Quantitative research of incentive effects of heat-metering policy on energy-saving behavior in Tianjin

Abstract

To promote energy efficiency and emission reduction, the Chinese government has invested large amounts of resources in heat-metering reform over the past decade. However, heat-metering, which can reduce energy consumption by 15% at least in developed countries, is still not well known in China. This paper quantitatively analyzed the arousal effect of heat-metering policy on occupancy behavior regarding energy saving utilizing statistics method based on measured data of heating energy consumption of approximately 20000 users from 2008 to 2012 in Tianjin. The statistical data showed significant difference on energy consumption between users based on metering and based on area. The energy-saving rate in the heating season increased significantly from 4.11% in 2008–2009 to 10.27% in 2011–2012 as the implement of the heat-metering policy. The difference in energy-saving according to various outdoor temperatures showed that the energy-saving of occupancy behavior was more significant in a warm season than in a cold season. It also showed that the impact of heat-metering policy would be more pronounced in generally insulated buildings (15.55%) than in better insulated units (6.45%). Besides, this paper proposed some feasible suggestions for the formulation and implementation of heat-metering policy in northern heating areas of China.

Keywords

heat-metering / occupancy behavior / heating energy consumption / policy

Introduction

With the rapid development of China’s economy, the increasing high demand for better indoor environment quality (IEQ) leads to a variety of high energy consumption. Approximately 30% of the total energy is consumed by buildings in China and it is still increasing [1]. According to recent data, the area of northern heating region in China has exceeded 7.5 billion square meters, consuming approximately 14.3 million tons of coal, accounting for 40% of the total building operation energy [2]. So the energy-saving potential in space heating is enormous. The current policy, which charges heating bills based on the heating area can neither meet people’s IEQ requirement nor motivate people to change their behavior to save energy. To promote energy efficiency and emission reduction, the Chinese government has invested large amounts of resources in heat-metering reform over the past decade.

Heat metering based on households not only makes the relationship among the occupant behavior, energy consumption for space heating and heating fee clear but also fully mobilizes the enthusiasm of occupancy behavior to save energy [3]. Previous studies have shown that user behavior may have a significant impact on energy for space heating [4-6], but quantitative analysis of this influence has not been done yet. Since building energy efficiency has been upgraded in the past years in terms of building features itself (such as high performance insulation, optimization of the operation strategy), there is little space left. This highlights the impact of occupancy behavior on building energy-saving from the other side.

Based on previous literature, many factors affecting the building heat consumption have been intensively studied such as building thermal characteristics, operation strategy, etc. However, the impact of heat-metering policy on user behavior has been seldom studied. In general, people think that heat-metering can make user save more energy in theory. But in fact, the energy-saving rate is only 15%-24% in some cities in Germany [3]. de Meester T et al. [7] analyzed the effect of relevant factors (life-style, occupation, and family numbers) on users’ energy-saving behavior by means of modeling, yet they failed to conduct quantitative analysis of various other factors. Xu et al. [8] claimed that the total flow variations can be reduced by 10% from the system of charging the heating fee based on metering compared with the one based on heating area. Santin O G et al. [9] argued that it is not accurate only by utilizing modeling and software simulation; therefore, a conclusion was drawn that the characteristics of users’ behavior can effect heating energy consumption by approximately 4.2% based on the analysis of the actual data for 1000 users. However, the difference between the sample types to be chosen are quite significant, and the data cannot reflect the independent impact of the heat-metering policy. Xu et al. [10] found that users’ behavior have changed at all in almost any form upon implementing charging the heating fee based on heat-metering by the research of the heating energy consumption in heating season for a 9 floor apartment building from 2007 to 2008. In this case, the heat-metering policy has less impact on heating energy consumption. Since the policy has been implemented shortly from 2007 to 2008, the energy saving awareness cannot be easily changed in a short period of time. It is hard to judge the reliability of the conclusion.

For this reason, this paper presents the actual data of heating energy consumption for 20000 households in 4 heating periods from 2008 to 2012 in Tianjin. It aims to quantitatively analyze the incentive effects of heat-metering policy on users’ energy saving behavior.

