|
Solar Thermal Guide |
|
|
Affordable Solar Hot Water Heating
Colorado School of Mines and iCAST
Senior Design Team Spring 2008
Department of Engineering
Team Members: Allen Baybayan, Stephen Pronovost, Tully Gallagher, Troy Widener, Aaron Martinez
Team Faculty Advisor: Jeff Will
Interested in affordable solar hot water heating? Do you have a do-it-yourself mindset and can scavenge pieces here and there? If you answered yes to these questions then you must read ahead…
Our Senior Design team was able to create an affordable and efficient means of capturing the sun’s energy to heat water for home usage. For approximately $500 out of pocket expenses and some know-how, we were able to efficiently capture the sun’s energy to save ourselves $11* a month on water heating. That’s a savings of $132* per year with an estimated payback of 3.3* years!
For your convenience we have posted what we did, how we did it, and our theoretical analysis for you to consider. For a few days worth of work you too can get this system built and start saving yourself some money.
*Based on 2008 numbers. These numbers are estimates since actual monthly savings will vary with hot water usage and availability of the sun
Introduction
Our client – iCAST (International Center for Appropriate & Sustainable Technology)– is concerned about the cost and payback of present solar hot water heating systems. Our team was tasked with finding an efficient and economic solution that would make capturing the sun’s thermal energy more affordable and attractive to the average household income. Our objective was to design a system that is cost effective (less than a 7-year payback) and affordable (less than $500 out-of-pocket costs to the homeowner). By designing for these two constraints we hope to make solar hot water heating more attractive to the average homeowner.
Our team initially considered three different collectors and delivery systems. After our initial analysis of calculated efficiencies and payback periods we decided to pursue a traditional flat panel design due mostly in part its simplicity. A closed loop delivery system was chosen to supplement this panel type which will allow the capability of year-round operation.
We were able to keep out-of-pocket costs to $450* and the estimated payback of this design is approximately 3.3 years which equates to ~$136/year savings (based on a 41.5% efficiency). We then determined that at this efficiency we could’ve spent up to $950 on our system and still fall within our 7-year payback period. In order to keep our costs down we had to scavenge as much materials as we could. We were fortunate enough to have donated to us: the double pane glass for our collector, wood used to build the collector, the PIC microcontroller chip and programming board, and a gas hot water tank. If it weren’t for those donations our costs for this system would’ve have been much higher.
*Does not include cost of storage tank, collector glass, wood used to build collector, and testing materials
Flat Plate Collector
For our solar panel design we went with a flat plate collector. We constructed a single 81”x 37” panel out of a 4’x 8’ sheet of plywood and 2 10’ lenghts of 2x4. We then ran ~89’ of 1/2” copper tubing (each riser was 74” long and each end piece was 1.5” long) inside of the panel in a serpentine ‘S’ pattern to capture the heat and send it to the heat exchanger (Figure 1). The copper tubing is strung together via 90 degree elbows in an array of 14 copper tubing risers. The copper tubing (as well as the whole panel) was painted flat black to ensure that the absorptivity of the solar radiation on the metal is as high as possible.
We were fortunate enough to get a 35”x 80”x 1” double pane piece of glass to seal the top of our collector. We estimate that the donated double pane glass saved us ~$270 (this is assuming a cost of $14/ft2 of one-inch thick double pane glass) off of our construction costs which made a huge difference in our payback period. The benefit of double pane glass over a single pane is that it reduces heat loss from inside the collector by insulating the outside temperature from the inside temperature with a space of air like a thermos bottle or your oven door.
When the panel was all put together we sealed each seam with silicon sealant to ensure that minimal amounts of heat escape through the seams. We also bought some insulating foam and placed it in the bottom of the panel to help trap the heat and minimize heat losses within the panel.
Figure 1 – Flat Plate Collector with Serpentine Copper Tubing for Heat Transfer
Figure 2 – Larger View of Flat Plate Collector with Serpentine Copper Tubing for Heat Transfer
Step by Step Instructions for Building the Panel
1. Measure dimensions of glass panel. Our panel was approximately 35”x 80”.
2. Add two inches to each dimension of the glass and cut sheet of plywood to these dimensions. We cut our sheet of plywood so it measured 37”x 82”.
