El-Sitgeti- Electrical point 4-Pumping, irrigating and sawing

Electrical point 4: Pumping water, irrigating and running a 3-phase 3KW table saw

 

It is now getting more interesting. In this post I describe how to run a 1.5 KW (2HP) submersible pump and a 3KW table saw directly from the solar panel arrays, without connection to batteries, using a variable frequency drive.

Pumping water

Most of the water resources of the farm come from a natural spring located at a high elevation point on the farm. It provides for drinking water as well as irrigation water. The stream of the spring varies over the seasons being obviously lower in summer, but never stops though. The water is saved in a 70m3 water pool, which was built in the year 1938. Since all the farmed land lies at lower elevation than the pool, the crops are irrigated from the water pool by gravity. An irrigation network runs over the farmed area of the land. 

 However, even if this system allows for irrigating without the use of energy, water sill needs to be pumped upwards in some specific cases:

1)   Throughout the winter there is often a surplus of water in which case it is pumped upward in a 20m3 reservoir.

2)   For personal consumption (washing machine, shower, etc…) the water is regularly pumped upward in a PP 1m3 reservoir, in order to provide for sufficient water pressure out of the tap.

These tasks were carried out with a petrol surface pump until recently, which was replaced by an electric 3-phase submerged pump.

The pump

Submersible pumps are most often used for wells and especially deep wells. They are designed is such a way as to fit into tight holes (typically 10-15 cm diameter) and to pump water at high elevations (typically higher than 100 m). Therefore they consist of small diameter cylindrical shaped enclosures fitted with a watertight motor which drives a number of stacked turbines on its axle. The more turbines there are the higher the pressure the pump can reach and thus the higher up it can pump water.

We purchased one such pump from DAB, an Italian manufacturer, and not being from the field I have to say I am very impressed by the technology. It costs a mere 500 euros, it is made completely from stainless steel parts, is incredibly small (see picture), has a rated power of 1.5 KW and can pump water at a rate of 6m3/hour and keeps that rate at elevation up to 90m! It feels to me that these pumps have a very high usefulness/price ratio! They are particularly attractive for our application, as they do not have any problems of air intake, as is often the case with surface pumps and can simply be placed at the bottom of the pool on a simple stand that comes with it.

View of the spring and water pool. The pump lies at the extremity of the pool, one can appreciate how small the pump is. The right picture shows the pump from a closer distance.

 The variable frequency drive

As I have already mentioned in the previous posts, variable frequency drives (VFD) are one of the most important piece of equipment in our installation. Lets explain briefly what they are used for and how we adapt them to our system.

VFDs are used extensively in industry and their use is specific to the drive of 3-phases motors. 3-phases induction motors, by conception, have their speed limited to a single RPM value, which is directly linked to the frequency of the 3-phase mains supply (50 Hz or 60 Hz). This is were VFD come very handy, as they allow controlling the speed of the motor as well as a lot of different parameters such as torque, acceleration and have PID speed control features, load compensation algorithms, etc… In a standard AC system, in order to be able to vary and control the speed of the motor, the VFD rectifies the 3-phase input to a DC supply of around 325V (in the case of a 220V VFD) and then recreates a variable frequency alternating 3-phase voltage to feed the motor with using power IGBTs. The trick is that to drive a motor the VFD does not need to recreate a pure sine wave voltage, rather the output of the VFD is a Pulse width modulated signal of varying frequency. By varying the shape of the signal the VFD controls both the RMS Voltage applied to the motor windings as well as the frequency (see here for a very nice explanation on VFDs).  This is very important as when a velocity lower than the nominal velocity of the motor is required not only the frequency of the 3-phase must be lowered, but also the RMS voltage to avoid the motor from withdrawing more current than it is rated for. The voltage/frequency ratio generally follows a linear law but most VFDs have many different voltage/frequency curves options.

What is the advantage of using VFD in DC photovoltaic system (see here for additional info)?

