Examining our previous electricity bills shows that we use about 22kWh of electricity per day. Most offline people consider that consumption to be outrageous.
If I ever get my electric car project going (which largely depends on my staying here, rather than moving to where I can walk to work) that figure will have another 10-20kWh per working day added to it.
Our electricity use can be divided into two parts: the fixed load of the items that are on 24/7, and the intermittent loads that are on part of the time. First I did some measurements of the fixed loads, and the result was that we consumed about 330W or 8kWh/day on fixed loads - of which about half is computers. Our computers consisted of two PCs which draw about 40W each, and an elderly Cisco switch which draws about 80W.
The rest of the house consists of various things: two fishtanks with airpumps; various things "on standby", like two PC monitors, a video, a microwave, a doorbell, a cooker and a boiler; and transformer boxes for phone chargers, shavers, and the like. There is also an interconnected fire alarm system which consumes a maximum of 18W for each of 12 units. Perhaps it's surprising that these add up to less than 160W, but they do.
The other load not measured is the 'fridge/freezer, which claims to use 527kWh/year. So that's another 60W, or 1.44kWh per day.
A programme of energy conservation has reduced the base load to about 100W, or 2.4kWh per day, mainly by replacing the computers with a dedicated ADSL router and switch and an old laptop instead of the rack mount PC.
The intermittent loads are mainly white goods and lights. The main cost of this is the white goods: we have a drier which my wife runs for about 3 hours a day: that's about 6 1/2 units a day. Washing and drying clothes and the dishwashers probably consume about 1/2 of our current electricity usage.
Replacing the lights with CFLs and using the washing line more has reduced our intermittent loads to maybe 10kWh per day.
Average figures for houses at our lattitude suggest that a 1970's-construction house will need about 13000kWh annually, while a modern house will need about 4000kWh annually. Most of that is in the winter months - so January might need 1000kWh (3000kWh) per month, or 33kWh (100kWh) per day, or an extra 1.4kW (4.2kW) for heating. With a little extra insulation, we might be able to get away with doubling the requirement for the winter months. However, energy for domestic heating is best stored in a thermal store, since these are a tiny fraction of the cost of an electrical store.
A detailed examination of the insulation possibilities for the house we're considering suggests an average winter heating bill of maybe 84kWh per day, assuming an average winter temperature of 5oC and an inside temperature of 18oC. A thermal solar panel system might provide 20% or so of that, but the electricity system is going to need to generate an average of 3kW surplus through the winter months. Fortunately, winter winds make this quite possible.
For offline power, the normal setup is to generate DC power using renewable sources, and then to store that power in a battery for when it's needed. If the power comes from wind and sun, then the worst case is when the wind is not blowing and the sun is not shining: eg a high pressure overcast. If we want storage for these conditions, we need to have a battery pack capable of delivering this power.
There are other energy sources available: one of the most attractive we've been considering is tidal. There are two tides a day, and a tidal generator can generate tides on both the ebb and flow: so that makes four times a day when the energy is available. Where we want to live - western Ireland - has one of the fastest tidal flows in the world, so a tidal generator might be feasable. However, none of the properties we are currently considering are suitable for tidal power.
Another source of power is streams. Falling water generates a lot of power, and it also has the advantage that the power is always available. Even a mill capable of providing that 100W base load would make a really big difference, especially to battery life. A mill capable of providing enough power to provide peak load would be great, since then the batteries would not be required, but streams like that are rare.
A number of sources of energy are considered on separate pages:-
The ones considered most suitable for this stage of planning are wind and sun.
So, to deliver our current energy requirements for five days of high pressure overcast, we need 5000Ah worth of 12v batteries. At the moment we have 2000Ah worth - enough for two days - but I want to put some of them in a car.
|Base usage(kWh/day)||Fridge/Freezer||intermittent(kWh/day)||Total||Min power(W)||Ave power(W)||1-day storage (Ah@12v)|
The maximum load can be calculated in a number of ways, but my personal inclination is to calculate the amount that a grid connection would offer. Our house has 2 ringmains giving 30A each, and two lighting mains giving 15A each. It should also have a cooker point giving another 30A, and there may be a dedicated connection to the car charging point for another 30A. So the total would seem to be 150A. However, the company fuses are typically rated at 40A to 100A. So an invertor that delivers 10kVA will probably be sufficient, and an invertor that delivers 16kVA would certainly be so.