3. Cut two lengths of 2x4 equal to the long length of the sheet of plywood. We cut two 82” lengths of 2x4.
4. Cut two lengths of 2x4 equal to the length of the short side of the sheet minus 3. We cut two 34” lengths of 2x4.
5. Use table saw or router to rip slits for glass panel on each of the 2x4 lengths. Make depth of the slits ½”. The glass panel will fit into these slits when the collector is sealed. We used a table saw to rip 1” thick, ½” deep slits down the length of each 2x4.
6. Attach the two long lengths of 2x4 and one short length of 2x4 to the sheet of plywood by screwing through the bottom of the plywood sheet up into the lengths of 2x4.
7. Cut three lengths of wood equal to the inside width of the collector. These lengths of wood should be ¾” thick. The copper manifold will sit on these lengths of wood inside the collector. We cut three 34” lengths of wood that were ¾” thick and 1-1/2” wide.
8. Cut a 2” wide piece of insulation the length of the inside of the collector. We cut a 2” wide piece of insulation that was 34” in length.
9. Butt the piece of insulation up against the inside of the short end of the collector and attach it to the sheet of plywood using the silicon sealant.
10. Take one length of wood cut in Step 7 and butt this up against the insulation and attach with silicon sealant. The 1-1/2” wide side should be facing up.
11. Measure to the middle of the collector and another piece of insulation that will fit into this area. Butt the insulation up against the wood strip of Step 10 and attach with silicon sealant. Our piece of insulation measured 36” x 34”.
12. Take a second length of wood cut in Step 7 and butt this up against the second piece of insulation and attach with silicon sealant.
13. Measure to within at least 5” to the end of the collector and cut a final piece of insulation to fit this area. Butt up against wood strip of Step 12 and attach with silicon sealant.
14. Take final length of wood cut in Step 7 and butt this up against the third piece of insulation and attach with silicon sealant.
15. Use the remaining silicon sealant to seal the edges of the insulation to the sides of the collector panel.
16. Paint the entire collector black using flat black spray paint and/or flat back enamel paint. Make sure to use oil-based flat black enamel paint.
Now on to the manifold…
17. Cut lengths of copper pipe equal to the length of the plywood sheet minus 8. Our lengths of copper pipe were 74” in length. Keep in mind that you may have to piece some lengths together using the ½” copper couplings. Cut as many lengths as can be evenly spaced ~2” apart inside the collector. We used 14 lengths of pipe in our collector.
18. Cut lengths of copper pipe equal to 1.5” in length. The number of long length pieces will be one more than the number of 1.5” pieces. We had 14 long lengths, so we had 13 1.5” pieces of pipe.
For those who have never sweated copper pipe…Don’t worry, it’s not hard. We had never done it before, but managed to get the hang of it rather quickly. Refer here or here for instructions on sweating copper piping.
19. Sweat the ½” copper elbows to the ends of each of the 1.5” pieces of copper tube. Make sure that the free ends of the elbows are parallel with each other. We had 13 1.5” lengths, so we needed 26 elbows.
20. Use the ½” copper couplings to piece together any long lengths of tube. We had some shorter lengths of pipe (i.e. – 34” and 40”) that we connected using the couplings to piece together a couple 74” riser lengths.
21. Sweat the long lengths of pipe to the short lengths with the elbows connected to them forming an S-shape serpentine manifold.
22. Finally, sweat elbows to the two ends of pipe that have nothing attached to them. These two risers will be the inlet and outlet of the collector. Make sure to sweat the elbows to they are both perpendicular to the rest of the manifold (i.e. – when the manifold is lying on the ground, these two elbows should be pointing directly up.
