Utilizing excess energy from solar panlel - not a dumpload question ;-)

[solga]

Some people in this and other forums make a distinction between a dump load, considered as a place to either use OR waste power to prevent overcharging of a battery bank or to keep a load on a wind or hydro turbine to keep it from running out of control when no power is actually needed. For wind, hydro, or shunt type battery charger, this load is not optional it is required.

When the specific goal is to take available power that would otherwise be wasted, as for example, the extra PV potential output when an MPPT controller does not need to use full output, it is more often called an opportunity load. It is completely optional in that if it is turned off or disconnected, nothing bad will happen to the system.

The most common opportunity load is water heating. This can be done directly with DC from the power source, but that requires finding heating elements with non-standard voltage and resistance specifications to make best use of the power. A few people have put forward the idea that as long as you have an inverter running off the batteries anyway, you can just as well power your opportunity load from the inverter whenever the battery bank is fully charged. The input current to the inverter will simply come directly from the panels rather than forcing a charging and discharging of the batteries.

The simplest way to control this is to use an output from the CC which indicates that it is in Float to drive a relay which energizes the heating element. That will not let you use surplus power during Absorb. That would be more complicated.

An idea which I have not seen explored would be to take off the excess power before the MPPT CC by sensing the current being drawn from the panels by the CC and pulling additional current into an opportunity load when the CC current is below the current Imp value. But that would require some way of knowing how much light is hitting the panels, or at least what the current Imp value is.
And too rapid a response could cause instability in the MPPT algorithm of the CC.

If you do a google search for power management systems, you will eventually find a company that sells something to the UK market that regulates the opportunity load to exactly null out any power returned to the grid in a grid tie system. The only reason this makes sense it that the UK tarifs pay the user more for each KwH generated and used on the premises than for the same amount of power generated and returned to the grid. So the concept is around, just not apparently in the context of off-grid.
At some point the extra complexity and risk of failure is not justified by the energy saved. Also, it requires the CC and the inverter to be closely coordinated.

The problem with a resistive load, of course, is that it has a fixed rather than a variable resistance. So you have to either control the voltage applied to it or switch the load on and off in multiple segments. The best way to deliver a variable power to a resistive load is via a DC-to-DC converter or an inverter combined with an SCR type current regulator.

And, no a simple diode network cannot do this since the CC needs to be able to draw power back from the battery at all times and sense the battery voltage, while the inverter needs to see constant power or trip out from low battery voltage or overload on its output.

I have a similar need in Ga. GTI. NO COMPSATION for excess generation, so retaining generation to use in load is very important. Other wise a loss

[Stu]

Why can not the man put his heater on full 500W and then let all the dump or extra load go directly from the panels to the heater? Ok this would probably make the heater jump in all directions (hotter and lower) but it would not explode?

Since the batteries are still going to need a float charge available continuously, you would have to figure out how to send some of the panel output to the controlller, and some to the heater. And it would be changing in real time. Not to mention the Absorb stage opportunity would be missed. I had considered taking this approach at one time, but now that I see how hard the Outback seems to be working to maintain the right voltages to battery I am glad I left this control in the hands of the professionals haha!

[Roil]

To the question of accuracy and milliseconds - is it necessary to be so spot on?

How about the following.

The Arduino pays attention to the status of the controller:
- Charge state: Bulk, Absorption and float
- Battery voltage
- Watt to the batteries (or Amps)

As a test case we can set the following equipment:
1000W of solar panels
1000W of load (Heater) (proportionaly adjustable - maybe as a DC consumer via a inverter)
12V battery - but let's focus on watt and watt hours and not volt amps (yeah I know should have been 24V)


When in bulk - Do nothing

When in Absorption pay attention to Watt every 5 min, the start watt in absorption mode indicates available solar energy at this time (and should probably stored as value) but of course if could rapidly change in both directions), when watt has been reduced with let say 50W, turn on the heater and apply 10% ( of the 500W example load) and measure watt again. If Watt into the batteries has increased with 50W then we now the sun has been able to cope with the additional load. This last parts should then be a loop, and could run every 5 min and adjust up and down accordingly.

While in absorption mode the natural max load will probably be in the area of 90% of the solar panel max power.

In an increase in load make the regulator move back to bulk, we now we have applied to much load, and can adjust it.

I guess my main idea here is to allow the load to be both "too low" and "too high", allowing some energy to be wasted, and some energy to be drawn from the battery bank.

