1、NA-04-5-3 Numerical Evaluation of Alternative Residential Hot Water Distribution Systems Evelyn Baskin, Ph.D. Member ASHRAE Roberto Lenarduui, Ph.D. ABSTRACT Subsequent to space conditioning, domestic hot water is the second largest consumer of energy in American homes. Improvements have been made i
2、n the energy eficiency ofwater heating equipment, but fm gains have occurred in the distri- bution system S energyperformance. Energy and water waste associated with distribution system performance can be a significant fraction of the total expenditure of delivering and generating hot water. Improvi
3、ng hot water delivery systems will reduce the energy usage and system performance. Results from numerically simulating,four diflerent hot water distribu- tion systems in three diferent system locations in a conven- tional house reveal that system conjiguration and location as well as pipe material a
4、nd insulution noticeably impact the overall performance of the distribution system. Model results based on two usageprojiles reveal that: In most CPVC distribution systems, signijcantly more energy is wasted from previously heated water remain- ing in the pipe than heat loss through the pipe walls w
5、hile hot water is flowing. Changing the assumed hot water use patterns can dramatically change the perfor- mance of an individual system and its performance com- pared to copper systems. Demand recirculation systems reduce the waiting time for hot water as well as the energy and water waste, while c
6、ontinuous recirculation systems also reduce waiting and water consumption but increase energy consumption about 600%. Parallel pipe systems made of PEX reduce the waiting as well as energy and water waste compared with con- ventional branch and trunk systems. Robert Wendt Keith A. Woodbury, Ph.D. Ad
7、ding pipe insulation does not have a significant impact on heated water energy waste or piping heat loss for hot water delivery initiated from a “cold start but can be effective ifhot water uses are clustered in a short time frame. Distribution systems located in the attic (buried in attic insulatio
8、n) should not be insulated because it looses additional energy when standard pipe insulation is added. Copper pipes have -10 times more heat loss when they are located below thegoor slab in the soil (clay) for “cold start” compared with a clustered use hot water delivery. INTRODUCTION AND SIMULATION
9、 DESCRIPTION A numerical model of hot water distribution systems was developed that allows analysis of various types of piping configurations with and without insulation. The model was derived from a model by Stewart et al. (1999) that calculates and compares the hourly heat loss and exit temperatur
10、e of a copper or steel straight pipe (100 ft maximum) in still air. In this model, the systems may be exposed to a convection envi- ronment, buried in attic insulation, or buried beneath a floor slab. The distribution system model is Windows-based and versatile. The simulation is written in a progra
11、mming language with a graphical user interface (different program- ming language). The temperatures in the fluid, pipe, and insulation are calculated by applying a finite element technique to two heat transfer equations. The temperature distribution in the fluid (T(x,2) shown in Figure i) is simulat
12、ed by the one-dimensional energy transport in the axial direction of the piping system Evelyn Baskin is a research engineer in the Building Equipment Group, Robert Wendt is a research engineer in the Residential Buildings Group, and Roberto Lenarduzzi is a research engineer in the Analog and Digital
13、 Systems Group, Oak Ridge National Laboratory, Oak Ridge, Tenn. Keith A. Woodbury is an associate professor of mechanical engineering in the College of Engineering, University of Alabama, Tusca- loosa, Ala. 02004 ASHRAE. 671 Figure I Pipe and flow temperature distribution. with lateral heat losses t
14、o the pipe wall. The temperature distri- bution in the pipe wall and insulation, T,(v,x,t), is calculated using two-dimensional calculations, coupled to the one- dimensional pipe solution through a heat transfer coefficient. Mathematically the (axial) temperature distribution of the fluid is governe
15、d by wherep is the perimeter of the pipe, A, is the cross-sectional area, and k, cp, and p are properties of the fluid. The heat loss from the fluid to the pipe wall is modeled using a heat transfer coefficient stated as where T,(x,t) is the temperature of the surface of the pipe. The temperature di
16、stribution in the pipe and insulation is calcu- lated from the solution of the ho-dimensional heat conduc- tion equation in radial coordinates: (3) where the radial variation in k must be retained (to allow for insulation over the pipe) but the axial variation in k is ignored. T,(r;x,t), the solutio
17、n for the temperature in the pipe andor insulation and the temperature T,(x,t) in Equation 2 are equal at the pipe inner radius: Equation 2 is used to couple the solution for T,(cx,t) to that for T(x,t). The boundary condition on Equation 3 is an external convection environment that has a known refe