Now that the manifold is finished, back to the collector…
23. Lay the manifold on the collector and mark places where the two elbows of Step 22 touch the bottom of the collector.
24. Drill 5/8” holes at each of these spots.
25. Sweat ~3” long pieces of copper tube into each of the elbows from Step 22.
26. Take the manifold outside and spray paint black. It may take several coats.
27. Lay the manifold back down on the collector and put the pieces of tube from Step 25 into the holes drilled in Step 26.
28. Attach the manifold to the wooden strips from Step 7 using ½” pipe clasps and ¾” or 1” long screws.
29. Attach thermistor temperature sensor to the manifold near one of the elbows attached in Step 22 using electrical tape or some other adhesive.
30. Drill small hole in bottom of collector and run wiring to the thermistor. Wire up thermistor.
31. Slide glass panel into the slits of the collector sides.
32. Slide the remaining end side piece of the collector cut in Step 4 onto the collector so that the glass panel fits into the slit of this length of wood.
33. Attach the final end piece to the collector using screws and screwing up through the bottom of the collector.
The collector is ready to go!
Closed-loop Delivery System
For our delivery method we chose a closed-loop delivery system that uses a 50% mixture of food grade propylene glycol as a heat exchange fluid. This type of system allows us to more effectively store the energy from the sun. The closed loop system with food grade propylene glycol will provide for year-round operation which gives us the opportunity to capture more heat year-round. The closed loop system (see Figure 1) consists of a heat exchanger in the pre-heat tank, a pump, and a microcontroller in addition to the water heating system already in place. We used a 115 VAC, 1/35 horsepower, 3250 rpm Taco pump (005-F2). The microcontroller will measure and analyze the temperature in the collector and the temperature of the water in the preheat tank. The microcontroller will turn the pump on and off, depending on whether we are able to capture enough heat in the collector to transfer to the storage tank for a temperature gain.
Figure 1 – Closed-Loop Delivery System
Figure 2 – Pump and Water Input Head Tank
However, our heat exchanger turned out to be almost impossible to implement in our storage tank. Routing the heat exchanger was an excruciatingly painful process since we were given a gas heater tank. There is no easy way to get the 1/4” flexible copper tubing into a gas water heater and what we had to do is route the flexible tubing through the top of the tank (where the sacrificial anode normally is) out the bottom outlet of the tank and hope to find the other end to pull out of the tank. This will transfer the heat gained from the collector panel to the water in the preheat tank. When the gas heater requires water in the inlet, the pre-heated water will be at a higher base temperature requiring the gas heater to use less energy. If we were given an electric gas heater we would’ve been able to pull out the heater core and easily install the heat exchanger.
Figure 3 – Pre-heat Storage Tank/Heat Exchanger
For our heat exchange fluid we decided it would be appropriate to use a 50/50 food grade propylene glycol and water mixture. This will keep the collector from freezing and will give us the option of running the closed-loop system year-round without worrying about freezing and drain-back. Also, the food grade propylene glycol doesn’t require a double-walled heat exchanger since it will not poison the drinking water. For testing purposes we used city water as our heat exchange fluid.
Some closed-loop systems use a drainback or an expansion tank but we opted not to use either to keep our costs down. The drainback tank isn’t necessary since we don’t have to worry about freezing and the expansion tank was dismissed since our fill tank can accommodate any fluid expansion.
Figure 4 – Flat Plate Collector and Closed-Loop Delivery System
Microcontroller
The objective of the microcontroller is to turn the pump on when there is sufficient heat in the collector to send to the heat exchanger. When the temperature of the fluid is at an acceptable level, the pump is turned on to circulate the antifreeze through the collector and the heat exchanger. In order to perform this task we used a PIC 18F4580 microcontroller. Click here for the code that we used. The microcontroller was programmed to read in 5 different voltages and convert them to temperatures. The microcontroller will log and analyze these data points to determine if the pump should be turned on.
The microcontroller and developmental kit can be found here. The compiler software that was used for this project is extremely expensive. An alternative that may be used can be found here. Microchip offers "MPLAB" free of charge, but this is an assembler, not a C compiler. There is a demo complier (good for 60 days) offered free of charge. The PIC18F4520 development kit has everything else that will be needed. The chip is nearly identical to the chip used for this project. For the PIC18F4520 datasheet, including pin assignments, click here.