Data and the method of analysis

Composition of sample

The data used in this paper includes heating energy consumption in 4 heating seasons from 2008 to 2012 in Tianjin, a municipality located in the eastern longitude of 117°11′ and in the northern latitude of 39°10′ in subtropical continental climate which is cold in winter and hot in summer. The maximum temperature is approximately 40.9°C and the minimum temperature is –15.2°C. The variation of typical outdoor meteorological parameters is shown in Fig. 1. The heating season in Tianjin is from Nov. 15 to Mar. 15 of the next year. The average temperature of a typical heating season is –0.2°C. Tianjin is one of the biggest cities using centralized heating supply in northern heating areas of China and is also the exemplary city of promoting energy-saving housing industry<FootNote>

Tianjin Construction Management Committee. Standard DB29-1-2004 Tianjin Energy Efficiency Design Standards for Residential Buildings. Tianjin Urban Construction and Transportation Commission, 2004-01-01

</FootNote>. Up to the end of 2012, it has a heat-metering implementation area of 69.48 million square meters.

The data were collected from1726 apartment buildings (See Figs. 2 and 3), all of which had heat meter and thermostatic radiator valves (TRVs). The heating fee of 802 buildings was charged based on metering, and valid users were 9316 households while the heating of 924 buildings was charged based on area, and valid users were 13164 households. The percentage of building of types is illustrated in Fig. 4. Most of the buildings were multi-story and high-rise ones and middle high buildings accounted for less than one quarter. This proportion also reflected the status of the current distribution of residential buildings in China. The distribution of heating area is demonstrated in Fig. 5. The statistic data shows that a majority of the houses has a housing area of 40 m² to 89 m² while a minority of the houses has a housing area of 100 m² and above. The distribution of heating energy consumption per unit area in a heating season is depicted in Fig. 6. It is seen from Fig. 6 that 84% of the users have a heating energy consumption of 60 kWh/m² to 160 kWh/m² per unit area in a heating season. The heating energy consumption is centralized. The users of high heating energy consumption and low heating energy consumption account for less than 16% of the total users.

Method of data processing

Since the data used in this paper were collected from large members of users, a large amount of outliers occurred due to the damage of heat meter on user side and the effect of human factors from meter readers. These outliers would greatly affect the analysis results and might possibly lead to the incorrect conclusion by the researcher [11]. However, these outliers normally had a similarity: they showed a strongly inconsistency with other data [12]. To maintain a better quality of the statistic data, it is necessary to process the outliers.

There are many criterions to eliminate the outliers such as PanTa criterion, Chauventet criterion, Grubbs criterion, etc. This paper adopts the PanTa criterion to process the data because it is considered to be more applicable to more quantitative samples [11]. General speaking, if a Radom variation is affected by more random factors in which one of the factors cannot play an active role, the normal distribution will be submitted. In accordance with the principle of probability theory, in normal distribution, σ was taken as the standard deviation, μ as average value of all data. PanTa criterion stands for

(1)P\{\left|X-\mu \right|\le \mathrm{3}\sigma \}=\mathrm{0.9974.}

In this formula, the X value almost concentrates in the interval of (μ–3σ, μ + 3σ), and the possibility of the value beyond this scope accounts for less than 0.3%. The heating energy consumption as per unit area for users can be submitted to normal distribution. In this paper, the data are pre-processed by SPSS.

Methods of analysis

Due to the large amount of data used in this paper and the fact that all the users are independent, the method of ‘Independent-Samples t-test’ was used. The data were analyzed by SPSS, to know whether there was significant difference in heat consumption under different ways to charge. Based on this, further analysis was conducted on the effect of the outdoor temperature, and building thermal performance etc. on the heat-metering energy consumption. This paper overcame many interference factors by using a large number of samples to make reasonable inferences from a statistical point of view.

Results and discussion

Significance test of impact of heat-metering policy on user behavior

To verify the incentives of heat-metering policy in the early stages of implementation, this paper made a good use of the heating data in the heating season from 2011 to 2012 to analyze if there was any significant difference for the average value of heating energy consumption between the users whose heating fee were charged based on metering and whose heating fee were charged based on area.