1)   Similarly to Switched mode power supplies, VFDs’ working principle implies the rectification of the AC mains supply to a DC source used internally to produce the variable frequency 3-phase output. It can, therfore be used with a DC source. In practice it is as simple as connecting the DC BUS to the AC input, or in some brands (e.g. Fuji Electric) to the DC input connection. When the DC source is connected to the AC input it simply passes through the rectifier bridge and comes out of it unchanged (see scheme).

2) The VFD voltage working limits are very wide and most VFD handle input voltages from DCV200 to DCV400, adapting automatically the output with respect to the input voltage. This is particularly interesting for our system as the DC BUS varies during the day depending on various factors such as weather conditions, load on the BUS, etc…

3)   One of the most important features of the VFD is the starting acceleration ramp, which allows increasing slowly the speed of the motor on start up reducing to a great extent the current surge associated with the motor's start up. This feature is absolutely vital to off-grid PV systems, as it allows running motors of a nominal power virtually similar to the PV system nominal output. This would be absolutely impossible in a conventional PV system since the current surge generated upon motor start up can easily be 10 times its rated current and can easily exceed this value if the motor is under constant load (e.g fridges, we have recorded current surges 20 times the motor rated power in some cases).  It would therefore be difficult or most probably impossible to have a 1KW AC motor running on an AC inverter with less than 5KW nominal power without it disconnecting during motor start up due to over-current detection! Additionally, the start up speed ramp induces less stress on the motor mechanical parts and therefore increases its longevity very significantly. 

4)   The second most important feature is the ability to vary the frequency, and therefore the speed of the motor while in use. Reducing the speed, even slightly, induces a very significant reduction in power consumption and therefore allows for the use of the motor under non-optimum conditions, and allows adapting the consumption to external factors. This implies that motors can be used without the need for batteries, as would be the case in conventional PV systems. 

5)   Virtually all VFDs have a low voltage control board, allowing controlling their parameters (such as frequency, turn ON and OFF, etc…) with external relays or switches, or programmable logic boards.

6)   Last but not least they are pretty economical, and their lifetime is around 10 000 hours of continuous use.

The set up

As mentioned above, the pump is placed in the water pool. Therefore we built a pretty little stone structure besides the pool that hosts an electrical cupboard which contains the necessary equipment to run the pump, and a table saw (see picture). Indeed we use the same VFD to run a table saw for cutting wood lumber, when the pump is not in use.  

View of the stone structure and the electrical cupboard with the VFD on the left and the modified electrical box on the right.

The saw being the equipment having the highest power rating (3KW) we chose a VFD with a slightly higher rated power. We acquired one from Fuji-Electric (Frenic Multi model, 4KW) for about 380 euros. It has an incredible amount of functions, but we only used a few of them. Much cheaper alternatives are available, e.g. Chinese brand 3KW VFD from ebay for about 150 € (new)!

Similarly to previous electricity points we modified an electrical box with a voltmeter/ammeter and switches.  The VFD is directly fed with the DC BUS wired through a fuse. The output of the VFD is wired through a conventional 3-phases switch, which allows selecting which equipment between the pump and the saw is being used. The saw is connected to the VFD output via a conventional three-phases plug.

VFD control

It is worth mentioning that for a 350V DC BUS voltage, only VFDs with 3-phase input of 200V (manufactured for north America), and VFDs with single phase input of 200V (manufactured for Europe) will work. They will provide a maximum value of 220V RMS output (the output voltage is always referred to as between phases). This voltage output is compatible with 220V/380V 3-phases motors, however the winding of the motor will require to be wired in a delta configuration. In the case of the pump which has a sealed connection box, the rated voltage is specified on order (3-phases 230V in our case) and the pumps comes ready to be plugged in.
 
The VFD is controlled via a contol box for the saw and by a timer placed in the electrical cupboard for the pump. We modified the original saw’s control box and placed and ON/OFF switch, an emergency switch and a mini reset switch (bought from RS components). All the switches are connected to the main control box via a RJ-45 (internet) cable. The switches actuate relays whose output is directly connected to the VFD control board. 

