from 6: 00 to 19: 00, and the average outdoor air temperature during this period can be used as one of the factors affecting the electricity consumption of the terminal.The annual variation of electricity consumption in terminal buildings is similar.Daily electricity consumption is strongly related to daily passenger flow, weakly related to outdoor daytime average air temperature (6:00 ~ 19:00), and has no correlation with outdoor wind speed and relative humidity.the terminal.
(2) According to the electricity consumption characteristics of the main equipment systems of the terminal and applying the Pearson correlation analysis method, it is proposed that the electricity consumption of the terminal can be divided into three categories, namely, the basic electricity consumption related to the building scale of the terminal, the variable electricity consumption I related to passenger flow, and variable electricity consumption II related to outdoor air temperature and passenger flow.The order of the three categories of electricity consumption is variable electricity consumption I > basic electricity consumption > variable electricity consumption II; Among them, variable electricity consumption I account for about 49% of the total electricity consumption, and basic electricity consumption accounts for about 28% of the total electricity consumption.(3) Based on the three classification principles of main equipment systems and the regression analysis method, the daily electricity consumption prediction model of the main equipment systems in the terminal is proposed, and the measured results verify the effectiveness of the model.(4) Combined with the terminal daily electricity consumption prediction model and K-means cluster analysis method, five classification principles corresponding to the daily passenger flow scale of the terminal are put forward, and the evaluation reference value of the daily electricity consumption of the terminal is given, including "limited value", "admission value" and "advanced value".The research results can provide an evaluation reference for the terminal electricity consumption.

Research involving human participants and/or animals
This article does not contain any studies with human participants performed by any of the authors.

Informed consent
For this type of study informed consent is not required.

Fig. 1
Fig. 1 Monthly passenger flow of the terminal

Fig. 4
Fig. 4 Electricity consumption of the equipment systems Figure 6 reflects the daily electricity consumption changes about the three categories in 2022.As shown in the diagram, the basic electricity consumption is relatively stable.Variable electricity consumption I in the transition season is reduced, which is more caused by the obvious decrease of passenger flow from March to May and September to October in 2022.Variable electricity consumption II in the transition season is obviously less than that in winter and summer, with obvious seasonal characteristics.

Fig. 5
Fig. 5 Correlation coefficient of the equipment systems

Fig. 6 Fig. 7
Fig. 6 Daily changes of passengers and three categories of electricity consumption in 2022

Fig. 8
Fig. 8 Daily variable electricity consumption I and passenger in 2022

Fig. 9
Fig. 9 Variable electricity consumption I regression curve

Fig. 11
Fig. 11 Comparison between calculated and actual daily electricity consumption

Fig. 12
Fig. 12 Changes of annual electricity consumption of terminal with passenger the prediction model

Fig. 14
Fig. 14 Box diagram for grading evaluation of electricity consumption of the terminal

Table 1
Correlation analysis of influencing factors

Table 2
Calculation Conditions

Table 3
SC and CH coefficient index Fig. 13 Clustering result of k = 8