LUA – Virtual Solar Battery: simulate the presence of a battery for storing harvested solar energy

[megamarco83]

hi, i modify the code like this:

Code: Select all

--Manually create and/or define below (newly created virtual) devices and 'User Variables' by user given names
local electricity_meter = 'Immissioni'					-- Existing P1 Electricity device name in 'Utility'

--To be created virtual devices in the hardware section of Domoticz:
local solar_battery = 'Virtual Solar Battery'				-- (1) Custom Sensor name for the 'Virtual Solar Battery'
local solar_battery_idx = 73
local battery_use = 'Solar Battery Usage'				-- (2) RFXMeter counter name for the 'Solar Battery Usage'
local battery_use_idx = 74
local lost_energy = 'Lost Solar Energy'					-- (3) RFXMeter counter name for the 'Lost Solar Energy'
local lost_energy_idx = 75

--User Variables:
local battery_capacitylevel = 'SolarBatteryCapacity'		-- User Variable name for the 'Solar Battery Capacity'. Type 'String'. Value in Wh (4kWh battery = '4000')
local total_electricity = 'TotalElectricity'		

Image1.jpg (52.5 KiB) Viewed 4075 times

Image2.jpg (29.17 KiB) Viewed 4075 times

Image3.jpg (17.66 KiB) Viewed 4075 times

i write as local electricity_meter = 'Immissioni'


but i receive this error:

2018-11-29 15:01:30.038 Error: EventSystem: in virtual battery: [string "--Script_time_solar_battery.lua..."]:65: attempt to perform arithmetic on global 'sPlusT1' (a nil value)

[megamarco83]

this is how are defined my electic sensors:
and i choose "immissioni" to put in the script, that are the amount of energy that i put on the grid (only positive value, if i take from the grid, i will have immissioni=0 and "prelievi" that show the value of enery take from grid.
i have also "produzione" = enery produced
"consumi" = home energy consumption (without considering the production, the "pure" consumption of the house)

Image6.jpg (37.55 KiB) Viewed 4075 times

Image7.jpg (74.43 KiB) Viewed 4075 times

[stuiow]

I don't have a p1 meter, but i do have 2 z-wave meters, one recording grid energy usage and another monitoring solar energy.

Can i make use of these meters to simulate a solar battery?

I know very little to no LUA at all, but hoping someone can help me with this little project to get me started.
I have all these z-wave sensors and without using LUA, i am not making the most of them. Really need to get programming with it.

[crazynight]

I thought I had set everything up right, but getting this error.

2019-06-09 19:13:19.987 Error: EventSystem: in script_time_Solar-Battery: [string "--Script_time_solar_battery.lua..."]:65: attempt to perform arithmetic on global 'sPlusT1' (a nil value)

Any guidance would be appreciated, this is a really great idea almost a try before you buy scenario.

Thanks

[jake]

I thought I had set everything up right, but getting this error.

2019-06-09 19:13:19.987 Error: EventSystem: in script_time_Solar-Battery: [string "--Script_time_solar_battery.lua..."]:65: attempt to perform arithmetic on global 'sPlusT1' (a nil value)

Any guidance would be appreciated, this is a really great idea almost a try before you buy scenario.

Thanks

Do you use the original script from the opening post?
Do you have a p1 meter that outputs 4 and 2 values every time? (Delivery high and low and Return high and low, plus current delivery and return power)

[jake]

I thought I had set everything up right, but getting this error.

2019-06-09 19:13:19.987 Error: EventSystem: in script_time_Solar-Battery: [string "--Script_time_solar_battery.lua..."]:65: attempt to perform arithmetic on global 'sPlusT1' (a nil value)

Any guidance would be appreciated, this is a really great idea almost a try before you buy scenario.

Thanks

Do you use the original script from the opening post?
Do you have a p1 meter that outputs 4 and 2 values every time? (Delivery high and low and Return high and low, plus current delivery and return power)

I have access now to my own script, since I'm back home.
Although you haven't answered my previous 2 questions, I guess the answer on nr 1 is 'yes', since line 65 is about sPlusT1 calculation.