 

View of the saw and close up view of the control box.

The control box includes as well a timer (timer prototyping board acquired from ebay: 6€!), a switch which allows selecting the frequency control mode –automatic or manual– and a potentiometer which allows varying the frequency when the latter mode is selected (see pictures below). Selecting the automatic mode with the switch activates a VFD built-in frequency control method, which basically adjusts the frequency linearly with the DC BUS input value. The function allows reducing the speed of the saw or the pump automatically in case of a sudden decrease of the DC BUS voltage, possibly due to an increase in the motor’s load, or to a temporary decrease in solar irradiation.  Although the manual mode has no real usefulness for the saw, it is very useful for the pump when it is used for direct irrigation as it allows varying the flow of water and the pressure on the irrigation lines. It also allows pumping water in cloudy conditions by adjusting the frequency as to maintain the highest possible voltage and current having a simple look at the voltmeter/ammeter. This is equivalent of finding the MPP manually!

View of the VFD control box (right), and close up view of the modified electrical box (left).

El-Sitgeti-Electrical point 3-Yurt

El-Sitgeti-Yurt- Charging a Ni-Cd battery and heating up water with a DCV source 

The Yurt is the point of where most household appliances are used. Both ACV 230 and DC BUS are distributed here. The ACV 230 comes from point 4, which will be described in a later post.

The DC BUS is used here for 2 purposes:

1) Charging the Ni-Cd battery set

The charger consists of a simple SMPS bougth from ebay. They are sold as LED strip power supplies, and they are sold in various current ratings. We used here a 50 A (600W) type (50€), which is more than sufficient as the charge current will hardly go over 30 A. However having a current rating twice the nominal current will increase the longevity of the power supply and avoid over-heating. The output voltage is controlled by a potentiometer and it is adjusted to a value slightly above the float voltage as to provide for a maximum charge current of about 25 A and minimum water loss when the batteries are in full charge. Since the power supply does not have a feedback control loop the output voltage gradually increases according to the DC BUS voltage and thus with the solar irradiation intensity  (e.g. when the sun starts shinning DC BUS is low, and thus the charge voltage remains below the float voltage), and starts pumping current into the battery only when power starts being available from the panels.

The Power supply is wired through an electrical box that combines AC and DC parts. The voltmeter allows monitoring both the DC BUS voltage and the charge voltage by simple pressure on the left switch (see picture). The current value shown on the ammeter is the current consumption on the BUS, but the charge current can be easily calculated from that value.

Electrical schematics of the box. Note than when the switch is actuated to monitor the charging voltage the ground of the secondary part of the SMPS is connected to the ground of the DC BUS (primary). There was no way around this (using a simple switch at least), but we believe it does not raise any safety concerns as the time when the two grounds are connected only last for a few seconds.

View of the electrical box and battery charger. Inside view of the box and view of the charging voltage being monitored (pressing the left switch).

2) Heating up water with an IGBT controlled boiler

This is the first appliance of the blog to be modified to work directly on the DC current rather than on AC. If you read my previous posts you will easily understand why it can work.
Boilers are mainly composed (at least the most basic of them) of a mechanical thermostat (or thermal switch) and a heating element, practically a resistor with a resistance a few tenth of ohms (typically of a rated power of 1.5 to 2KW). The heating element is specifically designed so that its maximum current rating is achieved under a 230V RMS voltage. If we were to connect that heating element to the DC 350V it would simply overheat and eventually be damaged because too much current would pass through the resistor. The way of avoiding this is to chop our DC source at given frequency and to modulate the pulse width (or duty cycle). Since the RMS value of any variable voltage is given by the integral of the wave over its time period (see figure below), by feeding our heating element with such pulse width modulated (PWM) signal, with an amplitude of around 350V (the value of the DC BUS), we will actually be varying the RMS voltage value we are applying to the heating element. By making sure the RMS value doesn't go over 230V we will ensure the heating element will not withdraw more current than its actual rating.