For the temperature sensors we used five NTC 5k-ohm thermistors. The NTC (Negative Time Constant) thermistors will decrease in resistance as temperature increases. The 5k-ohm rating for the thermistor is the resistance that the thermistor will be at 25°C (this is known as R25).
Figure 1 - Thermistor
Steps to Obtain Thermistors
1. Thermistors can be found online in many different web catalogs such as Digi-key, Newark, and Omega just to name a few.
2. We bought our thermistors from Newark (part number 90B6326).
3. There are many different types and ratings for thermistors. You should choose which type will work best for your application. Click here to choose from multiple types and ratings of thermistors.
4. Though we used Newark you should shop around to get the best deal. These are relatively inexpensive but if you could find them locally or get free shipping that would end up saving you some cash.
Steps to Calibrate Thermistors
1. Whichever distributor you decide to use, they should have the datasheet available for the thermistor of your choosing.
2. To build our calibration curve we used the table of temperatures (°C) vs. Rt/R25 values (see Figure) given here RL1004-2910-97-D1.
Figure 2 - Temperature vs. Rt/R25 table found on thermistor datasheets
3. From the datasheet we copied the Temperature and Rt/R25 values for 0°C to 100°C into a spreadsheet (we used Excel).
4. We then solved for the Rt (Rthermistor) value. Our R25 value was 5k-ohm so we simply multiplied our Rt/R25 value by 5000.
5. With the Rt value we can now use a simple voltage divider equation to solve for our voltage out vs. temperature
6. For our source voltage we used 5 volts dc and our R1 resistance was 4.7k-ohm.
7. To solve for Vout we used the following voltage divider equation
1. Vout = 5Vdc*Rt/(Rt+4.7kO)
2. You can look at our thermistor data spreadsheet to see how we did it
8. Once you have all the values for Vout you can create a graph and plot Temperature on the y-axis and Vout on the x-axis.
9. With the graph built you can now use a trend line or best fit line to solve for your equation that relates temperature to Vout (see Figure 3)
1. Thermistor resistance vs temperature is highly non-linear so we had to use a 6th order polynomial to get a close fit to our data.
10. With the equation obtained you are now able to read temperatures from the voltage output of the voltage divider circuit
We considered plotting certain temperature points vs. Vout but realized that the relationship between resistance and temperature is very nonlinear, especially at high and low temperatures (see Figure 3). So to be safe and not assume linearity or extrapolating to an assumed value we used the given data (GE Type Number RL1004-2910-97-D1) to build our calibration curve.
Figure 3 - Thermistor Temperature vs Vout
Wiring the Microcontroller Circuit
For the 5Vdc power supply we used an old cell phone charger that rectifies 120Vac to 5Vdc. We initially used a 9V battery with a 5Vdc voltage regulating transistor (LM7805) but realized that the battery didn’t last long. The thermistors will change in resistance as temperature changes thus changing the voltage outputs of the voltage divider cicuit (See Figure 4).
The temperature sensors were placed in five strategic locations to give us a good peripheral of how temperatures are flowing throughout our system (see Figure 5). The sensors were placed inside the pre-heat tank, on the inlet and outlet of the heat exchanger, inside the flat plate collector panel (on the outlet pipe), and one on the inlet of the collector panel. For a larger view of our wiring diagram click here.
The temperature sensors on the tank inlet and the collector outlet can show us any line losses that may occur. For our collector outlet temperature we mounted the thermistor in the collector panel on the outlet piping to monitor both the collector temp and the outlet temp at the same time. The thermistor that we were primarily concerned about was the one inside the pre-heat tank since that is the one that shows us how hot we are able to get the preheat water up to. In order to get our sensor in there and to keep it from rusting we used a metal epoxy and mounted it to the old thermostat probe. This seemed to work really well.