The heating energy consumption per unit area for heating fee based on metering and area can be taken as two totalities and the sample was obtained individually. Therefore, the method of ‘independent samples t-test’ could be adopted. The null hypothesis was that there was no more significant difference for the average value of heating energy consumption between the user whose heating fee was charged based on metering and that of the user whose heating fee was charged based on area. The result of basic statistical calculation under the two conditions is shown in Table 1. It is seen from Table 1 that the average value of heating energy consumption for heating fee based on metering is 105.72 kWh/m², with a standard deviation of 33.29 kWh/m² and that for heating fee based on area is 117.80 kWh/m², with a standard deviation of 37.18 kWh/m².

There is a certain difference between the average value of heating energy consumption for the users whose heating fee was charged based on metering and that of the users whose heating fee was charged based on area, as seen in Table 1. To judge the deviation related to sampling error or system deviation, a test for significant difference between the two values was conducted by using the measured value for F test of Levine, t-test and P-value for the corresponding probability. The first step is to judge if there was any significant deviation between the two population variance by using F-test. The second step is to judge if there was any significant deviation between the mean scores of the two totalities by using t-test. The result of the calculation is tabulated in Table 2. The measured value for F statistics is 64.988 with a P-value of a corresponding probability of 0.000. Taking the significant level α as 0.05, since the P-value of probability is below 0.05, the significant deviation between the two population variance is considered for the existence. According to the result of t-test, the measured value of t statistics is 17.408 with a corresponding two 0-tailed probability P of 0.000. Now that the P-value of probability is below 0.05(α = 0.05), there is a significant difference for heating energy consumption between the users whose heating fee were charged based on metering and the users whose heating fee were charged based on area.

Above analysis indicates that significant effect on promoting user energy-saving behavior is achieved though the policy of heat-metering has been just implemented for only a few years. Besides, this paper made detailed analysis of whether the variation of this effect would be consistent and continuous as the duration of heat-metering policy went on, and whether the effect on users’ behavior was more obvious on weakened when the outdoor temperature was changed.

Continued influence of heat-metering policy on user energy-saving behavior

An analysis was made of whether the user’s energy-saving behavior would be promoted or rebounded since the implementation of the heat-metering policy based on the statistics data in the four heating seasons. As illustrated in Fig. 7, the average value of user’s energy consumption is different in each heating season due to the fluctuation of meteorological parameters over the years. However, for all the heating seasons, the average value of users’ heating energy consumption for the heating fee based on metering are obviously lower than that for the heating fee based on area. This further corroborated the result of initial analysis in Section 3.1. Fig. 7 shows that the energy-saving rate of user behavior increased year by year under the incentives of heat-metering policy. The energy-saving rate is approximately 4.11% in 2008-2009 but is 10.27% in 2011-2012. Thus, it can be seen that the policy has an increasing influence on users’ behavior along with the time.

To further confirm the change characteristics of user energy-saving behavior, a record of sampling tracking was conducted based on the setup of indoor temperature for users of heat-metering. The sample selected is a housing estate for implementing the heat-metering policy. There are 998 households in the housing estate. The investigation of the sampling was conducted for 301,186 and 265 households respectively from 2008 to 2009, 2010 to 2011 and 2011 to 2012. The mean indoor temperatures of each household in the whole heating season were recorded. Quantitative distribution of users in different temperature ranges is depicted in Fig. 8. The peak value in 2008-2009 was between 20 °C and 22°C for 223 households, accounting for 74% of the total sampling. The mean indoor temperature of 20 households (more than 7%) was higher than 22°C.

The survey was blocking-up due to some special reason in 2009-2010, and the indoor temperature distribution for households was, therefore, not obtained. Nevertheless, it is found that the indoor temperature of the user showed a downward trend compared with the data of the next two years. Take the heating season of 2010-2011 as an example: the peak value of 147 households (79%) was in the scope of 19°C-20°C. The investigation results of 2011 to 2012 show that the sampling of indoor temperature was in the range of 18°C-19°C, accounting for 78%. The mean value of indoor temperature was below 18°C, account for 8% of the total number of households (22 households). From indoor temperature of the three years, it can be seen that, as for the user whose heating fee is charged based on metering, the significant arousal effect of heat-metering policy is to promote the user by controlling the indoor temperature reasonably. For example, when the user is not at home or the indoor temperature is high, some users set the indoor temperature at a lower value to save energy. The same conclusion was drawn by Liu et al. [2]. When the indoor temperature is high, 88% of the users whose heating fees are charged based on metering reduce the indoor temperature by adjusting the temperature control knob rather than by opening the window.