RMS voltage of a PWM signal respect to its duty cycle. As seen in the formula, the RMS value is proportional to the square root of D which corresponds to the duty cycle.

We can even extend this concept further, by actually varying the duty cycle not only to limit the RMS value, but to limit the current consumption to that available at that particular moment in time. That is, if solar irradiation is low (and therefore the available current on the BUS is low) we would decrease the duty cycle as to adjust the current drawn from the heating element to the available current from the solar panels at maximum power point (MPP). This will have the effect of decreasing the temperature of the heating element though, but will still offer the advantage of heating water however at a slower pace. If the duty cycle was to be fixed to an RMS value of 230V, the boiler could only be turned on a sunny day at around noon, for it not to drive the voltage of the BUS lower than the MPP voltage.

To achieve such task we designed an boiler controller device with the above-mentioned characteristics (120 €). It consists of an op-amp comparator which compares the voltage of the DC BUS to a threshold set by the user (by means of a potentiometer) and an IGBT driver which feeds the gate of a power IGBT with a square wave PWM signal. the IGBT driver varies the duty cycle so that the value of the DC BUS does not drop bellow the set threshold. Note that the voltage threshold is an indirect way of limiting the current passing through the heating element. It is not as accurate as actually limiting the current to a set value, but it does the job as the voltage threshold will be set to a value close or above the MPP voltage, and therefore no more current than that available at MPP will be drawn from our panels.

Front view of the boiler controller.

The device comprises a digital thermal switch which turns on or off the heating depending on the temperature threshold set by the user. Thus, the boiler was stripped down completely and the original mechanical switch was completely removed and replaced by the temperature sensor that came with the digital temperature switch (a 5Kohms NTC resistor). It is as simple as that! And in practice it works extremely well: With a 50 liters boiler (only 65 € in our local hardware store!!!) and with  the temperature set to 67.5°C, it only takes between 30 to 45 min to heat that volume from 40°C to 67.5°C early in the morning (the heating element has a 2 KW rated power). And the beauty of it is that it reaches that temperature even on the cloudiest weather conditions! It just heats up the water more slowly. The process is fully automatized so you never have to worry whether you are going to have hot water at the end of the day. As long as the DC BUS rises over the set threshold it'll start heating up and adjust the heating rate according to the current available on the BUS. The threshold is set according the the priority we want to give it. If it's set to a low value it will have a high priority and lower priority if it is set to a high value.

View of the yurt connection point (3), with the electrical cupboard at the bottom, and the boiler (50 L, 2 KW) on top of it. The right picture shows the boiler controller in the cupboard. The DC BUS input is wired through a 10A fuse to protect the heating element in case the IGBT gets short circuited (nominal current at 230V RMS voltage is approx. 8.5 A).

No more cold showers!!!

El sitgeti-AC and DC Connection cupboad

El Sitgeti-Connection point 2

Nothing special about this point, it is just a connection point were the upper part AC and DC lines are connected to the lower part AC and DC lines. Though I thought it still deserves a picture for the beauty of the integration of the cupboard into the stone wall!!