Figure 4 - Voltage divider/temperature sensor circuit
Figure 5 - Microcontroller Wiring Diagram
System Modeling /Initial Calculations
In order for us to justify our design we did some preliminary calculations and modeled the amount of energy we would be able to collect out of our flat plate collector. We took the annual average daily solar radiation for Colorado given by NREL (see Figure 4 and 5) and used that to calculate our predicted collector efficiency and payback.
Figure 1 – Average daily solar radiation per month for a flat-plate collector facing south – From NREL
Figure 2 – Colorado Daily Average Insolation per Month
NREL provides tabulated date for the insolation of an average day of each month. This data is plotted in Figure 10 and is what we used to perform our calculations. To provide ourselves with an accurate payback calculation we had to factor in the daily usage of hot water for an average family of four. This means that we had to account for hot water usage during certain hours of the day to see how our collector could meet that demand. For instance, hot water typically gets used in the early morning hours but there is no sun available so the water would have to be heated by conventional means. In order to accommodate this we built a daily usage spreadsheet to forecast the load that we’d be able to meet (see Figure 6).
The calculation of the approximated efficiency of our system involved several variables.
Es – the energy from the sun incident upon the collector (Insolation)
Er – the energy from the sun that was reflected away by the glass on the collector (Insolation Losses)
Ec – the energy lost from the collector due to convection (Collector Losses)
El – the energy lost from the insulated copper pipes between the collector and the storage tank (Line Losses)
Et – the energy lost from the tank due to convection (Tank Losses)
Ew – the energy collected in the water in the storage tank (Energy Saved)
Eu – the energy in the water that had been used (Gallons Used)
Es was found from the NREL data plotted in Figure 10. This data was tabulated on hourly increments so all our calculations were also done on hourly increments. Each data point from the NREL data was multiplied by the square footage of our collector surface area (Ac). Therefore, Es = Ac(data point)
Er was found from the reflection coefficient of our glass. For double paned glass, this value was 20%, which simply means that 20% if the incident energy is reflected away. Therefore, Er = .2Es.
Ec was dependent upon the U-value of the collector glass, the U-Value of the collector’s insulation backing, the temperature inside the collector and the temperature outside the collector. U-value is the thermal conductivity, and in the case of glass and insulation these values are tabulated by the manufacturer. We used double paned glass with a half inch gap between the two panes. The U-value (Ug) that we used for double paned glass was 1.08 Btu/(ft2 hour F°). The U-Value (Ui) that we used for the ½” thick foam insulation was .26 Btu/(ft2 hour F°). The energy lost from the collector was found by adding the two U-values together and then multiplying the result by the square footage of our collector surface area (Ac) and the difference between the temperature inside the collector (Tic) and the temperature outside the collector (Toc). Therefore, Ec = (Ug + Ui) Ac (Tic – Toc).
El was dependent upon the Line-Lose coefficient of the ½” insulated copper pipe (LL), the length of the copper pipe (Lp), the difference between temperature inside the copper pipe (Tip) and the temperature outside the copper pipe (Top). The Line-Lose coefficient is a tabulated value which for ½” insulated copper pipe is .089 Btu/(ft hour F°). Therefore, El = LL (Tip – Top) Lp.
Et was dependent upon the U-value of the storage tank (Ut), the surface area of the storage tank (At), and the difference between the temperature inside the tank (Tit) and the temperature outside the tank (Tot). We used a hot water heater tank to store our water and a typical R-value for a hot water heater tank is 14 (ft2 hour F°)/Btu. But, to do our calculations, we really want a U-value (for consistency mainly), and you may notice from the units that Ut = 1/R, so the U-value that we used was .07143 Btu/(ft2 hour F°). Therefore, Et = Ut At (Tit – Tot).
Eu was found by approximating how much water was used by a family of four and at what time of day it was used. When water is used, it is drawn from the tank at a temperate Tit, and an equal amount of water flows into the tank at a temperature Tg ˜ 50 F°. The energy removed from the tank as a result of usage is the volume of the water that was used (V) times the density of water (?) times the specific heat capacity of water (C) times the difference between the temperature of the water that was used and the water that flowed into the tank. Therefore, Eu = V ? C (Tit – Tg). This energy is necessary for keeping track of the temperature of the water in the tank, but it does not play a role in the efficiency calculation because it is not lost energy. The whole purpose of this system is to provide hot water that can be used.