18、rence temperature: -k-l ar rz = h(T,(r,x,t) - T,) (5) The radiation is handled on the exterior by a radiant heat transfer coefficient expressed as Figure 2 SOIL/ATTIC material of thickness thick surrounding pipe/insulation. Piping systems surrounded by a large layer of attic insu- lation, soil, or c
19、oncrete slab are treated in the model as a finite radial thickness of the external material. This is basically the same as if the pipe (with or without pipe insulation) is further insulated with a thickness of attic insulation (piping buried beneath attic insulation), soil, or concrete (piping burie
20、d in soil underneath the slab). The condition is depicted in Figure 2. The layer of surrounding material is characterized by a thickness parame- ter, thick, and this thickness of material is assumed to be all around the pipe. The outer surface of the composite cylinder is assumed to be subjected to
21、a convective/radiative boundary condition. It is assumed that the simulation time is rather short, and the temperature on the outside of this large cylinder of added material will not change substantially during the simu- lation. Therefore, the solution will not be affected if one surface of the mat
22、erial is exposed to convection and the others are semi-infinite (as in the case of a buried pipe) or if one surface has convectionradiation to a lower temperature than the other (as for attic insulation). The outer radius boundary is assumed to be at a constant temperature during the operation of th
23、e hot water system. Both the constant temperature that is assumed and the radius of the material are user inputs. 672 ASHRAE Transactions: Symposia When there is no water flowing in the piping system, the model exercises one of the pure conduction options using the equations above. If the flow rate
24、is specified as zero, the initial fluid tem- perature is taken as that of the environment and the pip- ing is treated as a fin on the hot water heater. If the flow rate is specified as any value less than zero, a special calculation is performed whereby the initial fluid temperature is set equal to
25、the supply temperature or a specified temperature, and the heat loss during the cool- down is calculated. For no flow in the pipe, a new heat transfer coefficient accounting for the heat conduction from the fluid to the pipe was developed by using a correlation based on an analytical solution for he
26、at conduction in a solid cylinder that is subjected to a step increase in temperature at its surface. The no-flow heat transfer coefficient applies to all configurations. Heat loss is computed using the conduction equation (Equa- tion 3) plus the new heat transfer coefficient. During flow conditions
27、 all of the above equations are used. The time between a cluster draw is calculated as a no-flow cooldown of water in the piping between draws. The no-flow cooldown temperature is used as the pipe andor insulation surface temperature in the subsequent draw in the cluster. During hot water use, a sma
28、ll depth of the soil and attic insulation is pene- trated by heat and this same depth is affected during cooldown. Since the depth is small, it is not used when the cooldown piping surface temperature is calculated for the subsequent cluster draw. SIMULATION METHOD The model solves for the temperatu
29、re distribution in the water, pipe, and insulation along the length of the pipe as a function of time using a finite element technique capable of modeling various piping configurations, the entire piping layout, and hot water use events. The use events can originate from a cold start (water in pipe
30、at temperature of surround- ings-most wasteful) or determined by entering a daily use profile that defines the cooldown periods between events. The simulation can be used to do comparative studies, such as: Establishing the heat loss differences between different pipe materials Identifjing the impac
31、t of an insulated versus non-insu- lated pipe Calculating the effect of various pipe diameters on the outlet water temperature Calculating the waiting period for hot water to amve at a fixture. The simulation requires the following data to calculate the heat loss and outlet water temperature: the pi
32、pe parameters (length, inside diameter, and wall thickness); the pipe and insulation properties (thermal conductivity, specific heat, and density); the water flow rate; the insulation thickness; and the distribution system location. The model accepts spreadsheet inputs files, which define the daily
33、water use and events based on the physical properties of the pipe sections and usage pattern. Pipe and insulation property data are automatically selected based on pipe and insulation type as specified in the spreadsheet. Each pipe section is defined by five columns of data. There is a limit of 50 s