[Roil]

Why can not the man put his heater on full 500W and then let all the dump or extra load go directly from the panels to the heater? Ok this would probably make the heater jump in all directions (hotter and lower) but it would not explode?

Heater is 220V AC, but even if it would have been a AC 12V heater the Tristar MPPT need a battery as powersource, and can not be replayced by a load like a heatr

[rainworm]

heater

Why can not the man put his heater on full 500W and then let all the dump or extra load go directly from the panels to the heater? Ok this would probably make the heater jump in all directions (hotter and lower) but it would not explode?

[Olberg]

Here is a reply from the manufacture on the subject:

"When used with a PWM control, the Power-io will act in a similar fashion to a variable switch. For example, if connected to a visible light bulb, at a very low PWM speed, you will see rapid ONs and OFFs. If connected to a heater, these ONs and OFFs determine the amount of heat rise, such as you have inside an electric oven in your house.

At higher PWM, you eye will not see the light bulb go on/off/on/off. You will see the average light, even though the bulb is actually going on/off/on/off rapidly. Your eye is only accurate up to about 100 Hz. (An AC lightbulb in your country is probably running at 50 Hz, but we consider that 100 Hz switching since there are 100 half cycles in a 50 Hz power line.)

Our product can go on/off/on/off up to about 15-20kHz speed. Our maximum speed is obtained with a strong control signal.

Your Arduino is a medium strength control signal. At about 5V, 40mA, you will be able to PWM us at about:
50 microseconds ON, 80 microseconds to remain ON, then OFF. Thus, you will be able to reach a maximum speed of about 1-2 kHz, which is still incredibly fast.

Power-io is often used with solar applications. I have been told that we are part of an interesting system in the Antarctic.

Your attached PWM chart is accurate. First, you determine your overall PWM speed (above) then the Arduino determines how much of that cycle will be an ON and how much will be an OFF."

Stu:
Yes. Programming it will be a challenge!
It must be in real time. I will have to continuously read the Canbus. Use the incoming PV current from the cc and read the battery voltage and current.
Pv current (recounted to match read battery voltage) - Battery current + current to the SSR( opportunity current) = 0 is my theory.

/O

[Stu]

[ATTACH=CONFIG]2808[/ATTACH]Stu: Fantastic with web monitoring like that. Something I certainly will look into after I’ve installed my system.

Stu, Inetdog and Roli:


Am I thinking totaly bonkers or could perhaps the solid state relay Stu’s using (lika a HDD-06V75) be controlled directly with PWM (analog pin) from the arduino?


/O

No problem from what I know. The challenge is knowing how to determine the PWM rate. The Outback takes care of this, insuring the Absorb and Float are always maintained at proper voltage. From watching my system in action, I can say that this is a constantly adjusting real time effort as I watch the PWM rate vary on the display of the Outback. You could replicate this of course with the Arduino by watching the voltage at the battery during these charge cycles and increasing the PWM rate until voltage drops then back off until it returns. Feels like some complicated programming though to insure you don't start creating some crazy feedback loops! Think about a cloud passing in front of the sun for a second and imagining how you would accomodate this in real time without compromising the charging voltages.

[Roil]

Nice small detail on these to screenshot.

Float level and still high output from the solar panels.

Only difference between them: A load of 200W via a Victron inverter, still manually remotely applied thought ;-(



TriStar MPPT - Live Data (1).JPGTriStar MPPT - Live Data.JPG

[Olberg]

pwm.gifStu: Fantastic with web monitoring like that. Something I certainly will look into after I’ve installed my system.

Stu, Inetdog and Roli:

I´m still a bit in the dark here but perhaps you guys could help shine some light and fill in the gaps:

Am I thinking totaly bonkers or could perhaps the solid state relay Stu’s using (lika a HDD-06V75) be controlled directly with PWM (analog pin) from the arduino?


Arduino: 5V max 40mA
SSR: Min turn-on voltage/current 4.25 VDC / 10 mA


When it comes to controlling the load:

Arduino:
Arduino's PWM frequency at about 500Hz, the green lines would measure 2 milliseconds each. A call to analogWrite() is on a scale of 0 - 255, such that analogWrite(255) requests a 100% duty cycle (always on), and analogWrite(127) is a 50% duty cycle (on half the time) for example.

