1、NA-04-5-4 (RP-1172) Online Domestic Hot Water End-Use Database Dale K. Tiller, D.Phil. Gregor P. Henze, Ph.D., P.E. Member ASHRAE Xin Guo ABSTRACT Thegoal ofASHRAE Research Project II 72 has been the development of methods and tools to characterize total domes- tic hot water (DHW) consumption accord
2、ing to individual end use. An online web-enabled database entry and query system has been developed to provide worldwide access to project results. This paper describes the structure and function of this web-enabled database entry and query system. INTRODUCTION In the eleven years since its 1991 deb
3、ut, the World Wide Web has revolutionized the sharing and management of infor- mation. Booking an airline ticket, finding the telephone number of a friend in another city, and comparison shopping are all much easier today than they were a decade ago. Web- enabled database entry and query systems pro
4、vide the back- bone for these tools. This paper describes the structure and function of a web-enabled database entry and query system, developed under the auspices of ASHEUE Research Project 1 172, that provides statistical summaries of measured domes- tic hot water use. Two methods have been develo
5、ped to characterize DHW consumption, theflow trace signature analysis method and the temperature-based event inference method. The flow trace signature analysis method is well-documented (DeOreo et al. 1996; Fanney 1990; Hiller 1998, Lowenstein and Hiller 1996, 1998). This method allocates water flo
6、w to individual end uses based on the unique “flow trace signature” that occurs when hot water is drawn at different end uses (e.g., the duration and timing of measured flow will be different for dishwashers, clothes washers, faucets, and showers). Lowenstein and Hiller (1998) extended the flow trac
7、e signature analysis method by supplementing measured flow at the hot water tank with measurements of pipe temperature at strategically located branches in the hot water pipe network. This supplementary information helped better discriminate concurrent hot water draws, which were difficult to discri
8、minate using only flow trace signature. While additional temperature measurements at strategic locations in the pipe network helped to more accurately allo- cate water flow, Lowenstein and Hiller (1998) noted that temperature measurement points located at the hot water draw location itself would hav
9、e provided more specific localization of end use. The temperature-based event inference method (Henze et al. 2002) adopts this strategy, llocating hot water draws to specific end uses (e.g., dishwasher, clotheswasher, faucets, etc.) based on the change in pipe temperature that occurs at the end use
10、when hot water is drawn. Henze et al. (2002) attached a fast-responding temperature sensor to the pipe at each end use, measuring pipe temperature every second, while a flowmeter measures flow from the hot water tank. As hot water flows to specific end uses, the pipe temper- ature at respective end
11、uses where water is drawn will rise sharply. The change in pipe temperature provides an indirect measure of where in the pipe network hot water is being drawn. Two papers provide technical details on the analytical methods and software tools that have been developed to allo- cate domestic hot water
12、draws to specific end uses (Henze et al. 2002; Tiller et al. 2003). Henze et al. (2002) reported results from a month-long pilot study of a single residence that compared flow trace signature analysis and temperature-based event inference. Temperature-based event inference was able to correctly allo
13、cate 97.1% of the observed hot water draw events to different end uses, while flow trace signature analy- Dale K. Tiller and Gregor P. Henze are associate professors, and Xin Guo is a graduate student in the Architectural Engineering Department, University of Nebraska-Lincoln, Omaha, Neb. 02004 ASHR
14、AE. 682 - sis was able to correctly allocate 90.6% of hot water draws to different end uses. Although the specific method used to allocate hot water draws to end use is incidental to the development of an online database, the ability of database queries to resolve and summa- rize usage trends for in
15、dividual end uses will depend on whether the data collection and analysis method can allocate hot water draws to specific end uses. The database described in this paper has been populated with DHW usage data collected using the temperature-based event inference method, but any data collection and an
16、alysis method can be used, as long as the summaries of each hot water use event follow the same structure as used in the main database table (described below). This paper describes the structure and function of the web-enabled database entry and query system that has been developed to provide access
17、 to results from project RP-1172 and serve as a repository for other DHW usage data. The paper will first outline the generic structure of web-enabled database systems. This will be followed by a description of the DHW usage data that populates the database, the rationale behind including these spec
18、ific data, and an overview of the analysis methods and approach that are used to produce these data. Finally, the paper concludes with a description ofthe functions and queries currently supported by the database. INTRODUCTION TO WEB-ENABLED DATABASE SYSTEMS Web-enabled database entry and query syst
19、ems provide the backbone that allows websites to provide and display dynamic content. Web browser software sends a request to a remote computer, then retrieves, interprets, and displays a relatively small file consisting of text and layout instructions encoded in “hypertext markup language” (otherwi
20、se known as HTML). None of the dynamic interactivity offered by web- enabled database entry and query systems is possible with plain-text HTML. Three additional software components are required to add dynamic database entry and query, as follows. embedded in HTML, the functions of which are to commu
21、ni- cate with the database itself). Database server software is also available for all current operating systems: different database server products support a wide range of functions and capa- bilities. Query Module The web and database servers do not generate the dynamic content that is displayed a
22、s a web page. Since HTML only specifies the layout and appearance of text and other elements in a .browser window, it has no direct method of communicating with the database server and contents. The query module is the link that supports communication and information exchange between the static HTML
23、 web pages and data contained in the database. The static HTML pages that send queries to the database contain additional embedded instructions that have been written to communicate with the database. Any computer software code that has been written to query a specific database can be called by the
24、embedded instructions, as long as the query results that are returned by this program meet the standards for correct display in the web browser. Although custom computer software can be written for database query and display, most websites that incorporate database functions rely on commercially ava
25、ilable query modules. Query modules are also available for all current operating systems, and different query module products support different functions and capabilities. Some query modules include graphic capabilities (presenting results of database queries in the form of histograms or pie charts,