The ideal arrangement is to have a continuous power of the maximum amount we need, 24/7 all year around. In real life, this will never happen.
A minimum arrangement is to have at least the average amount of power we use provided by the energy sources, averaged over a period of a few days as covered by the batteries. This means getting meteorological records for the local area.
Also, wind power is generally cheaper than solar power, per installed kW.
Storage is the most expensive part of a home power system. Storing electricity requires expensive batteries, which are worn by the process of storing energy. As far as possible, electrical storage should be avoided.
Here are some strategies for avoiding electrical storage.
Certain loads do not need to be run all the time: they can be run when it is convenient. An example of this might be a washing machine: laundry can collect for a day or two while we wait for the wind to blow. On a windy day, when the batteries are charged, running the washing machine does not wear the batteries, since the power comes straight from the windmill to the washing machine.
In the same way, things like well pumps can be operated by two float valves, not one. The first float valve prevents the header tank from being emptied, but the other only operates when the wind is blowing and the batteries are charged, and is used to top the tank off to the highest possible level. In this way the energy required to lift the water is stored as lifted water, not as electricity.
There are other ways to store energy than as electricity. Washing machines, dishwashers and tumble dryers, for example, use heat in their operation. They can generate that heat from electricity, or they can collect it from another source.
By using a heat store to store heat, the electricity consumption of these appliances can be massively reduced, so lowering the amount of stored electricity required to run them.
The only sensible way to store electrical energy is in batteries. I have in mind a battery bank with a nominal terminal voltage of 60V. This gives the following arrangements:-
|Chemistry||Number of cells||Service life||Number of discharges||Discharged voltage||Nominal voltage||Float voltage||Maximum voltage|
|Lead Acid||30||10 years||300-2000||48v||60v||68v||90v (ZDV charging)|
|Nickel Iron||50||80+ years||5000+||50v||60v||80v||90v|
Storing our typical requirements of 12kWh per day requires 200Ah. So a five-day store will require 1000Ah. Our base load will only need 1/10 of this, though, so if loads can be juggled so they are only used when the sun is shining or the wind blowing, a five-day battery bank could last more than a week.
The cost of the energy comes in two parts: the generation cost and the storage cost. It is unlikely that the storage cost would be much less than 15p/kWh, which is a little over double that of grid electricity (but it was about 2 1/2 times when these pages were first written a year ago). So obviously we can save battery wear and so cost if we can time our use so that it coincides with the time the energy is available.
Using the above methods of avoiding storage, we should be able to save 75% of the energy use: mostly running washing machines, dishwashers and dryers. That might enable us to reduce the daily storage requirements from 12kWh per day to 4kWh per day: or 20kWh for five days' storage. That's nice, because I have 24kWh of storage in the back yard.
The most popular form of batteries for home power storage consists of lead and lead oxide in sulphuric acid. Lead and lead oxide are heavy metal poisons that accumulate in our bodies and damage our internal organs. Sulphuric acid is mildly poisonous and very corrosive. The batteries are unlikely to last more than 2000 discharge cycles even with light use - call it ten years tops.
A better battery for lasting is nickel-cadmium cells, which last almost maintenance free for 20+ years. The electrolyte is potassium hydroxide, which is corrosive when concentrated but not poisonous, and lithium hydroxide, which is mildly poisonous and psychoactive, but otherwise no worse than potassium hydroxide. Nickel is not poisonous either: lithium, potassium and nickel are all very common in the environment. The problem is cadmium, which is a very nasty heavy metal poison.
Now if you replace the cadmium in a nickel-cadmium cell with iron, you decrease the efficiency of the cell and increase its self-discharge rate. Now all of the ingredients are common in the environment. This type of cell, the Edison cell, was one of the first storage batteries available back in 1910. Some of the originals are still in service. The Edison cell may be within the capabilities of the home constructor - here is a page discussing home construction of Edison cells.
The normal setup of a pile of batteries and an invertor is all very well, but there's a problem with that, too. Most invertors consist of a transformer and some kind of switching circuit. (Some consist of a switching circuit and a filter.) Generally with wound components like transformers and filters the energy lost is almost constant, regardless of load. This means that transformers (and filters) are at their most efficient when they are running at 100% of their rated capacity.