Do you have a P1 meter with multiple outputs? What does your electicity device outputs when you look it up in the menu 'Setup' - 'Devices'? I search on 'P1', since that is the type of hardware. My device name is 'Electricity' and the output is 6 numbers, seperated by a semi colon sign.
Did you type the right device name on line 13 of the script? Take care of capital letters, because they make a difference

[JPSke]

@Jake,
very nice script, thank you.
It could be useful to take the inverter maximum power in account.
When using battery power, that is the limit and when the house needs more power, surplus must come from the grid.
This means that less battery will be used and in some occasions the battery will not be completely empty, when the next day PV production restarts.
This could also give an indication which inverter to use: f.i. a 3.3 kW will be cheaper, but a 5 kW will be more efficient. But just how much and is the extra investment paying back?
KIndest regards,
JP

[jake]

@Jake,
very nice script, thank you.
It could be useful to take the inverter maximum power in account.
When using battery power, that is the limit and when the house needs more power, surplus must come from the grid.
This means that less battery will be used and in some occasions the battery will not be completely empty, when the next day PV production restarts.
This could also give an indication which inverter to use: f.i. a 3.3 kW will be cheaper, but a 5 kW will be more efficient. But just how much and is the extra investment paying back?
KIndest regards,
JP

Yes, this is an interesting approach. I had thought about it before, but never came around to implement it yet. Currently the script checks once a minute for energy change on the meter. When there is a power peak in between, I won't be able to catch it. However, I think that the effects will be limited.
For myself I have a dzVents version of the script running for about a year. I will add the feature in this version and publish the improved script.

[JPSke]

@ jake
Thank you for considering.
I agree, in most cases the PV production will be in line with the battery charging limit, especially when using a Hybrid inverter for PV strings and battery combined. In my case however it is a bit different: we have two PV inverters, one 5 kW and a newer one 4,4 kW. On a sunny day we produce 9 kW or more between 10 am and 3 pm. The inverter we hope to install for the battery will be AC-coupled (no PV strings attached) and to remain affordable not more powerful than 5 kW. So right now the script says the 15 kWh battery is full around noon, but that is not realistic because the 9 kW will be topped to 5 kW for the battery and the rest will go into the grid. It would be nice to have a more precise charging schedule and a correct value for "solar energy lost". For discharging the same goes, but comsumptions over 5 kW are rare, so the error will be much smaller.
JP

[jake]

Hi JP, your message is a little confusing to me, I wonder if you use kW for both power and energy.

I have 1 initial question: when a battery inverter had a power limit of, let's say 2kW is that both for charging and discharging? Or is that different from each other?

[JPSke]

Hi Jake,

I use kW for power and kWh for energy. I will try to explain my line of thought more clearly.
If you install a battery system, there are two power limiting factors: the maximum power of the inverter and the maximum power of the BMS (battery management system). The inverter has a fixed maximum value which is the same for charging and delivering, but the BMS value can fluctuate, because it is depending on battery condition and temperature.
Lets forget about that for the moment. I plan to buy an hybrid inverter for 5 kW, so as long as my two PV-systems combined are producing not more than 5 kW on top of my consumption, everything goes into the battery. On a sunny day, around 10 am the production surplus exceeds 5 kW, so 5 kW goes into the battery and the rest goes into the grid (generating a fantastic profit of 0,04 € for 1 kWh, while when you have to buy the same kWh during the night you pay 0,21 €). After about 4 hours of full production, the battery (15,6 kWh) is full and from then on, everything wich is not consumed directly, goes into the grid. So the 'lost solar energy' value is a combination of both the topping while loading and the production surplus after the battery is full.
For your info: right now I have 5 offers for battery systems between 10 and 19 kWh with prices between 528 €/kWh and 860 €/kWh.

[jake]

Thanks @JPSke for the explanation. I was confused with the word 'inverter', because that could have to do with the solar inverters as well.

The introduction of the inverter 'power' adds the factor time to the script. Initially it didn't matter much when the script update and P1-meter update were 'out of sync', since it is simply a matter of comparing totals. Now it is important to check on the generated/consumed energy per time interval.

