Alternative Technology Assn Forums

1. 

   wlb
   Member

   Just fishing for ideas. It would be handy if something already exists, but I doubt it.

   I'm looking for a way to get an approximate indication of how much power is available and not being used from an off-grid system. That is, at any given moment, how many Watts are available from the arrays and not being used to charge the batteries and run other loads. Power that is available, but not being used.

   My system can tell me how much power is being delivered, but not what is available. I can't think of any way of doing it without knowing how the arrays perform under maximum load under various light conditions, and then comparing that with the actual light conditions, measured by other means. I think it would be quite complicated and difficult, but I thought it worth asking in case I've overlooked something.

   The reason I ask is that it would be useful information to have, for the benefit of the other members of the household (shall we say) who can't seem to grasp when it is okay to use a large electrical consuming device, and when it isn't.

   Posted Wednesday 7 Jan 2015 @ 4:09:53 am from IP #
2. 

   David
   Member

   wlb said:
   Just fishing for ideas. It would be handy if something already exists, but I doubt it.

   Ah a run down on your system would be helpful. however not critical.

   The amount available however is only relevant based on how much can be consumed at any instant. A grid-connect type system typically does provide the amount of power it's producing at any instant and typically displays it on the display. You would then need to know how much is being consumed within the premises to know how much is available or whether it is exporting excess at all.

   An off-grid system cannot tell how much is available because it/they must limit the amount of output power based on where it's being used.

   If batteries are charging, using x amount and the house is using xx amount, then a an off-grid system must only produce x + xx amount of power plus a tiny bit extra for other losses.

   The only time you will know for sure how much power you are able to produce is when the system decides to draw on the batteries to make up any short fall the array/s might not be able to produce at the time.

   Posted Wednesday 7 Jan 2015 @ 5:06:34 am from IP #
3. 

   johnmath
   Member

   wlb, you said you are off-grid, so you have battery storage? What monitoring do you currently have on your system components? Battery state of charge? Charge current? Load? Do distribute all of your power at 240 volts (i.e PV > charge controller > battery > inverter > appliances), or do you have low voltage or a mix of low and high voltage appliances?

   If I have made the right assumptions, you want to know when you have power from your PVs to spare because it isn't being used to charge the batteries as they are, or nearly, fully charged, and it isn't being used to run appliances either directly or through an inverter; and approximately how much spare capacity there is (which will be a function of theoretical output availability given the current insolation and angle of the sun, less any power for battery charging or other appliances in use, and taking into account the less than 100% conversion efficiencies).

   For example, if the suns shining, the batteries are charged and you're not in the shed welding, there might be, say, 2 kilowatts of spare capacity that could be used for drying teenage hair and playing Xbox on the home theatre.

   Posted Wednesday 7 Jan 2015 @ 5:19:52 am from IP #
4. 

   wlb
   Member

   David said: ... The only time you will know for sure how much power you are able to produce is when the system decides to draw on the batteries to make up any short fall the array/s might not be able to produce at the time.

   Exactly David. I know I can't measure it, for the reasons you've outlined. I was hoping there might be a clever way of estimating it. I can't think of one, other than comparing known maximum outputs for a range of insolation conditions and subtracting the present MPPT output, and that would be a job for a clever data logger.

   The system consists of 2 arrays (1.4kW & 1.75kW) charging a 24v bank of 2v flooded lead-acid cells (910Ah/C100) via a pair of networked Xantrex MPPT charge controllers.

   Although, perhaps I'm getting a bit ahead of myself. It might become irrelevant. I've been pondering this on and off for a few years and just went back to it today during my lunch break. Some time before winter, I should have the next upgrade completed and that will add 2 more arrays taking the array total to 6.8kW. In that case, it might be a simple matter of using one of the alarm relays on a charge controller set for a minimum output condition, and a clock. In other words, if it's after certain time of day (late afternoon) for a particular time of year, AND the array is delivering power below a certain threshold, an indicator says you can't use large appliances.