The above analysis results indicate that the initial implementation of the heat-metering policy had to be modified and completed in the first couple of years. A procedure of accepting and adapting slowly is necessary for users whose heating fees are charged based on metering. As time passes, the potential energy-saving behavior of the users will be enhanced and remain stable in a corresponding scope.

Effect analysis of various outdoor temperatures

To analyze the relationship between energy-saving potential of user behavior and outdoor temperature, in this paper, a heating season (2011-2012) was divided into four phases whose duration is one month respectively as Nov.15-Dec.15, Dec.15-Jan.15, Jan.15-Feb.15, and Feb.15-Mar.15. The heating degree days (HDD) is calculated as

(2)\mathrm{HDD}=T_B-T,

where HDD stands for degree and quantity by data (D.D), when T>T_B, HDD= 0; T_B stands for the referent temperature (°C), in this paper it is 20°C; and T the average temperature for a day (°C).

The HDD in the four phases were 536, 666, 704 and 503 upon calculation respectively. The statistics data for each phase were processed by the method of research as single-factor variable model, whose result is presented in Figs. 9 and 10.

As HDD increases, as shown in Fig. 9 (Outdoor temperature is decreased), the heating energy consumption per unit area of all the users increases gradually. In the four phases, the user heat consumption has the lowest value of 26.42 kWh/ (m²·month) in Nov.15-Dec.15 and highest heat consumption 32.58 kWh/ (m²·month) in Jan.15-Feb.15. In addition, compared with the slow changes of heating energy consumption of users whose fees are charged by area, it is more obvious for users whose fees are charged by metering with the outdoor temperature fluctuates.

Figure 10 shows the average heat consumption difference with the change of outdoor climate. It is seen that, affected by the heat-metering policy, user behavior energy-saving rate in these four phases decreases respectively, that is, Feb.15-Mar.15 (11.52%), Nov.15-Dec.15 (10.97%), Dec.15-Jan.15 (8.47%), and Jan.15-Feb.15 (7.88%), which indicates that as the outdoor temperature decreases, the difference between the heating energy consumption of two kinds of users reduces. That is to say, as the outdoor temperature decreases, the consistency of the behavior of the two kinds of users increases. Conversely, the difference of the energy-saving behavior increases as the outdoor temperature increases. The reason for this is that the users whose heating fees are charged based on metering are more willing to save energy by adjusting thermostat valve frequently. However, because the users whose heating fees are charged based on area have nothing to do with heating energy consumption; they are not motivated to save energy, especially in the cold period.

Effect in buildings with different insulation properties

Based upon the Energy Efficiency Design Standards for Residential Buildings in Tianjin promulgated by Tianjin Urban Construction and Transportation Commission, the building typically built in early 1980s is regarded as a benchmark. Compared with the baseline building, the stage 2 energy-efficient buildings designed between 1997 and 2005 are expected to save 50% of the energy and the stage 3 energy-efficient buildings designed after 2005 are expected to save 65%. The two categories of buildings are the main objectives for the heat-metering policy to be implemented in China. Table 3 lists the specification requirements of the overall heat transfer coefficient threshold of the buildings<FootNote>

Tianjin Construction Management Committee. Standard DB29-1-2007 Tianjin Energy Efficiency Design Standards for Residential Buildings. Tianjin Urban Construction and Transportation Commission, 2007-06-01

</FootNote>.

To analyze the difference of user behavior energy-saving ratio in the two category buildings, the samples were divided into two groups in this paper based on the above regulations. The analysis of the above problems was conducted based on the heating statistics data in the heating seasons of 2011-2012. There are 1008 energy-efficient buildings in stage 2 energy-efficient buildings, of which 427 buildings were the ones whose heating fee was charged based on metering while 581 buildings were the ones whose heating fee was charged based on area. There are 718 buildings in stage 3 energy-efficient buildings, of which 375 buildings were the ones whose heating fee was charged based on metering while 343 buildings were the ones whose heating fee was charged based on area.