The law of conservation of energy let us find out how much energy made it into the water in the storage tank. Therefore, Ew = Es – (Er + Ec + El + Et + Eu).
Each of these variables were calculated on an hourly basis in a spreadsheet daily usage spreadsheet, and the calculation was somewhat of an iterative process because the temperature of the water was dependent upon the energy stored in the water and the energy stored in the water was dependent upon the temperature of the water. So, we ran the calculations once with the initial water temperature (1:00AM) of 50 F°, and then reset the initial temperature to equal the final temperature (12:00AM) and ran the calculation again. This process was repeated until the final temperature remained equal to the initial temperature. This resulted energy in the stored water as a function of time which directly correlates to the temperature of the stored water as a function of time. The temperature of the water as a function of time is plotted in Figure 11.
Figure 3 - Daily Temperature Curves of Stored Water
The efficiency (e) of the system is the energy stored in the water (Ew), without the usage term (Eu), divided by the energy from the sun incident upon the collector (Es) times 100%.
Therefore e = [(Ew – Eu)/Es]*100%. This efficiency varied throughout the day and throughout the year so we calculated an average efficiency for our system for each month, and then over a full year, which turned out to be 48%.
So, we went through a lot of work to calculate this approximate efficiency. The next step was to test our system and compare the empirical data with the model that we worked so hard to create.
Test Results/Data Collection
To obtain the voltages below, a multi-meter was used and voltages were read across different leads (see Microcontroller for more information). Table 1 describes the different voltages and their associated temperatures and different points in our system.
Table 1 – Test #1 Temperature Data of our System
| Time |
Collector Inlet (Volts) |
Collector Inlet (F°) |
Collector Outlet (Volts) |
Collector Outlet (F°) |
Tank Bot (Volts) |
Tank Bot(F°) |
Tank Inlet (Volts) |
Tank Inlet (F°) |
| 9:30 |
2.87 |
63.34955041 |
2.65 |
71.53767927 |
3.08 |
56.71195225 |
3.08 |
71.1394778 |
| 10:00 |
2.05 |
97.34955041 |
2.02 |
99.30453443 |
2.94 |
61.0096216 |
2.94 |
111.237845 |
| 10:30 |
1.94 |
102.8356738 |
1.85 |
106.7013902 |
2.89 |
62.66767282 |
2.89 |
115.9280185 |
| 11:00 |
1.87 |
105.8534798 |
1.69 |
113.2250126 |
2.79 |
66.18344811 |
2.79 |
119.6427095 |
| 11:30 |
1.78 |
109.6139641 |
1.56 |
118.175727 |
2.64 |
71.93816621 |
2.64 |
123.2273586 |
| 12:00 |
1.69 |
113.2250126 |
1.5 |
120.3679828 |
2.55 |
75.63961455 |
2.55 |
124.6419209 |
| 12:30 |
1.69 |
113.2250126 |
1.43 |
122.8727145 |
2.43 |
80.808627 |
2.43 |
125.701848 |
| 13:00 |
1.64 |
115.1649358 |
1.47 |
121.447084 |
2.34 |
84.81513153 |
2.34 |
125.3483871
|
| 13:30 |
1.64 |
115.1649358 |
1.56 |
118.175727 |
2.24 |
89.34317656 |
2.24 |
125.701848 |
| 14:00 |
1.64 |
115.1649358 |
1.64 |
115.1649358 |
2.19 |
91.61972335 |
2.19 |
124.9951212 |
| 14:30 |
1.64 |
115.1649358 |
1.69 |
113.2250126 |
2.15 |
93.44016832 |
2.15 |
124.2886618 |
| 15:00 |
1.64 |
115.1649358 |
1.71 |
112.435857 |
2.13 |
94.34882386 |
2.13 |
124.2886618 |
| 15:30 |
1.6 |
116.684005 |
1.71 |
112.435857 |
2.08 |
96.61216847 |
2.08 |
123.5814933
|
Figure 1 shows these temperatures on a graph for our collection time (9:30 – 3:30, April 23rd.) As the Figure 1 shows, we were able to have a temperature rise of around 40o F.