   The house will also be rewired in the next 12 months or so and it had occurred to me that a simple technique might be to use UK or US power points where these appliances are used (e.g. hairdryer, vacuum cleaner, new Thermomix) and change the plugs on those appliances. Then simply isolate those power points after a certain time of day.

   johnmath said: ... For example, if the suns shining, the batteries are charged and you're not in the shed welding, there might be, say, 2 kilowatts of spare capacity that could be used for drying teenage hair and playing Xbox on the home theatre.

   You've got it John. That's exactly what I was looking for.

   The house is supplied by both 24Vdc (regulated) and 240Vac from a 3kW Xantrex inverter. The control panel for the MPPTs gives state of charge, charging current, voltage, and the inverter displays what it is delivering.
   The house was built in the early '80s when flooded cells were cheap (ex-Telecom) and inverters, solar regulators and PV modules were very expensive and it has just evolved from there. There is still predominantly 24Vdc lighting using modified 20W fluoro batten units (mini inverters), LED modules, and the odd LED and compact fluoro 24V bulb. The lounge/kitchen area has 240Vac lighting using CFL bulbs. The fridge is a 24Vdc eutectic home-made affair. The house has to be rewired in order to replace all the old 24V lighting. The delay has been that the roof has to come off to do it properly.

   Posted Wednesday 7 Jan 2015 @ 6:08:42 am from IP #
5. 

   johnmath
   Member

   My suggestion would be to estimate the actual PV potential by using one or more very small PVs (a watt would do), one for each of your arrays, at the same azimuth, elevation, direction and pitch as your arrays. Take the output of each mini PV through a mini charge controller to a dummy load consisting of "a never fully chargeable battery" simulator (otherwise known as a resistor and zener diode). Measuring the the voltage and current of your monitor panels and multiplying them together to calculate power would give you a scale value for the "real" arrays's power potential. The only serious complication would be partial shading of the real arrays, but if you could uses the real arrays' normalised output voltage and scale your monitor estimates accordingly, you might even be able to approximate for that.

   Now all you need to do is through some computation power against the data, sum the potentials, deduct the actual usage data from your inverters and charge controllers, and come up with some sort of display device.

   Posted Wednesday 7 Jan 2015 @ 11:18:26 am from IP #
6. 

   David
   Member

   johnmath said:
   Now all you need to do is through some computation power against the data, sum the potentials, deduct the actual usage data from your inverters and charge controllers, and come up with some sort of display device.

   Heck I think I'd rather just use one of the many control switching in the Selectronic SP-Pro to trigger circuits on/off when there's enough power and cut them when it drops say x% in SOC or whatever in program settings, or add an RD1 link it to do several options from one triggered event

   Posted Wednesday 7 Jan 2015 @ 11:37:57 am from IP #
7. 

   rockabye
   Member

   wlb said:
   Just fishing for ideas. It would be handy if something already exists, but I doubt it.
   ........

   The reason I ask is that it would be useful information to have, for the benefit of the other members of the household (shall we say) who can't seem to grasp when it is okay to use a large electrical consuming device, and when it isn't.

   This is where the test equipment attached to your head and the analytical capacity of the human brain work surprisingly well. A large digital voltmeter prominently displayed in the house is your best and cheapest option. Assuming a 24V system if it is around 28V and full sun outside most people have no problem understanding it's ok to turn on things. Trial and error with the instantaneous reduction in volts soon demonstrate what's possible. Turn it off and if the volts quickly return to float (28V) you soon figure out how much you can get away with.

   There are a lot of other factors like weather forecasts and time of day to consider but that's all part of the fun of being in control of your own energy production.

   Automation is good if you have the hardware but the human brain can do a reasonable job.

