|
|
Solar Panel Project. Conditions applying to the use of this design for a solar
water heater: 1) Date of publication 3 October 2008. 2) No transfer of Copyright is implied or offered by
the author. Permission must be obtained from the author to copy any of the
text or images within this web page. 3) The design methods and construction solutions herein
described are for open public access and are freely available for use by the
home constructor. 4) These plans and designs are offered for the
non-commercial construction of a solar panel: no permission is given by the author
for the commercial exploitation of any of the design or engineering solutions
herein described. 5) In attempting to construct a solar panel you accept
that the author of this web site offers no warranty and that you will
construct a solar panel from these designs entirely at your own risk. Health Warning: It is absolutely important that you understand that this
solar panel can only be used as a pre-heater to a hot-water storage tank, and
that supplementary heating will always be used to establish a hot water
temperature in excess of 60ºC. This requirement is in order to prevent the
possible contamination of the hot water supply with the bacteria that causes
Legionnaires Disease (Legionella pneumophila caused by the bacterium
Legionella haemophila). Background: Power from the sun: The average amount of energy available from the sun in Britain is between 2 and 3 units per square metre per day, and depends upon your location in the country (see NASA’s web site: click here). A unit of energy, in this instance, is the kilowatt-hour (kWh): where 1kWh is that amount of energy consumed by 1000Watts over 1-hour. For example a refrigerator may consume somewhere in the order of 50W. This power is 0.05 units of electricity per hour or 1.2 units per day, or a little under 440 units of electricity per year. At the time of writing this article electricity is charged at around £0.12 per unit, therefore the refrigerator will cost about £52 per year to run. A simple calculation would indicate that every square metre of surface directly exposed to the sun receives more than 24p worth of equivalent electric energy per day in Britain (2 units times 12p). The energy available from sunlight varies over the year: midsummer can exceed 5 units per square metre per day, mid-winter as low as 0.5 units per square metre per day or lower. The overall annual average, however, is consistently between 2 and 3 units per square metre per day, and depends simply upon your location within Britain.
The chart above is representation of the average energy capturing capacity for the solar panel described here. The data in blue is that monthly average air temperature for Edinburgh (average air temperature for the month across night and day as detailed by NASA), and to a reasonable extent will reflect the temperature of the coldwater supply. The data in red is the theoretical average daily maximum temperature of 90-litres of water (being the typical capacity of a domestic header tank; more on this later) at the end of a day’s solar-harvesting. In practice, on any particular day and averaged across the month, the water temperature will be lower than these monthly average values because of thermal losses in the header tank and the pipe-work to the collection panel. The control circuit described here is designed to prevent the water temperature exceeding 40ºC (i.e. the pump cuts-off) and, assuming that at times during the day hot water will be used resulting in cold water replenishing that used from the header tank, should prove to be about right for a target peak header temperature in summer months. If you would like to read more of the science behind the design of this solar panel then click here. Beware of exaggeration: When calculating how much to spend on installing a solar panel it is the overall annual average return that should be understood. All-to-often commercial solar panel manufactures will quote the output power (not energy) for their panels at the very height of summer on a completely cloudless day with the panel oriented to directly face the sun. Under such a method of calculating the output power each square metre could produce nearly 1000W, however this very attractive figure hides the fact that the real world can only provide about one-tenth of this amount of power when averaged over a few days. The most obvious restriction on the available energy is the fact that the sun is only around during the daylight hours! Therefore the overall average, 24, 7, 365, is around 100W per square metre in the UK. Cost-v-Return: The return on investment