Figure 1 – Test #1 Temperature Data of our System
Now that we had collected some data, it was time to compare it with the model that we had worked so hard to create. The data we collected was on half hour increments from 9:30 in the morning to 3:30 in the afternoon on the 23rd of April, 2008. Our model data was calculated on one hour increments and so we had to pick out a block of time from our model that matched. We chose the block of model data from 9:00 AM to 4:00 PM, and because the data was in one hour increments we interpolated to fill in the data points on each half hour. Also, the test data started at 9:30 not 9:00, and ended at 3:30 not 4:00 so we interpolated to fill in those data points as well. The interpolation simply served to smooth out the two curves for plotting purposes. The data is tabulated in Table 2 with the interpolated data points in bold. TTest is the measured value of the water temperature in the bottom of the tank (also Tank Bot in Table 1 and Figure 1) and Tmodel is the temperature in the tank that we calculated in the spread sheet (daily usage spreadsheet ) for the month of April.
Table 2 – Test vs. Model Data
|
Time
|
TTest (F°)
|
TModel (F°)
|
|
AM 9:00
|
52.4
|
59.7
|
|
9:30
|
56.7
|
64.7
|
|
10:00
|
61.0
|
69.7
|
|
10:30
|
62.7
|
75.6
|
|
11:00
|
66.2
|
81.5
|
|
11:30
|
71.9
|
87.4
|
|
PM 12:00
|
75.6
|
93.3
|
|
12:30
|
80.8
|
98.2
|
|
1:00
|
84.8
|
103.2
|
|
1:30
|
89.3
|
107.9
|
|
2:00
|
91.0
|
112.6
|
|
2:30
|
92.7
|
116.1
|
|
3:00
|
94.3
|
119.7
|
|
3:30
|
96.6
|
121.7
|
|
4:00
|
98.9
|
123.8
|
The test data and the model data for the temperature of the water in the storage tank from Table 2 is plotted in Figure 2. The test temperature was a little lower than the model temperature, but it wasn’t bad for an analytical model; off by a factor of 1.25 at the most.
Figure 2 – Test vs. Model Data for April
The next step was to compare the actual efficiency of our collector with the efficiency that we calculated in our model. From our model (daily usage spreadsheet) our average efficiency over the whole year came out to be about 48%. We also calculated an average efficiency for each month, and the average efficiency for the month of April came out to be about 45%. So, all we needed to do was find the actual efficiency of our system. To do this we used data from NREL for the same day that we did the test (23 April, 2008) shown in Figure 3.
Figure 3 – Solar Insolation data from NREL for 23 April, 2008
The area under the curve in Figure 3 from 9:30AM to 3:30PM is equal to the total number of joules/m2 incident upon our collector during the test; multiplying by the area of our collector gave us the total energy that was incident upon the collector. The energy that actually made it into the water in the storage tank was found by E = V ? C (Tf – Ti), where V was the volume of the water in the tank, ? was the density of water, C was the specific heat capacity, Tf was the final temperature of the water at 3:30PM, and Ti was the initial temperature of the water at 9:30AM. These calculations came out to be 37.87 MJ incident upon the collector and 15.72 MJ into the storage water. The efficiency was (15.72 MJ)/(37.87)*100% = 41.5%.
So, the actual efficiency of our test run in April was 41.5%, compared with the model efficiency of 45% for the month of April again shows that our model was to far off.
Proposed Optimization
When analyzing the overall performance of our system during the testing phase we noticed that the ¼” heat exchanger that we built into the storage tank caused the flow rate through the closed-loop section of the system to drop drastically. This reduction in flow rate is no doubt due to the constriction from ½” piping in the collector panel to ¼” tubing in the storage tank. We feel that solving this flow rate problem may help to increase system efficiency.