   Posted Wednesday 7 Jan 2015 @ 9:41:04 pm from IP #
8. 

   wlb
   Member

   Thanks John,
   I'd considered using small PVs to sample the light at different array locations for another reason, a couple of years ago, but I hadn't looked at it again since then because I'd found what I wanted to know by other means. I hadn't thought of using a Zener to mimic a battery and hold the voltage. That's a good idea; thanks.
   If I use bigger PVs, around 34W, then the only calculation needed would be to move the decimal point 2 places, and add the 2 values together. I could probably use a shunt on the DC supply to the inverter and use my data logger to monitor that and the solar samplers, and then do the calculations. But I'm probably getting ahead of myself, as that's more of a challenge type project for when I've finished the system upgrade.
   Partial shading is only a problem in the summer and before about 10am. It's caused by a huge gum tree about 50m from the arrays, so the transition when it occurs doesn't last long. The shadow moves down the array relatively quickly.

   David,
   Alas, the Xantrex MPPTs can give me relay outputs for all sorts of condition - except float charging! It's a ridiculous omission. Everything else about them, I like. My previous Outback MPPT had a float signal available. But I could use one of my original Plasmatronics regulators to monitor battery voltage and provide a float output signal.

   Rockabye,
   My built-in test equipment works fine. I know the system well enough that I have a pretty good idea of when the batteries should have reached float just by the prevailing weather. It's the others who don't seem to understand. I installed a large LED voltmeter in the kitchen about 4 years ago, but it didn't seem to help. And then when I added the 24V regulator to solve the problem of marginally designed fluoro inverters periodically dying, it became useless unless I rewired that part of the house. (The easiest place to install the regulator was in the battery shed, so all the meter now sees is regulated)

   Anyway, after the upgrade, we'll have all the system controls and instrumentation available in the kitchen. At present it's all still in the battery shed with the rest of the system. I put a trench and conduits through to the house a couple of years ago to bring remote controls and readouts to the house, but I since decided to completely relocate everything and do it properly, starting from scratch. As I said, it's a very old system that has simply evolved. The time has come to start again with a clean slate and do it the way I want it.

   I will certainly have a prominent voltmeter and battery SOC indicator in the house when the renovations are finished. And based on all this helpful input, I think the best way to convey the system status to the household is an indicator that shows when the batteries have reached float, and another that shows the estimated available power using the little solar sampling PVs. The full calculation system can wait until I have spare time and feel in need of a challenge.

   Using this arrangement, if the float indicator is OFF, then you don't use anything heavy duty. If the float indicator is ON, then you can see what's available from the arrays and see if they can handle what you want to use. If the float indicator goes OFF again, then you cease using that appliance unless it will only be brief.

   Float will come much earlier in the day once the new arrays go up, but I'm also planning to add air-con later so what initially may appear to be overkill in array sizing will be less so later on.

   Thanks for all the input and grey matter prodding.

   Posted Thursday 8 Jan 2015 @ 12:06:42 am from IP #
9. 

   johnmath
   Member

   PVs are essentially constant voltage current sources, where the current is proportional to insolation. That means the zener diode isn't strictly necessary and you could simplify the monitor system even further by just using a resistor value that matches the PVs optimum Vout / Iout.

   For example, if you had a 10 watt panel with Voc of 21.0 volts, it's Pmax voltage will be around 16.8 volts at a current of .65 amps. This represents a load resistance value of R = V/I = ~ 26 ohms.

   If you were to connect the above panel to a the nearest readily available resistor value of 27 ohms, at 10 watts, and measure the current through the resister, the current as a proportion of Pmax current would give you a very close approximation of available power under current insolation. Then just scale up to estimate the big array's power potential.

   The reason I suggested a very small panel is simply that the less current you are measuring, the less power your monitoring system has to dissipate. There is no penalty on accuracy by using a small PV, you just have to scale measurement to the large array correctly.

   Posted Thursday 8 Jan 2015 @ 1:37:28 am from IP #
10. 

    wlb
    Member

    Thanks John,
    I thought getting rid of the extra heat from bigger panels was a small price to pay for the convenience of easier scaling. Nice and simple to just shift the decimal point on the display. But it's just occurred to me that it would be easier to use an analogue meter and just recalibrate the scale. That makes the choice of sampling PV easier, and they can be smaller.

    Posted Thursday 8 Jan 2015 @ 1:54:29 am from IP #

RSS feed for this topic