1、4735 Study of Typical Meteorological Years and Their Effect on Building Energy and Renewable Energy Simulations H. Yang, Ph.D. Member ASHRAE ABSTRACT This paper investigates the generation of typical meteo- rologicalyears (TMY) and example weatheryears (EWs) for Hong Kong, and studies their efects o
2、n the simulation results of the performance of building energy and renewable energy systems, i.e., solar andwindenergysystems. According to vari- ous methodologies, dijerent TMYs and EWYs were calculated using Hong Kongs weather data from the past 22 years. The results were used for a building energ
3、y simulation andfor a hybrid solar-wind energy simulation, as two case studies, to study their efect on the simulation results. To validate the eflect, deviations of thesimulation results fordiferent methods were compared. The results show that the diference could be very sgnijcant: -20% for hybrid
4、solar-wind enew systems and a relatively smaller diference for building energy systems, *5%. A larger error was produced by using EWYs compared with TMYs for both the building energy simulation and the hybrid solar-wind energy simulation. This proves that gener- ating the right TMY is important for
5、meeting different needs and various application systems. INTRODUCTION The performance of environmentally driven systems, such as HVAC systems, solar thermal systems, solar power systems, greenhouses, and wind generators, is dependent on weather variables such as solar radiation, dry-bulb tempera- tu
6、re, wet-bulb temperature, wind speed, etc. However, these variables are neither completely random nor deterministic. In order to simulate the energy performance of an existing system or predict the energy performance of a new system during the design stage, suitable representative weather data are r
7、equired. Here is where the concept of a typical meteorological year L. Lu (TMY) comes in. A TMY provides a standard for hourly data on solar radiation and other meteorological parameters for a period of one year, representing climatic conditions consid- ered typical over a long period of time. The p
8、roperties of a TMY include: (1) the meteorological measures of the TMY (Le., solar radiation, temperature, and wind speed) should have frequency distributions that are “close” to long-term distributions; (2) the sequences of the daily measures for the TMY should in some sense be like the sequences o
9、ften regis- tered at a given location, which is often referred to persistent structure; and (3) the relationships among the different measures for the TMY should be like the relationships observed in nature. The most popular method for deriving TMY, first devel- oped by Hall et al. (1 979), is an em
10、pirical approach selecting individual months from different years using the Filkenstein- Schafer (FS) statistical method (Filkenstein and Schafer 1971). The final selection involved examining statistics and the persistent structure of daily dry-bulb temperatures and daily total global solar radiatio
11、n. Other studies (Pissimanis et al. 1988; Lam et al. 1996; Petrakis et al. 1998; Merter and Arif 2000; Zhang et al. 2002; Thevenard and Brunger 2002a, 2002b) derived TMYs for different cities. In these studies, different weighting factors of meteorological parameters were considered, while the persi
12、stent structure of weather data was neglected. Considering higher weighting factors of total solar radiation and dry-bulb temperature, the results from these studies are suitable for building energy simulations. But the persistent structure of weather data has to be considered in assessing solar ene
13、rgy systems. The above results are not suit- able for wind energy systems because they have little to do with wind. Lu and Yang (2002) developed the TMY for Hong H. Yang is an associate professor and L. Lu is a Ph.D. student in the Department of Building Services Engineering, The Hong Kong Polytechn
14、ic University, Hong Kong. 424 Q2004 ASHRAE. Kong, which can be used to assess applications of solar and wind energy. Their work also proved that determining proper weather parameters and their weighting factors is imperative during the development of TMY for different kinds of systems. TO study the
15、effect of differently developed typical mete- orological years and example weather years (Wong and Ngan 1993) on simulation results or assessment results, typical meteorological years and example weather years for Hong Kong were developed using different weighting factors of meteorological parameter
16、s as proposed by different refer- ences. With the derived TMYs and Ems, two case studies, namely, on building energy simulations and solar-wind hybrid systems (renewable energy), were established to find out the effect. Huang (1 998) studied the impact of different weather data on simulated resident
17、ial heating and cooling loads in some cities or locations ofthe U.S., using differently developedtypi- cal weather year data. In Hong Kong, two kinds of typical weather year data, namely, a typical meteorological year (Lam et al. 1996) and an example weather year (Wong and Ngan 1993), have been deve