in a solar panel is continuous: when averaged over the year, night and day, the return is more than 1p per hour per square metre, or around £100 per year per square metre. There is an economy of scale that applies to solar panel construction: a panel of any size requires one control circuit and one pump. The panel materials will increase with the size of the panel, but it is also the case that the cost of doubling the materials (copper pipe, aluminium plate, etc) does not necessarily double the cost. The solar panel I will describe here is 2 square metres and cost a little under £400 in materials. Therefore the average time to return the investment is around 2-years for electric equivalent, or around 4-years for gas equivalent, assuming a reasonable level of effort is put into insulating the header tank and connecting pipe-work, etc. Design overview: A solar-heat collection
panel measuring 2m x 1m is attached to a south-facing roof inclined at 50º
from the vertical. Ordinary reinforced garden hosepipe is used to carry water
from and to the cold-water header tank that, in turn, feeds the hot-water
tank. An immersion water-pump is used to circulate the water between the
header tank and the collection panel. Temperature sensors in both the collection
panel and the header tank control the operation of the water pump: if the
collection panel is at a higher temperature than the header tank then the
pump is switched-on, alternatively the pump is switched-off if the panel is
cooler than the header tank. The hot-water tank
must receive supplementary heating from either an electric immersion heater
or a central-heating boiler, and under all circumstances the temperature of
the hot water delivered to the taps, etc, must be 60ºC or higher. This is to
prevent the possible contamination of the hot water by bacteria causing
Legionnaires Disease. The collecting panel: A sheet of 2mm thick aluminium measuring 2m x 1m was purchased from e-bay for £88 including p+p. A 30m length of 12mm copper pipe (supplied as a single coil) was purchased from e-bay for £83 including p+p. 60 off 12mm copper saddles was purchased from e-bay for £10.75 including p+p. Blind pop-rivets suitable for fixing the saddles were obtained from B&Q for approximately £3 (as an alternative fixing method small self-tapping screws could be used, however using pop-rivets greatly speeds-up the construction process). Approximately 3m of 12mm reinforced garden hose, available at any garden centre. Two 12mm ‘Y’ pipe couplings, obtained from B&Q for around £5. Eight galvanised Jubilee Clips, available from B&Q for approximately £5. Four sheets of 3mm glass measuring 1m x 0.5m were obtained from a local glazier for £48. A tube of silicon sealing compound, ideally black or clear. A total of 9m of pressure-treated timber with a cross-section of approximately 38mm x 25mm, obtained from B&Q for around £15. A total of 6m of pressure-treated timber with a cross-section of approximately 10mm x 100mm, obtained from B&Q for around £15. Approximately 200 30mm stainless (brass or galvanised) woodscrews. A block of timber 25mm thick and measuring 56mm by approximately 100mm – I had an appropriate block in my garage. Polystyrene insulation boards approximately 50mm thick and 2 square-metres in area, obtained from B&Q for around £5. A tin of matt black paint – I used some Thompson’s Roof Seal (about £24 per tin) that I have lying around in my garage. Mounting the panel on the roof will be very specific to your particular circumstances, and I therefore do not suggest any one particular method. In brief the panel is constructed by bending the copper pipe into a series of horizontal lines and riveted to the aluminium sheet. A wooden frame is constructed around the panel, dividing the panel in to 4 equal areas; this frame will support the glass. A larger outer frame encloses the whole of the panel and holds the polystyrene against the back of the panel; this outer frame will also retain the glass. Holes cut into the side of the frame allows access for pipes and temperature sensor cable. When contemplating how best to bend the copper pipe it quickly became apparent that the bend radius would be too small for a run of single pipe. There would have to be sixteen rows of pipe, making the bend radius something in the order of 30mm: far too small a radius for 12mm annealed copper pipe! It was therefore necessary to fit the pipe in two 15m lengths and to have the pipe fixed to the aluminium plate as alternate lines resulting in a doubling of the bend radius. Clearly, where the pipe was bent there would have to be alternate overlaps, as the drawing below will show.