El Sitgeti-Connecting the panels arrays

Connecting the panels arrays

To start with, the panels are connected in series using 4mm2 solar cable and MC3 connectors (acquired from Solarhertz). The 4 lines are then connected in parallel  in a conventional electric box modified as to integrate 4 voltmeter/ammeters in order to be able to monitor the current consumption and DC voltage on each line. The voltmeter/ammeters were purchased from ebay (China), and they literally cost nothing (from 2€ to 8€)! There are mainly 3 models available (10 A with built-in shunt, 50 and 100A with external shunt). They operate on 0-100V range so the trick to measure voltages of up to 400V is to connect them through a simple voltage divider. For convenience we chose a 10 dividing factor so that they will show 1/10 of the real BUS value (see scheme below). The two resistors have values of 100K (2W, 1%) for R1 while R2 (0.5W) is adjusted between 11K a 12K to compensate for the internal impedance of the voltmeter (about 100K according to the manufacturer specifications). The right value of R2 for each voltmeter is simply found by trial and error, comparing the voltage with that of a precision voltmeter and picking a value of R2 that gives the closer to 10 dividing factor (standard values for R2 with 1% error are 11.3K, 11.5K and 11.7K). The voltmeters are powered by recycled phone charger SMPSs (1 € in my local second hand shop!). Any type would do as the voltmeter can be powered with a voltage range of 4.5V to 30V. The SMPS are directly connected to the DC, and will keep the output constant with a BUS input as low as 150V! In this case 10A built-in shunt ammeter were used, as the maximum current from each line is not expected to exceed 4.6 A (for an accurate measurement of current it is recommended to use these ammeters at half there actual rating).

Finally the 4 blocking diodes (acquired from RS components) were simply soldered onto the incoming cables and insulated with heat shrink tubing.

Electrical schematics of the connections box.

View of the connections box inside the electrical cupboard. Close up  and inside view of the box. Notice that the voltmeter indicates 1/10 of the actual voltage. The coma can be hidden by simply applying black paint over the LED "dot".

Note that while each ammeter shows the actual current consumption of each of the individual panel array, the voltmeters all show the global voltage of the BUS. If the voltage of a specific array is to be monitored, the lines of the remaining 3 arrays have to be disconnected (opening the fuse holder). The voltage of the array being monitored will show on all four voltmeters.

Mixed AC/DC off-grid photovoltaic power at "El sitgetí" organic farm (Catalonia, Spain)

El Sitgetí organic farm is located about 2 km away from the town of Bonastre (Catalonia, Spain) in a little valley away from civilization! The farm is located too far from the village to be connected to the electricity grid. And although his owner, Charlie, has relied on a truck battery to power his computer for the last 10 years it was about time to install a decent size micro photovoltaic power plant there! The whole project started by the acquisition of second hand solar panels (believed to be 3 years old, bought from a refurbished photovoltaic power plant, 80€/panel, a real bargain!). The system consists of 28 monocrystalline Silicon panels of 220W power (Voc=48V; Isc=5.3A, FF=72%), for a total of over 6 KW (though this is a theoretical value, which is never matched in practice)  

The 4 lines of 7 panels connected in series. 

So I haven't talked all that long about DC power just for the sake of it! Here is how we plan on taking full advantage of the DC power generated. The 28 panels are connected as follows: 4 lines of 7 panels in series are connected in parallel through 4 blocking diodes. Each line produces a theoretical maximum power point voltage (VMPP) of 336V (at 1 SUN) which corresponds approximately to the maximum point of the sine wave of the mains supply in European countries (230 x sqrt(2) = 325V) (see previous post for an explanation on that). 

Theoretical I-V curve of individual panels and of the 7 panel array

The DC source produced from the 4 lines of panels can therefore be used directly to power equipment running on DC (battery chargers, variable frequency drives, boiler, etc...). In practice the VMPP varies between 250 V and up to 400V (at 1 SUN), depending on various factors, such as  the external temperature, and the cloud conditions. We will refer to it as DCV350, or DC BUS from now on. Despite this variation in VMPP our system will still work perfectly as a DC source to power various equipment. 

Map of the installation

Map depicting the electrification of the farm

The total land area of the farm covers some 28 ha, which is pretty big. Luckily most it is forest and only about 2 ha are actually farmed. Given the location of the panels (placed at one of the highest accessible and most cleared point on the farm) with respect to where the electricity is to be distributed, some calculation were necessary to minimize voltage losses across the wires. First the location of the ACV230 inverter was chosen to be placed at middle distance between the most remote places where the AC has to be distributed. This leaves about a 100m distance to be covered by the current both ways before  being consumed. Taking into account a maximum power of 4000W (which is the maximum power of our inverter) this leaves us with 10 mm2 wire diameter to limit the voltage drop to less than 3% across the lines as is the norm in most European countries (click here for a voltage drop calculator). For the DC lines, the cross section of the wire can be downsized a bit because the voltage is higher that on the AC line. The losses in this case will not interfere with appliances operation as in AC in the case of a large voltage drop. But it should still be minimized to avoid power generated being lost to heat through the cables. A cross section of 6 mm2 was then used. 