18、loped for building energy simulations. According to their different weather parameters and weighting factors, the developed TMYs and EWYs are different, which affects the results of system simulations or assessments. This study considers the impact of different TMYs and EWYs on building energy simul
19、ations and solar-wind hybrid energy system assessments. DEVELOPMENT OF EXAMPLE WEATHER YEAR (EWY) AND TYPICAL METEOROLOGICAL YEAR (TMY) Methodology The Finkelstein-Schafer (FS) statistical method is usually used to compose TMYs and EWYs to achieve close perfor- mance of frequency distribution betwee
20、n past weather data and statistical data by comparing the monthly or yearly prob- ability of distribution function (PDF) with the long-term PDF in the month or year concerned. Considering the characteris- tics of renewable energy systems, four weather parameter indices are cited, namely, dry-bulb te
21、mperature, dew-point temperature, wind speed, and global solar radiation. Here, the hourly indices are compared by their distribution function, not by their maximum, minimum, andmean functions. The month or the year with minimum weighting sum (WS) value should be the optimal TMY month or EWY year in
22、 terms of close frequency distribution. The FS statistics are determined by M 1 WS(y,m) = - M WFjxFSj j=l 1 wq=lJ j= 1 M FS(j,m) statistics for the jth parameter of index 2j O, means the candidate year and m the candidate month) monthly (m) PDF value atX(i) value for the jth parameter index in the y
23、ear y monthly (m) long-term PDF value at X(i) value for the jth parameter index total number of different4 values, which is determined by the index values starting point, ending point, and increasing interval; for different indices, N is different according to relative requirements number of the hou
24、rly weather parameter indices, M = 4 average weighted sum for the month m in the year y weighting factor for the jth parameter The weighting factors are crucial for choosing TMY and EWY data from past weather data. According to five different criteria-namely, from Hall et al. (1979) Hall, Marion and
25、 Urban (1995) MU, ASHRAE (1997) ASHFUE, Wong and Ngan(1993) WN, andLuandYang(2002) LYI-theEWYs and TMYs were developed with corresponding weighting factors proposed from the five studies, as shown in Table 1. The former four studies mainly concentrated on building energy simulations with higher weig
26、hting factors of solar radi- ation and temperature (two crucial factors that affect building energy simulations), while Lu and Yangs (2002) criteria were suitable for solar and wind energy applications with higher weighting factors of solar radiation and wind velocity (crucial factors that affect so
27、lar and wind power generation systems). Weather data from the Hong Kong Observatory for a 22-year period (1 979-2000) were used in the analysis, including hourly dry-bulb temperatures, dew-point temperatures, wind velocity, and total global solar radiation for the location of 22“ 18 N and 114“ 10 E.
28、 Selection of “Typical Meteorological Year“ A typical meteorological year consists of twelve typical meteorological months. The month with the lowest weighting sum (WS), defined as the average weighted sum of all mete- orological parameters with certain weighting factors, should be chosen as the fir
29、st-ranking typical meteorological month. The persistent structure of weather parameters (especially solar radiation and wind speed) will affect the reliability and probability of wind-solar systems. The persistence has a significant effect on the results of renewable energy applica- tion assessments
30、 but no (or insignificant) effect on building energy simulations. Considering the sequence and persistent ASHRAE Transactions: Research 425 Table 1 - Weighting Factors Given to Weather Variables for Different Approaches* Weather Variable TMY Hall TMY MU TMY ASHRAE EMY WN Max. dry-bulb temperature 1
31、124 1/20 511 O0 114 Min. dry-bulb temperature 1 I24 1/20 5/100 Mean dry-bulb temperature 2/24 2/20 30/100 Max. wet-bulb temperature 1 I24 1/20 2.511 O0 1 I4 Min. wet-bulb temperature 1/24 1/20 2.5/100 Mean wet-bulb temperature 2/24 2/20 51100 Maximum wind speed 2/24 1 120 51100 Il4 Mean wind speed 2
32、/24 1/20 5/100 Total solar radiation on horizontal surface 12/24 5/20 40/100 114 Direct normal solar radiation - 5/20 - Hall means the approach using the weighting factors from Hall et ai. (1979) MU means the approach using the weighting factors from Marion and Urban (1995) ASHRAE means the approach
33、 using the weighting factors from ASHRAE (1997) WN means the approach using the weighting factors from Wong and Ngan (1993) LY means the approach using the weighting factors from Lu and Yang (2002) structure of weather parameters, the optimal months of the chosen ranks are also calculated based on L