The first stage of construction is to fabricate a bending block from a piece of timber approximately 25mm thick. Cut and plane a semicircle, radius 56mm, onto the block. This block will aid the bending of the copper pipe: obviously kinking the pipe will render it useless for this project, therefore great care should be taken when bending the pipe! The copper pipe is usually supplied as a coil, and I found it quite easy to gradually uncoil the pipe as I fixed it to the aluminium plate. In the drawing above you can see that the copper pipe starts and ends approximately 600mm from the right hand side of the panel. Using garden hosepipe to connect the copper pipe to two ‘Y’ couplings that bring the two copper circuits together as the pipes exit the panel. Begin by marking lines on the aluminium where the copper pipe is to be fitted: 16 horizontal lines evenly spaced across the panel. With a short length of copper pipe uncoiled use a saddle to secure it to the aluminium plate (using pop rivets or short self-tapping screws) approximately 600mm from one side of the panel. Do not fix the first saddle too close to the end of the pipe; this is in order to allow the pipe to be slightly lifted from the aluminium plate as you push the hose onto the end of the pipe. Gradually uncoil the pipe and fix it along the marked line at regular intervals: I used four saddles per horizontal. At the end of each horizontal use the wooden bending block to carefully form a 180-degree bend in the pipe. Repeat this process until you have reached the bottom of the panel, and then cut the copper pipe at a point approximately 600mm from the side of the panel. Start the second run of pipe in the same manner as the first, only this time as you reach the end of a horizontal you will have to lift the pipe away from the aluminium plate to form the bend and secure it on top of the first run of pipe. As you begin the next horizontal carefully encourage the pipe back into contact with the aluminium plate. Continue with this process until you have completed the copper pipe attachment; with luck you should have only a small amount of waste copper pipe by the time you have completed the last horizontal. Cut four lengths of garden hosepipe each 700mm long and push each length over the ends of the copper pipe. Use jubilee hose clips to secure each length of hose to the ends of the copper pipe, and now start on constructing the frame. Using pressure treated timber of cross section approximately 25mm by 38mm construct the frame as in the drawing above. Drill and use screws to secure the frame to the aluminium plate, and remember that the timber has to support the glass above the copper pipe; 38mm should be sufficient space to allow this. It will be necessary to cut notches in the timber through which the pipe will pass and to drill holes for the ‘Y’ fittings and the sensor cable access. I secured the wooden components of the frame to the aluminium plate using screws between each horizontal copper pipe for the central 3 supports and every 200mm of the outer frame; in total approximately 86 screws were used for this section of construction. It will be necessary to bring together the two copper pipe circuits in to one pipe for water supply. Trim the four lengths of hosepipe to the optimum length allowing them to couple with the ‘Y’ fitting, ensure that the fitting passes through the frame sufficient to allow the attachment of the external pipes. Use jubilee hose clips to attach the hose to the ‘Y’ fitting. I used epoxy resin to secure the ‘Y’ fitting into the frame, but not before making doubly sure that the ‘Y’ fitting protruded sufficiently beyond the frame to allow fitting of the external pipe-work! It is advisable, whilst the panel is still reasonably manoeuvrable, to do some testing for leaks in the jubilee joints. The method I used was to prop-up the panel with it resting on one of its long sides (much as in the drawing above, but it’s not necessary for the panel to be fully vertical) and connect the lower ‘Y’ fitting to a mains-pressure water tap. The upper ‘Y’ fitting was connected to my garden hose, which has a jet spray head that can be adjusted to shut-off the water flow. By fully turning on the water tap and adjusting the jet spray head it is possible to bleed the air from the panel’s copper pipes. Once all the air has be removed from the panel’s pipes, and a steady jet of water is passing from the spray head, shut of the water at the jet spray head (remembering to leave the water tap on) and leave the panel for several hours. Should any leaks appear then tighten the appropriate jubilee clip. It is a good idea to make very certain at this stage that you don’t have water leaks; it’ll be a lot harder to attend to any leaks after the panel is fitted! Once you are very happy that the panel has no leaks, disconnect the water supply and fully drain the panel’s pipes. By this stage the panel is becoming rather heavy, and the next task specific to manufacturing the panel should be done only after the panel has been lifted into its final position on your roof, or wherever. As the final mounting location of the panel is very specific to your particular circumstances, I do not suggest any particular mounting method. However there are general requirements that apply no matter where you have in mind for the panel: it should ideally face due south (for northern latitudes) and be inclined at an angle equal to your latitude, i.e. Edinburgh is 56 degrees north, therefore the panel should be lifted from the horizontal by 56 degrees and face south. The polystyrene insulation has to be secured behind the panel by some means until the outer frame has been constructed around it. Using 100mm by 10mm pressure treated timber cut a length of 2m and hold it on the underside of the panel, as in the drawing below, and mark the timber for trimming to the required width.