Changes I am making to the script:
- Store previous values in a history aware persistent value
- Replace the script trigger from 'time' to 'device'
- addition of a script trigger interval (per default I use 60 seconds, the script is only run when the last script execution >= 60 seconds ago)

And yes, the energy that will be above the capacity of the inverter will be counted as 'lost energy'

[JPSke]

Hi jake,
thank you for scripting.
The inverter power factor will also limit the use of the battery while delivering. If the house needs more than the inverter can deliver, the rest must come from the grid, making the battery charge last longer. However, this will not make a huge difference, but if you can take it into account it will make the script more accurate.
BTW: I do not have a regular P1 hardware device, but my P1 port is connected to an Arduino, which reads meter values (every second), calculates delivery, consumption, quarter hour peak evolution and synchronizes with the Smart meter's internal clock (for peak accuracy). The data is sent once every minute to the Domoticz Raspi over a MySensors USB interface.
This means I had to make changes tot your script and get the meter values from variables instead of P1-devices.

[jake]

My 'solar battery usage' is showing 3220kWh. In the log page of the device on the right top corner I can check when I started this device: 23-02-2017. 4 year and 1 month. When I subtract March of this year (40kWh), I have 3180kWh in 4 year, roughly 800kWh per year.

When new legislation will allow the energy companies to only pay a fee for the returned energy (more or less 6ct/kWh) it means that every kWh usage has a value of 21-6 = 15ct

800kWh x 0,15 Euro/kWh = 120 Euro per year in savings while simulating a 5kWh battery. This is a long ROI when I see what a 10kWh is costing per kWh. I asssume a 5kWh system will be more expensive per kWh, so 528Euro/kWh --> 600Euro/kWh?

ROI is 5kWh x 600 Euro/kWh = 3000 Euro / 120 Euro/year savings = 25 years. Time for new batteries

Only when it is possible to sell energy to the grid for a good price, it might be useful to invest (more) in batteries. For now a bigger battery than a 1-2 days worth of consumption makes no sense, since you can only top up a battery once. Only when you can sell the daily surplus to the grid, it will be worth to increase the battery size in order to 'catch' more/all solar energy on a daily base.

[JPSke]

Hi jake,

my situation is completely different. My yearly production is 11.500 kWh and last invoice period my consumption was 4.500 kWh on top.
We have an EV (charger 11 kW), no gas, so a heatpump for central heating, a solar domestic water system with a 3,1 kW heater element, dishwasher, washing machine, laundry dryer, kitchen appliances (steamer, airfryer, hotair oven, microwave, induction hob...), additional electric heaters in shower and bathroom.... Now the meter is not turning backwards anymore we will have to pay for 70 % of our own production, because the year round self comsumption of the PV system is 30%. This means 11.500 x 0,7 x 0,17 € (I get 4 cent for delivery and pay 21 cent for consumption), so 1.368 €. After subtraction of the "prosumertariff" of 680 Euro, this gives an extra cost of 688 €/year. With a battery we can reach a self consumption of 60%, so the extra cost will be 782 € - 680 = 102 €.
Profit: 586 € /year. Cost of the system minus government premium: 6.700 €. ROI: 11 years. Not excellent, but, 2 buts:
1. if you have the money in the bank, they pay you 6.700 x 0,21 % = 14 € per year (compare that to 586 €). If you have to make a loan, that is a different story, but maybe also profitable).
2. from next year on, in our region, the distribution costs will be peak dependant. Without action our monthly quarterhour peaks will be around 16 kW, meaning we will pay 5 times more for distribution costs than an average consumer or one without a smart meter. With smart switches and other domotic tricks I have been able to limit peaks to 6,5 kW, but with the battery I can shave peaks to 3,5 kW, which will also give a saving of something like 200 €/year. The overall picture is mildly positive, in my opinion.

[jake]

Wow, what a differences from country to country. I assume that you have taken into account that during the year that a majority of consumpton is during the winter, while generation is in the summer? Is your EV around the house during peak moments of energy production?
I have seen an API on this forum where the user was able to control the charging capacity of his EV, also to reduce the chance of using the grid for charging.
These peak moment checks, can they be forcasted? In other words, can you disconnect or reduce consumption during that check? Or will you try to smooth the electricity load over the whole day (for instance, stop/reduce charging the EV when the heat pump / oven is on?)

Will the government also differentiate in price based on production/consumption on the grid? In that case it might be of interest in the winter months to charge your battery from the grid as well on cheap moments.

Your financial calculations would look OK to me. I also had/have a 10 year ROI on my solar panels in 2012 (I just checked: in total 14.054kWh x 0.21 = 2900 Euro, which is just above my net investment of 2800 Euro (in total I had +/- 1200 Euro of subsidies). Ha, I didn't even realize that I am past my initial investment. From now on this means free energy!!!)