34、u and Yangs method. Typical meteorological years for Hong Kong are generated according to the different methods, as shown in Table 2. The selected candidate months from different approaches are dissimilar. TMYs from Hall et al. (1979) and Marion and Urban (1995) consist of the same typical months, e
35、xcept for September, because these two approaches used the same weighting factors. The developed TMYs are also different according to different criteria, which have impact on energy performance assessments or predictions of energy system performance. The weighting factors of meteorological param- et
36、ers play an important role in the deriving of suitable TMYs for different systems. Different weighting factors affect the selection results of the typical meteorological months. If the results from Lu and Yang (2002) are taken as a refer- ence, the approaches from Hall et al. (1979) and Marion and U
37、rban (1 995) have the biggest deviations of WS because the smaller weighting factor of wind is considered in their studies, as shown in Figure 1. The results from ASHRAE (1 997) and Wong and Ngan (1993) have smaller deviations. From this point of view, the TMYs derived by ASHRAE (1997) and Wong and
38、Ngan (1993) are relatively better for assessing renewable energy systems. Selection of “Example Weather Year” The example weather year is compiled by applying the various methods to entire years instead of to individual months. According to the definition, different example weather years are selecte
39、d from the 22 years worth of weather data from Hong Kong, and the deviations of WS for different methods are shown in Figure 2. Figure 2 shows that the results of WS from Hall et al. (1 979) and Marion and Urban (1 995) are similar, with the same deviations, while the deviations from others vary bec
40、ause of the different given weighting factors of weather parameters. This proves that the weighting factors for different parameters are critical to the selection of EWY. Using different weighting factors leads to different results of EWYs. The five candidates of example weather years (in ranks) are
41、 selected according to the five different methods, as shown in Table 3. The year 1989 is chosen as the first-ranking candi- date according to the ASHRAE method but only the fourth- ranking using the approaches of Hall et al. (1979) and Lu and Yang (2002). TWO CASE STUDIES FOR DIFFERENT TMYS AND EWYS
42、 The simulation results of building energy simulations and solar wind power generation systems are usually affected by local weather data. To validate the influence of different derived TMYs and EWYs on building energy simulations and hybrid solar-wind applications, two case studies, namely, buildin
43、g energy simulations and solar-wind hybrid systems (including solar systems, wind systems, and hybrid solar-wind systems), were investigated. A Case Study for Building Energy Simulation A multi-function, nine-story, concrete, modem commer- cial building, with a total floor area of 102,400 m2, includ
44、ing a supermarket, cinema, coffee shops, restaurants, shops, a clubhouse, etc., was simulated for the energy consumption of its air-conditioning system under Hong Kongs weather condi- tions. The simulation package, HTB2 (Heat Transfer in Build- ings, Version 2), was applied for the simulation work a
45、s an investigative model of the thermal performance of the build- ing. With humid and hot weather year-round, only a cooling load was considered in this study. 426 ASHRAE Transactions: Research Table 2. Candidate Months Selected Using Different Proposed Methods _ Ranks Years Selected Jan Feb Mar Apr
46、. May June July Aug Sep Oct Nov Dec Five candidate years selected using the weighting factors by LY 1 St 1997 1995 I999 2000 1997 1994 2000 1986 1998 1989 1985 1996 2“d 1999 1996 1984 1993 1981 1995 1989 1996 1996 1998 1989 1993 3rd 1995 1985 1998 1979 1995 1984 1988 1995 1995 1993 1996 1992 4th 200
47、0 1998 1995 1999 1984 1979 1995 1985 1980 1987 1984 1987 5 1980 1997 1997 1980 1991 1992 1992 1997 1986 1982 1997 1989 Five candidate years selected using the weighting factors by Hall 1995 I998 1999 1919 1981 1994 1989 1986 1998 1998 1985 1996 2nd 1999 1995 1998 2000 1997 1995 2000 1996 1980 1984 1
48、989 1993 3rd 1997 1990 1984 1980 1984 1992 1990 1989 1996 1996 1997 1982 4th 1981 2000 1993 1993 1998 1991 1986 1984 1995 1989 2000 1983 5th 1980 1982 1995 1992 2000 1993 1995 1985 1981 1982 1990 1990 Five candidate years selected using the weighting factors by MU 1 St 1995 1998 1999 1979 1981 1994
49、1989 1986 1980 1998 2985 1996 2nd 1999 1982 1998 1980 1997 1991 2000 1989 1998 1984 1989 1982 3rd 1997 1990 1984 2000 1984 1992 1990 1996 1996 1996 1997 1983 4th 1981 2000 1993 1993 2000 1995 1998 1984 1995 1989 2000 1993 5h 1980 1995 1992 1992 1998 1993 1986 1992 1981 1982 1990 1990 Five candidate years selected using the weighting factors by ASHRAE 1 st 1995 1998 1998 2979 1997 1994 1989 1986 1980 1984 1985 1993 2“d 1997 1985 1999 i980 1981 1991 2000 1996 1996 1989 1989 1996 3 rd 1999 1982 1993 i993 1984 1995 1990 1989 1998 1998