It is important that the small shelf, see above drawing, protrudes sufficiently to allow the glass to be supported during the glazing procedure. Once trimmed to size, secure the timber to the underside of the panel using screws every 200mm and screwing into the wooden frame on the panel. Next add similar wooden trim to the other three sides: it isn’t necessary to have a shelf for the top and sides trim, therefore the timber can be cut to be flush with the frame. When cutting the trim for the side with the pipe and cable holes remember to cut suitable holes in the trim. It is advisable to cut the pipe access holes at a diameter of around 30mm, this will allow easy access to the ‘Y’ fitting when you come to fit the external pipe-work. It is at this stage that you need to address the temperature sensor fixing, this is covered later but by clicking your mouse here you can get a sneak preview. The next task is to paint the panel matt-black. I used “Thompson’s Roof Seal”, however any matt-black paint will do as long as you use a suitable primer for wood and aluminium. You want as deep a colour as possible, therefore apply several coats of paint. Once the paint is fully dry you can start the glazing. I found it easier to first mount the two centre sheets of glass; however depending upon your particular circumstance glazing might well have a different order. Use the silicon sealer to run an unbroken line of sealer around the frame that will support the first sheet of glass. Carefully place the glass on the support shelf and slowly offer-up the glass to the frame. The sealer will spread as the glass compresses the silicon; gently encourage this process until the glass is fully bedded home. Ensure that the glass is sealed throughout before starting on the next sheet of glass. Continue this process to complete the panel. The pump, pipe-work and collector sensor A “Hozelock Cascade 1500” was purchased from e-bay for £37 including p+p A Hozelock hose fitting Two (possibly more) galvanized jubilee clips Several metres of garden hose was purchased from a local garden centre Several metres of 3-core 3A electric cable was purchased from B&Q A few metres of pipe insulation was purchased from B&Q In brief the submersible pump is placed in the header tank that feeds the hot water cylinder, and hosepipe is used to connect the water outlet from the pump to the bottom of the collecting panel. Another length of hose is used to return the water from the collecting panel back to the header tank. A suitable route for running the pipes and sensor cable from the collection panel to the header tank will have to be identified: no particular method will suffice for all situations, it is therefore for you to identify the best route in your property. The advantage of using a submersible pump is that the electrical power it consumes will itself contribute to heating the water, whereas an external pump would simply loose its heat to the air! The pump I used consumes 20-Watts of electricity whilst operating, however this power consumption is dwarfed by the heat pumping capacity it provides: at full power the collection panel will generate nearly 2000 Watts of power, i.e. 100 times the electricity consumed by the pump! The temperature sensor in the collection panel (described in detail later) is connected via a length of 3-core cable to the electronic control circuit. This cable is routed along with the pipe-work from the collection panel to the header tank of your hot water system. The header tank also has a temperature sensor connected by another 3-core cable; it is therefore convenient to route both of these cables together back to the electronic control circuit. The pipe-work should be insulated, especially those section which are outside the building, using commercially available pipe insulation. Tip: The end of the water return pipe should be held beneath the water level in the header tank in order to prevent a “leaky cistern” sound of running water. However, in order to prevent a siphon-lock occurring, drill a 5mm hole in the return pipe about 50mm above the water level in the header tank, this will ensure that when the pump stops the collection panel and associated pipe-work will drain of water. The electronic control circuit The electronic control circuit takes in signals from the collection panel and the header tank to establish whether or not the collection panel is at a higher temperature than the header tank, and consequential to these measurements will switch the pump on or off. There are, however, three conditions under which this basic operation will be overridden: First: if the header tank is at a higher temperature than 40 degrees C the pump will be forced off irrespective as to whether the collection panel is hotter than 40 degree C. This level is adjustable and can be disabled, but is included within the function of the circuit to prevent the header tank reaching too high a temperature. I arbitrarily chose 40 degrees C for no other reason than gut instinct. Second: if the collection panel is at a lower temperature than 0 degrees C the pump will be forced on even though the collection panel is colder than the header tank. This level is adjustable and can be disabled, but is included within the function of the circuit should it prove not possible to completely drain the collection panel (as it is in my case) to