The plan is to have a split DC power bus, running at between +/-24v and +/-50v, depending on battery chemistry and state of charge. All components connected to the bus must accept this sort of voltage range.
The full load delivered would be 15kW, which at +/-24v is 315A. So the bus must be capable of delivering this current.
For the PV array, this would be fed directly from the blocking diodes on the array, and the array would be split around earth. For typical cells, that would mean a string of 180 cells with an earth tap halfway through and a blocking diode at each end. For Atlantis Energy Sunslates, that would mean 34 units as two chains of 17, each delivering about 4.6A at +/-49v. The roof is 14 rows of about 30 slates, apparently of the same size as the Sunslates, so we should be able to get ten strings on the roof, for 46A peak, or 4.5kW installed. That'll not do much for the dumpload, but it'll provide plenty for the batteries and enough to run intermittent appliances.
For the windmills, a 55v 3-phase output would be delta rectified to produce +/-47v, which should be plenty enough for charging. For governed windmills, a sensor would detect overcurrent and/or underspeed, and signal this to the dumpload system.
A rectifier will accept a 110v split phase input (55v-0-55v AC) from a generator, to allow emergency power to be fed onto the bus. That will provide +/-77v, which will be reduced by a smoothing and stepdown circuit to provide +/-48v. This input would normally be expected to provide 3kW: that is, just over 30A. This input would not normally be expected to operate the dump loads, just charge the battery -- but a generator may well heat the heat-store through the waste heat from the engine.
Alternatively, this input might come from a continuous-rated transformer from grid supply or 240v generator, if this was the most convenient way to run the system.
Storage on the bus would consist of a battery and charger. The battery would provide power to the bus whenever the bus voltage fell below the battery voltage (from +/-30v down to +/-24v in the case of lead acid cells), and when the voltage on the bus significantly exceeded battery voltage, and the batteries were less than fully charged, the charger would drive the batteries with the appropriate current. Lead acid batteries would stop accepting charge at about +/-34v in a normal float charge regime, but Edison cells might go up to +/-45v.
When the charger concludes that the voltage is sufficient, current will be reduced to a float charge level appropriate to the chemistry, and the charged state of the batteries will be indicated to the dumpload system. At this point the voltage on the bus is likely to rise abruptly.
All the loads on the bus are via the mains voltage invertor. But since the control of these loads is separate, they are considered separately.
The main load on the bus is a 15kW invertor. This delivers 63A at 240v, and provides power to the whole building. It will use a modular system to maintain efficiency at low loads, probably as a 200W module and a series of ten 1.5kW modules. At the start of a half-cycle only the 200W module will be connected, then output waveshape monitoring circuitry will detect when the module is unable to cope, and more modules will be switched in as necessary.
The modules will span the split bus, and so must be capable of withstanding maybe 100v on the input, but must still be capable of functioning at as little as 48v. Conversely a 1.5kW module must be able to sink 32A to function at low voltages. Since the drawn current may not be exactly sinusoidal, and different half cycles may have different loads, we must ensure that the modules can deliver more than their rated average capacity for a few cycles.
The output from the modules will be combined through an output switch unit (to isolate the windings of unused modules from the mains supply) and earthed on one side. The neutral will be taken off this earth point and the live off the other side, and will be fed to a "consumer unit" fed off a 63A RCB.
The "consumer unit" will provide all power to the house: the only mains connections to the invertor will be via the L,N,E connections that come out of the "consumer unit". This ensures the RCB can protect the whole system.
The dumploads will consist of two parts: a "cheap rate supply" that is switched on when the battery is detected to be fully charged and maintained for, say, an hour afterwards; and a system of dump loads that are engaged when the windmills or PV array signals that excess power is available. The dumploads consist of up to five 240v immersion heaters in the heat store, each run off a 16A MCB, and will heat the water to a maximum temperature of 90oC.
Since the dumploads consume almost all the power the invertors provide, they will only be engaged when there are no other loads on the system. But this can be determined on a cycle-by-cycle basis, allowing the dump loads to soak up excess power without overloading the invertors or wiring, or tripping the breakers on the consumer unit.
This page is some notes on Domestic Power from Renewable Sources, and is written and maintained by Simon. At this stage these pages are constantly under revision. Thoughts and comments are welcome.