26、 for example). Graphics capability is also available using third- party software, which ifused must itselfbe configured to func- tion with the specific query module being used. Web-Enabled Database System Figure 1 depicts a conceptual overview showing how these three components interact in any web-e
27、nabled database, including the system developed for project RP-1172. The web Web Server A web server is required to process web page requests received from remote users and return the appropriate files to be displayed in the web browser on the remote computer. Web server software is available for al
28、l current operating systems. Some popular web server software packages are available for free download from the Internet, and numerous other commer- cial web server applications are available. Any internet service provider (ISP) that hosts web pages provides this component. Not all ISPs support the
29、next two key components. Database Server A database server is required to support web-enabled database entry and query. The database server hosts data sets and responds to queries that are passed from other programs (such as web pages containing additional coded instructions Figure 1 Overview of Web
30、-enabled database entry and query system. ASHRAE Transactions: Symposia 683 server (1) responds to requests from remote browsers. If the request includes embedded code that communicates with a database, the code is passed to the query module (2), which in turn passes the inshuctions to the database
31、server for process- ing (3). The database server returns the results to the query module. The query module integrates this information and returns formatted HTML containing text, any calculated statistical quantities, and associated graphics back upstream to the remote browser. The contents of a dat
32、abase can also be updated using these tools, following at least two methods. A web page displaying a static HTML form containing the appropriate data entry fields can be updated manually by a remote user (similar to filling out a form at any website). The information on the form is checked for integ
33、rity by the query module and then passed to the database server, which appends the new information to the database. Alternatively, batch mode update is also possible and is preferable when large amounts of data are involved that would not be practical to enter manually. In this case, datafiles are u
34、ploaded to the website hosting the database, as directed by a static HTML webpage. The query module checks the integrity of data contained in the files and passes the new data to the database server, which appends the new information to the database. The web-enabled database developed for RP-1172 in
35、cludes and requires all three components (web server, data- base server, and query module). While a specific set of commercially available software tools have been used to produce the functionality described in this paper, it is impor- tant to note that a similar system could be developed and deploy
36、ed using other commercially available tools. The DHW usage reports described in this paper are based on data collected over the fall of 200 1 and winter of 2002 from residences located in greater Omaha, Nebraska. These resi- dences differed in age, occupancy, and the number of hot water end uses in
37、each residence. The number of days of available data varied for each residence, as follows: residence 1 - 161 days; residence 2 - 1 10 days; residence 3 - 150 days; residence 4 - 131 days. In total, 552 days of data were included in the database. The next section provides an overview of the data acq
38、ui- sition and analysis methods that were used to collect these data. More complete details are provided by Henze et al. (2002) and Tiller et al. (2003). OVERVIEW OF DATA ACQUISITION, ANALYSIS, AND TRANSFER Temperature-based event inference allocates measured total hot water flow to specific end use
39、s based on the change in end-use pipe temperature that occurs at the end use when hot water is drawn. Type-T thermocouples were attached to each hot water end use in each house studied. These wires were passed back through the residence to a personal computer (PC) based data acquisition management a
40、nd control system. A flowmeter was installed at the water pipe leaving the hot water tank in each residence, and the signal from the flowmeter was also connected to the data acquisition system. The data acqui- sition system interface from the computer to the input signals from the thermocouples and
41、the flowmeter used the Universal Serial Bus (USB). The data acquisition system polled each measurement point (end-use pipe temperature and flow) every second, resulting in about 18 megabytes of daily data. Bandwidth limi- tations make it impractical to transfer this much data on a daily basis. Raw d
42、ata files are analyzed on the data acquisition system itself using a custom software program. This software program applies the temperature-based event inference method to the raw data file, identifiing instances where measured flow is accompanied by increases in pipe tempera- ture. The program defi
43、nes each such instance as a hot water use event and writes a small text file describing the duration and flow volume for each draw event represented in the larger data file. This file, called the “detection file,” is described in more detail in the next section. Henze et al. (2002) and Tiller et al.
44、 (2003) present results that establish the reliability and validity of the method and associated software tools to correctly allo- cate hot water draws. The detection files must be transferred from the remote site to a single computer before they can be uploaded to the database. This automated proce
45、ss occurs daily, using the task-scheduling features of the host computer that is acquir- ing the data. A dial-up connection to the Internet is automat- ically initiated in the middle of the night, and the data files are e-mailed to an Internet POP3 (Post Office Protocol Version 3) server, using a co
46、mmand line e-mail client. The detection files are automatically retrieved from the Internet POP3 mail server the next morning, using the same command line e-mail client but with a different set of command line parameters that direct the program to retrieve e-mail instead of sending it. The command l
47、ine e-mail program is directed to automatically place any attached data files into a specific directory on the receiving computer: since the e-mail client used in this work can automatically attach and detach the files to a specific directory, both the data send- ing and retrieval processes are full
48、y automated. Once the detection files have been retrieved from each site, they,are manually uploaded to the database, using a pass- word-protected web page at the website. This is the only step in the process that requires human input, but it could also be automated. Once the data have been uploaded
49、, all queries to the database report the most current data contained therein. DHW DATABASE End-Use Taxonomy The detection file that describes daily hot water draws uses a specific set of keys that together compose a taxonomy of DHW draw events and end uses. A mutually exclusive and exhaustive list of descriptors was developed to describe all possible DHW end uses and locations, using a triplet of indi- vidual keys, as follows. 684 ASHRAE Transactions: Symposia The first key describes the End Use Type, and it has the following possible values: DW = dishwasher; CW = clothes washer; SK = sink;