We tested just the collector panel portion of the system by taking the storage tank and heat exchanger offline. If you refer to Figure 4 on the “Closed Loop Delivery System” page, instead of running tubing from the outlet of the collector into the top of the storage tank, we ran the tubing right back into the orange bucket. Figure 1 below shows the results of this test. Notice that after only one hour of operation the Thermometer reading (the Thermometer was measuring the temperature of the water in the orange bucket) jumped from ~56 °F to ~105 °F. Furthermore, the temperature increased to a maximum of ~136 °F by 12:30 pm.
Figure 1 – Temperature Output when Bypassing Heat Exchanger
These findings, coupled with the flow rate problem, gave us the inspiration for a new proposed heat exchange system. Our recommended solution is to remove the heat exchanger from the storage tank and incorporate it into the orange bucket. In this case, the fluid flowing through the new heat exchanger would be the stored water in the tank. This would require purchasing an additional circulating pump, but we feel the pump needed for this circulation would be inexpensive to add to the system. In essence, the new heat exchanger would be submerged in the fluid that is continuously circulating through the collector panel. Refer here for an illustration of what this new small storage tank with heat exchanger would look like. Circles 12 and 14 represent the tank and internal heat exchanger. Using the data from Figure 1, it can be seen that this new heat exchanger would be sitting in fluid above 120 °F for a good portion of the day. Small storage tanks with built in heat exchangers are produced commercially, but their costs are in excess of $800. Click here for an example of a commercial tank w/ storage tank.
Of course, this recommendation would require the added complexity of designing and constructing a home-built small storage tank with internal heat exchanger, but we feel that using this type of heat exchanging system may increase the efficiency of the system.
List of Materials and Breakdown of Costs
Flat Plate Collector Cost Breakdown
All of the above materials with prices were purchased at Home Depot. We had the glass panel, plywood, and 2x4s donated to us. The cost of the wood should not be more than $25. The double-paned glass panel was part of a sliding glass door that was once in a restaurant. We would suggest looking for glass panels in junkyards or salvage shops.
Closed-loop Devlivery System Cost Breakdown
The storage tank that we used was a 40 gallon gas hot water heater that was donated to us. We recommend checking local salvage shops for old water heaters. However, make sure that the water heater does not leak. Another place to look for used water heaters is Craigslist (http://www.craigslist.org). We found people either giving their old water heaters away or selling them for less than $100. The Taco pump and flanges were purchased online at PexSupply (http://www.pexsupply.com/). Additional information about the pump can be found here. The data sheet for the pump can be found here.
Microcontroller Cost Breakdown
The 120Vac to 5Vdc rectifier that we used was an old cell phone charger that was donated by a team member. The wiring accessories consisted of terminal board, terminal leads, Velcro, electrical tape, and crimp-ons.
Overall System Cost: $450
Contact Us
Jeff Will - Team Faculty Advisor
Email:
This e-mail address is being protected from spam bots, you need JavaScript enabled to view it
Homepage: http://diamond.gem.valpo.edu
Aaron Martinez - Physics and ME (system modeling)
Email:
This e-mail address is being protected from spam bots, you need JavaScript enabled to view it
Stephen Pronovost - EE (microcontroller)
Email:
This e-mail address is being protected from spam bots, you need JavaScript enabled to view it
Allen Baybayan - EE (microcontroller and webpage design)
Email:
This e-mail address is being protected from spam bots, you need JavaScript enabled to view it
Links:
Colorado School of Mines
http://www.nrel.gov/solar/
Colorado School of Mines List of Senior Design Projects
|
|
|
Workforce Center Training
iCAST's Green Workforce Center (GWC) Staff Training is online with a discount for NAWDP members
This online training is based on a four-hour program and is designed to help educate workforce center employees on ‘green jobs’. Taught in an upbeat, interactive manner, .... Click here for more info.
Services
Click below to find out how iCAST can best serve your community:
|