Figure 11 depicts the energy-saving rate of user behavior in buildings with different insulation properties. It can be observed from Fig. 11 that the heating energy consumption per unit area in stage 3 energy-efficient buildings, no matter the heating fee is charged based on metering or based on area, is lower than stage 2 energy-efficient buildings. However, there is a great difference between the users’ energy-saving rates of two types of buildings. The heating energy consumption of users whose heating fee are charged based on metering is lower than that of users whose heating fee are charged based on area by 15.55% in stage 2 energy-efficient buildings. However, the rate is only 6.45% for stage 3 energy-efficient buildings. For user, the main reason for this difference is that the stage 2 energy-efficient buildings have a greater energy saving potential. Therefore, the incentive effect of the policy on energy-saving behavior for users in the buildings with general thermal performance is more obvious than users in buildings with better thermal performance.

Conclusions

The heat-metering policy did play an important role in saving energy regarding occupancy behavior even though it has only been implemented for a few years. The users’ energy-saving behavior rate increased steadily in four consecutive heating seasons from 2008 to 2012. This indicates that users’ energy saving awareness has grown greatly to the point that saving energy has already become a habit due to the implementation of the incentive heat-metering policy for several years.

The heating energy difference between users whose heating fee were charged based on metering and whose heating fee were charged based on area shrinks as the outdoor air temperature drops, from 11.52% during mild cold season to 7.88% during extreme cold period. When it becomes colder, the two kind of users’ behavior tend to be the same. From Fig. 10 it can be seen that when it becomes warmer, the users whose fees are charged by metering are more willing to save energy. The users, motivated by the heat-metering policy, adjust the thermostat timely or take other positive method to save energy.

The effect on the user of energy-saving behavior under the policy of heat-metering is different among buildings with different thermal performance. The rate of behavior energy-saving of users living in better thermal performance buildings is 6.45% while it is 15.55% for users living in less insulated buildings. This indicates that the implementation of the heat-metering policy in buildings with general thermal performance gets better energy-saving result.

It can be concluded from this paper that the effect of the occupancy energy-saving behavior on heating energy consumption cannot be ignored. In those regions where the heat meters are installed, especially in mildly cold regions or in regions where residential buildings have poor thermal performance, the demonstration and publicity of the policy should be strengthened.

References

[1] Building Energy Research Center (BERC). Annual Report on China Building Energy Efficiency in 2012, Tsinghua University. Beijing: China Architecture & Building Press, 2012 (in Chinese)

[2] Liu L, Fu L, Jiang Y, Guo S. Major issues and solutions in the heat-metering reform in China. Renewable & Sustainable Energy Reviews, 2011, 15(1): 673–680

[3] Joachim W. Handbook of Heat Metering in Germany. Beijing: China Architecture & Building Press, 2009 (in Chinese)

[4] Branco G, Lachal B, Gallinelli P, Weber W. Predicted versus observed heat consumption of a low energy multifamily complex in Switzerland based on long-term experimental data. Energy and Building, 2004, 36(6): 543–555

[5] Ouyang J, Hokao K. Energy-saving potential by improving occupants’ behavior in urban residential sector in Hangzhou City, China. Energy and Building, 2009, 41(7): 711–720

[6] Haas R, Auer H, Biermayr P. The impact of consumer behavior on residential energy demand for space heating. Energy and Building, 1998, 27(2): 195–205

[7] de Meester T, Marique A F, de Herde A, Reiter S. Impacts of occupant behaviours on residential heating consumption for detached houses in a temperate climate in the northern part of Europe. Energy and Building, 2013, 57: 313–323

[8] Xu B, Fu L, Di H. Field investigation on consumer behavior and hydraulic performance of a district heating system in Tianjin, China. Building and Environment, 2009, 44(2): 249–259

[9] Santin O G, Itard L, Visscher H. The effect of occupancy and building characteristics on energy use for space and water heating in Dutch residential stock. Energy and Building, 2009, 41(11): 1223–1232

[10] Xu P, Xu T, Shen P. Energy and behavioral impacts of integrative retrofits for residential buildings: What is at stake for building energy policy reforms in northern China? Energy Policy, 2013, 52: 667–676

[11] He P. Some methods of deleting inordinate values from measuring data. Aviation Metrology & Measurement Technology, 1995, 15(1): 19–22 (in Chinese)

[12] Bickel P J, Doksum K A. Mathematical statistics, Oakland: Holden-Day, 1977

RIGHTS & PERMISSIONS

Higher Education Press and Springer-Verlag Berlin Heidelberg