prevent freezing the water within the panels pipe-work. Third: at intervals of approximately 15 minutes the pump is forced off irrespective as to whether the collection panel is hotter than the header tank. This is a necessary function should your panel’s heat transfer efficiency be very high: in circumstances where there the solar gain is less than the system losses the circuit should switch the pump off. However if the solar gain is only just less, then the panel -v- header-tank differential temperature may not drop sufficiently to switch the pump off. Under these circumstances by forcing the pump off on a regular basis a true indication of the energy gathering capacity can be established: the panel’s temperature will change and will indicate whether or not the pump was chilling or warming the collection panel. The schematic diagram below is of the control circuit that I used in my solar panel, it is available in kit form from myself at a reasonable price, and is based around the MCP9700 linear temperature sensor manufactured by MicroChip. The kit form of this circuit is available from myself and comprises a printed circuit board (PCB), a detailed description of the circuit construction and all of the components except the 12V power supply unit (PSU) and fuse, the pump, the relay and associated RC filter. These additional components will depend upon your specific circumstances, however the additional costs are minimal and the devices are readily available on the web. In detail: a quad-operational amplifier (quad-op-amp LT1079 manufactured by Linear Technology Corporation Ltd) is used to provide the functionality of the circuit. One quarter of the LT1079, an op-amp, is used as a simple level detector comparing the voltages from the two temperature sensors, in the event that the collection panel sensor is at a higher temperature than the header tank sensor the output of the op-amp goes from 0V to +5V. The output from the level detector op-amp is delivered to a Schmitt Trigger configuration with a hysteresis of 0.5V to remove any propensity to jitter. The Schmitt Trigger input is biased to 0V that, in this configuration, equates to pump-off. Two op-amps provide the override conditions discussed earlier, each is provided with a small amount of hysteresis to prevent jitter, one (the 0 degree C set level) forces the Schmitt Trigger to +5V and the pump-on, and the other (the 40 degree C set level) forces the Schmitt Trigger to 0V and the pump-off. The timing function is provided by an LMC555 manufactured by National Semiconductor and is configured as an astable with period of approximately 15 minutes and a duty cycle of 10%. Therefore for approximately 1.5 minutes in each 15 minutes cycle the pump is forced off. As discussed earlier, this function will enhance the detection of even the smallest tendency for the panel to be a power-sink. Care should be taken to keep the circuit board clean around the LMC555 and its passive components; as you can see from the schematic the impedances are very high, and any leakage current will have a detrimental effect on function. The temperature sensors and their associated biasing, decoupling, signal impedance-tailoring devices and cable should all be constructed within the same immediate space, i.e. rats-nested, and the whole dipped several times in silicon sealer. This is obviously of particular importance for the header tank sensor, which must be submerged in the water of the header tank. The sensor (and all of the bits and bobs mentioned in the last paragraph) in the collection panel needs to be adhered to the panel with epoxy resin and then liberally coated in silicon sealer. This is something that has to be done before glazing and painting the collection panel (If this is a sneak preview you can go back by clicking here). The reference voltage device is a laser-trimmed band-gap ZXRE1004 manufactured by Zetex and provides a reference voltage of 1.22V for the level detection op-amps. Although I have used a 7805 regulator to convert from 12V to 5V it is not necessary to use such a high power device as the circuit operation current will be less than 100mA, and to all intents and purposes limited to the relay coil current. Without the relay coil current the circuit consumes less than 1mA! All of the components used in this design are readily available from the likes of Farnell Ltd and CPC Ltd, and numerous other companies. The relay will need to be capable of switching mains voltages and have a current capacity that exceeds the current required for the water pump of your choice. The coil of the relay will need an energising voltage of 12Vdc and must have a coil resistance greater than 120 Ohms. Tolerance of the passive devices need be no better than 5%, and without exception the resistors can all be low power, i.e. 100mW. The 100nF capacitor used in the suppression filter needs to be rated at 1000Vdc, but all of the other capacitors can have rated voltages of around 20Vdc. The two electrolytic capacitors, used in the two first-order filters on the level detector op-amp, should be tantalum. A 12V PSU can be of the “battery eliminator” type – much like a mobile-phone charger – and many are available from the companies above: be sure to select a supply that is clearly stated as being regulated. It only leaves me to wish you good luck and Happy Solar Harvesting!