SSR:
“High frequency switching test: 100% of relays are tested at 15 kHz, @ 50% duty cycle, @ 10V or greater control input, at >10mA, and a 10 volt or greater dc load. This generates a PWM switching speed of 33 microseconds "on" and 33 microseconds "off" or 15,000 ons and offs per second. With a strong control input signal, this provides: 15kHz PWM speed in the HDD-06V75 to HDD-6V15 models”

Is the 15kHz rating a maximum frequency and if I send a PWM @ 500Hz it will work?
Can a immersion heater be a PWM load?
Will the SSR work like a proportional control when used it in this way?

All the best
/O

[Roil]

Anders:

Olberg:

One obvious choice would be to drive a solid-state relay that has optically isolated control inputs. You could also drive a standard relay for on-off control of the opportunity load.
If you want to do proportional control, you could use a solid state relay or SCR/Triac circuit that is controlled by a pulsed waveform.
If you want to switch DC to the load, you will have a harder time, since you need to find DC rated switches that can interrupt the current on their own instead of AC rated switches and relays which count on the current passing through zero 120 times per second.

I have done some search for a proportional control unit. The all seem to have a cyclic time from around a minute up to half an hour. If load is set to 50% it sounds like the full load is connected half the time, so 30 sec full load, 30 sec no load. Am I lost?

[Stu]

The Arduino is a pretty handy unit for controlling a system if you like to write software and dabble in electronics.

Mine has a wifi board attached and sends sensor data to the internet through Cosm which I interface to a web page. Fun to be able to monitor the system anytime/anywhere. My system is pretty small and just intended for learning so I like to watch it work. It just powers my one room home entertainment center. You can see the data in real time here if interested:

http://www.4sightinc.com/stu/Solar.htm

Note the water temp reading from the heater is down lower on the page, we have had two cloudy days in a row now so water is getting a little cold!

[Stu]

Hi Stuart

Nice solution - Outback got it then!

I am looking for the same reward

Should I swap controller then? - hmm

How about some pics and specs of the solid state?

Anders

I really haven't looked into the other controllers so have no idea if they also support this. The outback is specifically designed to interface to a solid state relay this way.

I went with one of the recommended relays in the post I read, it is the HDD-1V50E-HS1.0
For my setup I have 600w panels, but only about 500w of "free" energy max at float so the 50amp rating was good for my 12 volt 600w heating element. Using higher voltages for the heating element would probably allow you to get away with less amperage required of the relay.

You can see/purchase it here:

http://www.power-io.com/products/hdd.htm

I know there are cheaper versions on ebay etc. but with all that current I decided to go with what looks like a pretty solid unit.

I got the one with the heat sink already attached. improper heatsinking of these things seems to be the greatest source of failure. The one it comes with is huge! I notice though when operating at max output it does start to get a little warm so sure wouldn't recommend doing without it.

Even though the relay goes directly to the Outback AUX jack, I have placed a low voltage mechanical relay between them so that the arduino can control whether the heater is on or not even if the outback is putting out the signal. This lets me turn it off if the water gets too hot.

[inetdog]

I struggle to understand your last part, but it sounds likt it is better to let the load be AC thus making it possible to be fully variable? Did I get it?

You have found the essence of what I said.

[Roil]

Now we are talking - yes please one of those.

By pulling the details from your controller, you case CANBUS, my case MODBUS (Tristar), and wright a software capable on computing how much power we can pass on to the opportunity load.

Sounds great.

I struggle to understand your last part, but it sounds likt it is better to let the load be AC thus making it possible to be fully variable? Did I get it?

[inetdog]

Hi Olberg

What is an "arduino"?

Anders

Anders:

An arduino is a small computer which is designed by some folks in the open software community. It is very inexpensive and easy to program compared to industrial process controllers and there are a wide variety of input and output adapter boards for it to allow you input almost anything that can be measured and control almost anything that can be controlled by a logic signal, a relay, or an analog voltage.
You can get assembled arduinos in a variety of price ranges and capabilities, or build you own.

Check the internet and these forums and you will see, for example, people who are giving away software written for the arduino which can interact with Modbus, Canbus and other inverter/charger control protocols.
It does help to know at least the C programming language if you are going to set it up on your own.

Olberg:

One obvious choice would be to drive a solid-state relay that has optically isolated control inputs. You could also drive a standard relay for on-off control of the opportunity load.
If you want to do proportional control, you could use a solid state relay or SCR/Triac circuit that is controlled by a pulsed waveform.
If you want to switch DC to the load, you will have a harder time, since you need to find DC rated switches that can interrupt the current on their own instead of AC rated switches and relays which count on the current passing through zero 120 times per second.