There are 7 points of connections and distribution, which are all marked on the scheme. Note that there is a set of battery in 4 locations. Battery sets of 12V are only used for illumination. All the light bulbs used in the farm are LEDs, so consumption is ridiculously low. All 12V sets are recycled batteries: one 20 years old NiCd , one  > 30 old NiFe, two truck batteries, but they still offer a minimum autonomy of 15 days in our condition. The 24V set is a refurbished (9 years old) Lead/acid set, which has only a small fraction of its original capacity, but as explained in my previous post, the all point of the mixed DC/AC system is not to rely on batteries as in conventional systems.
All the batteries are charged via the DC BUS using switched mode power supplies (SMPS). The 24V set is charged with a modified welder/inverter (working on the same principle of SMPS) and is connected to the ACV 230 pure sine wave inverter.

Panels location survey, tilt angle and shading

Location

First of all it is imperative to know where the south lies respect to where we want to install our solar panels, there are different way to determine it, as explained here. Then the way we proceeded was to print the sun path chart for our location by simply entering the exact latitude and longitude of the site on this web site. We then built a solar elevation gauge by following the instruction of this very useful link. Given that the farm is located in a valley with hills on both sides, the hours of sunshine are slightly limited by the fact that the sun only goes over the east lying mountain about one hour after sunshine, and similarly for sunset.

Solar path map. the green line at 11° elevation represent the hills located  eastward  and westward of the panels site

Tilt angle 

Given that we installed our panels on a flat surface it made sense to take advantage of it to mount our panels on variable titling angle mounts. Unfortunately given the height of the mount holding 3 lines of panels (see below to understand why we chose such type of structure)  it could only be manufactured with 2 tilting options. Although the angles of maximum energy production (for a 2 tilting angle system, with angle adjustment twice a year) are 16° and 54° in our latitude, we finally chose our system to be 17° (summer) and 34° (autumn, winter, spring). This was, on the one hand, to avoid the structure from being too tall and risking damage from strong winds. On the other hand these 2 angular positions are optimum for summer (17°) and spring-autumn (34°) which are the period of higher activity and thus higher energy demand on the farm (low temperature vegetable storage being the most import energetic demand in summer-spring-autumn), winter being a very quiet season and sufficiently cold not to require any additional energy to preserve vegetables. The price of the two mount was about 1200 € (bought from Alusinsolar, Spain) .

Shading

Shading is one of the most important parameter to be taken into account when placing the arrays of panels as even the slightest area of shade over a panel is likely to shut down the energy production of that entire panel (here is a nice study of the effect of shading on output power). In our case shading is even more critical since our system will be very dependent on the DC voltage of the arrays, therefore shading over one panel would reduce the DC voltage of the entire array and could limit or impede the operation of certain appliances. Given the area of flat surface available to us, two options were envisaged. the first one consisted of having four individual lines of panels placed on the ground spaced by 1.60 m distance, or a multiple array stand. After some trigonometry calculations it turned out that the shadow on the arrays resulting from the front row of panels could have been substantial especially in winter (morning and noon) when the sun's elevation angle is below 20° (see figures below, but rather that bothering with tedious, though not complicated, algebraic trigonometry, there is a better option to simulate shade using a free software SketchUp as shown here). To remedy this problem, one should have elevated the 2nd, 3rd and 4th row from the ground as depicted in the scheme below, in order to have the shadow passing bellow the panel's array of the line to its back .