34、ections of pipe per event. There is no limit on the number of hot water draw events. It is assumed that all the sections in one event have the same pipe material. The time needed to calculate each section depends on the length and location of the piping, usage pattem, speci- fied simulation time ste
35、p (-1 second), and maximum simula- tion time. The output results of the simulation are stored in a spreadsheet file and then further analyzed. The model requires a usage file to determine the hot water use pattern and an input file to determine the layout of the system. APPARATUSES: PIPING LAYOUTS I
36、N SAMPLE HOUSE A typical home (one-story, single family, 2010 fi2, three bedrooms, and two baths) is used as the sample case to eval- uate the model results. The house represents a typical single- story house that contains a laundry room, one bath with a combined tub and a shower with two lavatories
37、, and another full bath with a tubhhower and one lavatory. The kitchen includes a sink and dishwasher. The water heater is in the garage, and the distribution layout spreads the hot water consuming devices (point-of-use) throughout the house. Figures 3 through 6 show the distribution system layouts
38、by the dark black line on the house plan. SIMULATION PERFORMANCE RESULTS All draws are initiated at the hot water heater at 120F (48.9.C) and terminate at the point-of-use. Results from cold start (ail water in the pipe is at the surrounding temperature- most wasteful case) simulation runs using the
39、 conventional, demand recirculation and parallel pipe distribution configura- tions are presented in Table 2. The parameters used for the sample house simulations are shown in Table 1. The temper- atures are a seasonal average for California. The wait time for hot water (the time it takes for the wa
40、ter to reach a usable temperature of 105F at the point-of-use) and heat loss through the piping to the surroundings for each point- of-use are determined based on the pipe and water initiate temperatures. The draw profile is performed for a day and the monthly result is computed by multiplying the d
41、aily use result by 30. The results reveal substantial piping heat loss and wait time differences based on the pipe material, configuration, insulation level, and system location as described below and shown in Figure 7. Pipe Material: CPVC pipe has significantly less (-50%) piping heat loss reductio
42、n through the pipe compared to copper pipe for conventional (standard and central water heater location) and demand recirculation configura- ASHRAE Transactions: Symposia 673 QRUT ROOM BEDROOM 2 LIVING I DININO QARAOE Figure 3 Conventional hot water distribution system. ORUT ROOM BEDROOM 2 LIVWQ I D
43、INlNQ GAMG Figure 4 Demand recirculation layout. 674 ASHRAE Transactions: Symposia QRUIT ROOM BEDROOM 2 UVINO I atama GARAGE n Figure 5 Continuous recirculation layout. Figure 6 Parallel pipe layout. ASHRAE Transactions: Symposia 675 Pipe Main Branch Copper 314 Location Temperature PVC 314 (OF) PEX
44、318 Diameter (inch) Attic 76 Crawl Space (CS) 68 Soil 64 Water Wasted (Water Down Drain) Wait Time for HW (sec) Sample House (System Layout) Typical Max (gallons) I Slab Cu I 63 I 111 I 932 I 551,223 I 117,388 1 Energy Loss From (Btu) Previously Heated Water Wasted Pipe Attic Cu - Central 42 44 436
45、Attic CPVC - Central 41 42 426 Attic Cu 60 103 882 Attic Cu-Ins 60 104 883 Attic CPVC 57 99 839 1 I I 1 I I I 257,697 26,728 251,851 13,509 521,391 53,097 522,140 54,079 496.43 1 29,394 I Demand Recirculation l Attic CPVC-Ins cs Cu CS Cu-Ins 57 99 839 496,43 I 29,668 60 104 892 527,612 62,054 60 104
46、 892 527,612 60,371 CS CPVC CS CPVC-Ins Ins-insulation, CS-cruwlspuce, Cu-copper. I IW-hut water. Typical = median wait time ola11 daily profile draws, and Max IS the maximum wait tinir. of a11 daily profile draws 57 1 O0 849 501,902 33,231 57 1 O0 849 501,902 32,977 tions, both with and without ins
47、ulation. Using CPVC causes an estimated 50% energy loss reduction below that of copper pipe as shown in Table 2. The typical hot water wait time is about 5% less than that of the stan- dard conventional copper system. Since water waste is proportional to the hot water wait time, the CPVC pipe has ro
48、ughly 5% less waste water than that of copper pipe. ConJiguration: Having the hot water heater centrally located-reducing pipe length from the water heater to point-of-use (water heater not in the garage-centrally Slab Cu-Ins 60 104 884 522,890 82,956 Slab CPVC 58 1 O0 855 505,950 38,839 Slab CPVC-I
49、ns 58 1 O0 855 505.950 34,291 located in the house)-yields an energy reduction of roughly 50% compared to that of the conventional sys- tem (with the water heater in the garage). The parallel pipe configuration using PEX pipe material reveals sim- ilar reduced piping heat loss (-25%) compared to that of the centrally located water heater, but the maximum hot water wait time is about half that of the centrally located water heater. The parallel pipe system has -35% less piping heat loss and typical waiting time than that of the standard conventional copper system pla