All of the content of this web page, the text, the graphics, the design and engineering solutions, are copyright to the author. Permission must be obtained to reproduce any of the content of this web page.
Some of the science behind the solar panel design deltaTm: This is a measure of the panel’s maximum performance, unobtainable in any practical sense, and is defined as “the maximum change in temperature of the panel when exposed to maximum insolation”. When not exposed to the sun the temperature of the solar panel will settle to the local atmospheric temperature. In other words, like any object in your garden (any object, that is, which doesn’t have an internal heat source) at night its temperature will be governed by the air temperature around it: if the object is warmer than the air then convection will reduce its temperature, or if the object is cooler than the air then convection will warm it, and in all cases will tend towards a situation where the object is at the same temperature as the air around it. The speed at which the objects temperature approaches local atmospheric temperature is called the chill rate: if the air is still (not windy) then it is only convection that equalises the temperature, but if the air is moving then the rate of temperature equalisation increases with the wind speed – this is known as the chill factor. For an object that has a heat source, such as a building - your home for example in winter - then the internal temperature is higher than the external air temperature and heat flows through the walls of the object to be convected away into the air. The rate at which this heat (energy) is lost can be managed by attending to the insulation properties of the object: your home should have loft insulation for example to reduce the rate of heat loss. In all circumstances to maintain an internal temperature which is higher than the external environmental temperature requires power to be dissipated internally, and the amount of power required is directly proportional to the difference temperature, deltaT, and the chill factor. It is also the case that if a certain fixed power is consumed (again we can use a building as an example), the internal temperature will rise until the heat that is lost through the structure is the same as the internal heat source. Your home will have a certain heating coefficient which could be described in terms of “so many degrees Celsius per unit of power”: for example a typical home will be 2degC per kW, from which you can calculate that amount of power required to make your home a comfortable temperature. Lets say that the outside winter temperature is 3degC and you like to maintain a room temperature of 21degC, then the deltaT is 18degC and the power necessary to maintain this difference is 9kW (18 divided by 2). During windy periods the chill factor will apply making the apparent deltaT larger than simply the temperature difference, and therefore the power required to maintain a comfortable internal temperature increases accordingly. For a solar panel deltaTm is that difference temperature that is achieved, without cooling water passing through the panel (and no wind), when exposed to a known amount of insolation power. For the solar panel described here the deltaTm was measured to be 36degC per kW. Measurements were taken on a windless and cloudless midday in midsummer with the panel directly facing the sun: the in-shade air temperature was 18degC with the insolation factor (as described by NASA) of 980W per square-metre and the panel is 2 square-metres: the maximum recorded temperature of the inside of the solar panel was 88.5degC. The measurement of deltaTm is of importance when calculating the energy capture efficiency of the solar panel: we know that the maximum insolation energy impinging on the panel is 1.96kW (i.e. 2 times 980W), but that it is not possible to access all of this power because of thermal losses within the structure of the panel. We could ask, for example, what level of power would have been intercepted on the clear sunny summer’s day when the measurement of deltaTm was made if the header tank water temperature was at 40degC (i.e. at the arbitrary cut-off temperature)? In this particular case the panel’s temperature just prior to cut-off is 40degC and the air temperature is 18degC, therefore the difference is 22degC, and from our knowledge that deltaTm is 36degC per kW we can calculate that the amount of power being lost from the panel is: 22 divided by 36 and then multiply the result by 1kW, and this equals just over 0.6kW. Therefore the power that is being transferred to the water in the header tank is 1.96kW – 0.6kW that is equal to a little under 1.35kW. Another way to look at this figure is in terms of an efficiency factor: to harvest this power requires the consumption of only 20W (to power the pump), therefore the efficiency is 1350 divided by 20 and the result multiplied by 100%, that equals 6750%; which is not bad at all! Click here to go to the beginning of the page Click here to go to A T Services Home Page Click here to go to Lever Long Home Page |