Calculation of the panels height to avoid shading on the back row (system of individual panel arrays, not chosen due to the complexity related to elevating the panels from the ground and having a central pivot point)

 Finally a multiple array mount was chosen for its simplicity, though only mounts holding a maximum of 3 panels lines could be manufactured by our supplier, meaning that the first row would lie on an individual mount at ground level, spaced 2.10 m from the 3 panels lines mount. The shadow produced by the first row only affects the second row (the first row of the 3 panel mount), and only early in the morning and late in the afternoon the shortest days of winter (for roughly a month, from mid December to mid January). The loss on power generation is therefore not very significant, since at the time of shadowing the power generation is low.

Lateral view of the two mounts arrays adjusted to their autumn-spring angular position (34°). The individual mounts were connected together with twined copper wire (approx. 7 mm diameter each) and connected to the ground through a 1.5m copper coated steel rod to protect the system against lightnings.

Advantages and usage of DC source in off-grid photovoltaics

In my previous post I mentioned the advantages of a DC PV system. I will go through them in detail to clarify my point. Fist of all I should mention the typical voltage of the DC source we are working with is in the range of 300 to 350V, the value not being precisely fixed (in practice it can vary from 250V to 400V). This is to match approximately the maximum voltage point of the sinusoidal signal of the mains power supply in European countries, as to be able to adapt commercial appliances (made for AC230V) to DC. Such DC voltage is easily achieved connecting the adequate number of solar panels in series (e.g 7 x 48V; 10 x 36 V; 14 x 24V)

Maximum voltage point of an RMS ACV230 sine signal.

 1) A cheaper alternative to photovoltaic installations

This claim is supported by mainly two arguments. The first and most obvious one is that if DC generated by the PV array is directly used to power appliances we remove the need of a costly pure sine inverter. This is even a stronger argument when considering off-grid systems since they typically are in the 5-10 kW range and the price of a durable inverter-charger for such power typically exceeds 2000€. The second argument lies on the fact that with a DC system, energy consumption can be better optimized in such a way as to use energy when it is available, and store it in various different forms allowing downsizing the batteries (this point is developed bellow).
In practice many but not all the appliances can be modified to work on DC, so it is more convenient to install a mixed DC/AC system. In this latter case a pure sine wave inverter will still be needed, however its power output can be scaled down (and so its price) compared to a conventional system since most of the energy demanding work will be performed out of the DC. The AC will be used to power household appliances such as washing machines, dish washers, fridges, induction cooker... or any other appliances not fit for DC (or at least no easily modified to run on DC).

2) A more efficient way to use energy, using it when it is available, and storing it into different forms (electricity, heat, cold, potential energy).

This sounds pretty obvious! doesn't it? Well at low enough latitudes (typically those of the Mediterranean countries in Europe) there are sufficient sunshine days in a row not to be dependent on electricity storage except for illumination and nighttime.  Therefore there is no need for expensive deep cycle high capacity batteries.  All the energy demanding tasks are better carried out in day time and better carried out from the DC source directly as to adjust consumption with available energy, as for example:

-Running 3 phase motors from variable frequency drives allows varying the frequency (and thus the RPMs of the motor) as to adjust it to the available energy (e.g. on a sunny day full speed, on a cloudy day reduced speed)

- Heating up water, feeding the resistance of the boiler with a chopped voltage from the DC source, varying automatically the duty cycle to adapt the current consumption to the available energy (e.g. on a sunny day the water will heat up at a fast rate, while it may take several hours on a cloudy day, though still be warm at the end of the day)
- Controlling cold rooms or walk-in fridges (e.g. in a farm) with a variable frequency drive, adjusting the cooling rate with the available energy.
- Controlling the AC inverter current input and battery charging with a modified high power switched mode power supply, as to adjust it to the available energy.

3)  Automatize consumption

All the above tasks can be carried out in an automated manner. The simplest way to proceed is to modify the appliances in order to have them turning on/off according the DC voltage of the system (we call it the DC BUS) . That is the voltage of the PV array (or the BUS) evolves according to the solar irradiation and the load. The greater the load, the lower it goes following the I-V curve of the array (see figure). therefore, as long as the DC BUS is higher than VMPP, energy is available to use, appliances can then be given priority by setting their turn-on voltage to a higher or lower voltage. The current consumption of the appliance can then be adjusted (as explained above) to maintain the DC BUS to a desired value (in case current is drawn from another point on the BUS, or cloudy weather).

I-V curve of a hypothetical array of solar panel. The voltage of the BUS decreases with load "going up" the I-V curve.The evolution of the value of the BUS at a given solar incident light is proportional to the load. The operation of the appliances is performed according to the turn on threshold with a priority set for each appliance (the lower the threshold, the stronger priority).

For more precise appliances operation, maximum powerpoint trackers can be implemented to the equipement to accurately ajust the load to the available power at a given time.

Direct current (DC) for off-grid photovoltaics?  

A little historical reminder of the current standards

More than a century has passed since the end of the war of current. But what was it all about? And why did alternate (AC) won over direct current (DC) generation?  Well simply put, it was just about current production and distribution! For more info on this check out the Wikepedia page here. Well it makes perfect sense that AC currents won, given that at the time of Thomas Edison and Nikola Tesla solid state electronics did not exist and that power generation from sustainable (local) sources wasn't at all considered an important criteria! Indeed,  DC current distribution implied that, to avoid losses in transmission through the wires, current should have been generated in close proximity of the consumers households since there was no convenient method to elevate the tension for long distances current transport, and subsequently lower the tension at the point where it were to be consumed. AC current was much better suited to the latter task using power transformers, which had been developed at the end of the nineteenth century (remember Maxwell-Faraday low of induction, transformer can only work with an alternating current). Then AC won and household appliances were developed specifically to function on AC current standards (mainly 2 standards ACV110 and ACV230) (Check this video from the "Post Apocalyptic Inventor" for a very good explanation on current generation and transport and transformers; I strongly recommend you check his Youtube channel if you are interested in electricity related subject).

Since the times of Edison and Tesla things have evolved quite a bit, especially in the area of power electronics, which at that time did not exist. With the invention of power transistors (and especially MOSFETs and IGBTs) lowering a DC voltage source has become much easier. The principle still relies on feeding an alternate source to the primary winding of a transformer, except that now the alternate current is fabricated by chopping a DC voltage source at a given frequency using a power transistor. This principle, most commonly known as  "inverter" is used nowadays in virtually all domestic appliances power supplies (so-called "switched mode power supplies", see here for an outstanding tutorial video series) which supplanted the old fashion primary transformer based power supplies. In other words, instead of using the properties of the 50/60Hz AC mains supply to reduce the voltage in our house hold appliances, the AC input is actually rectified to a positive DC voltage internally and subsequently chopped by means of a power transistor at a higher frequency (typically a few tenth of KHz) and fed through a transformer out of which the voltage comes out reduced to a given value. This method has many advantages, the first one being the extremely high achievable transformation yields (>90%), the very significant reduction of the transformer size (and therefore inducing a reduction in weight), and the ease of output voltage control.There are many other appliances which work on a similar principle, like variable frequency drives, air conditioners, welders....

So I'm getting to the point: One may ask himself, why would we question a current standard which has been established more than century ago? Taken out of context, there is no real motivation for that, but in the case of photovoltaic solar cells, where the current generated is DC type, and that power is generated in close proximity to where it is being consumed, there is a clear advantage if the current could be used straight out of the solar panels, rather than having to feed it to an expensive sine wave inverter which converts the DC output from the solar panels to a usable AC power source. In the following blog we describe the development of an off-grid photovoltaic system which takes advantage of DC power generation to provide for:

-A cheaper alternative to photovoltaic installations

-A more efficient way to use energy, using it when it is available, and storing it into different forms (electricity, heat, cold, potential energy).

-Automatize consumption according to the DC voltage output.