[JPSke]

Here in Flanders our government has gone crazy. They are saying they want a green future, but are harassing citizens who are trying to realize it.
Since March 1st the kWh counter is not turning backwards anymore for those who have PV and a digital counter (100.000 cases). The other 300.000 PV owners without a digital counter still benefit from the fact that the Ferraris counter is turning backwards and will not get a digital counter before 2025. That says something about everyone equal for the law and discrimination.
The new trick to extract money from our pockets is the capacity tariff. Consumption cost will remain based on the amount of kWh, but distribution costs will be calculated according to the quarter hour with the highest consumption in one month. An average of the last 12 months is used for the calculation.
There is no difference between a day or night peak, so charging the car overnight to avoid peaks, is not an option. Only one quarter with high consumption in each month is enough to double, triple or even quadruple the cost for distribution, which is already very high. The 21 cent for 1 kWh only contain 4 cent for electricity production, the rest are taxes and distribution costs.
They say they want to avoid investing 5.000.000.000 Euro's to adapt the grid for the future, so everyone must contribute and those who burn no fossil fuels and have an EV and use only electricity are targeted, while the polluters are left alone and in some cases even are sponsored to pay their fuel bills.
So far for the frustrations.
We are lucky to have been retired so we can charge the car during the day when the sun is shining and use our appliances when there is plenty of power.
I use the Arduino to forecast the peak. The P1 port gives me data every second. The peak counter is synchronized with the meter's timestamp (which is ± 20 seconds behind world time in most cases). The value of consumption for each second is divided by 900 (number of seconds in a quarter hour) and added to the running total. That plus the average of consumption in the time past for that quarter hour x the remaining seconds gives me a prognosis. Domoticz is than used to switch devices off or on or stop charging the car to keep the peak under control. So far I was not able to get it under 6,5 kW, but I hope by fine tuning the system and the aid of the battery to get it under 3,5 kW. The exercise I am doing now is to find out which size of battery will be the most cost efficient.
The high consumption in the winter and the high production in summer is something that can only be compensated by using another power storage system, like hydrogen, but that technology is not in reach for hobbyists right now and maybe will never be. That is why I am so mad with the government. For thirty years they are debating about closing the nuclear plants and investing in renewable energy and now that it is going to happen, the grid is not ready. Stupid politicians!

[jake]

So, if I read you correct, the peak tariff is based on highest energy consumption (kWh) and not on highes Power (kW). From a technical perspective Power would be more reasonable, but using a 15 minute energy consumption averages it out to avoid a single peak to base the tariff on.

However, I assume you will replace water boilers in the kitchen by camping versions (of 600-900W) The same for other kitchen appliances? Just kidding.
Altogethre quite a hassle I guess to work around it. Your solution is good for your wallet, but for the energy company it doesn't have a technical benefit. You consume the same amount, at the same rate (kW), but split it on/off over many quarters. Many appliances don't even give an option to use different power levels, except microwaves and cooking stoves, although the latter need a certain amount of heat to prepare the food properly. An oven, on the other hand, is in my case also an 16 Amp on-off device (year 2008 model), but that would help to change that in a more continous lower power level.

Are you able to control devices in the house by altering their power level, or is it a brutal on-off by a smart socket or intermediate plug? I wish our freezer would have a controllable thermostat, for instance. During peak moments of solar production I would lower the temperature and at night I would go to the upper (food safe) limit.

In my house we also use somewhere around 30% of our own solar energy. High consumers like an induction cooking stove are used when after the solar system has it's peak production. During those peak hours only the fridge/freezer are continous consumers, plus some electronics here and there.

[JPSke]

The peak is the sum of all power that is consumed in 15 minutes, so it is kW, but you can calculate it by using the kWh for those 15 minutes and multiply that by 4. That is because the h in kWh contains 4 quarters. If you switch a 1000 W stove on for 15 minutes, you will have a peak of 1 kW but a consumption of 250 Wh. So multiplying the kWh by 4 gives the kW peak. On the other hand, theoretically you can have a consumption of 15 kW during 1 minute and zero for the other 14 minutes and that would also give a peak of 1 kW and a consumption of 250 Wh. So it is not the momentarily power peak that counts, but the average of that quarter hour. What you put into the grid during those 15 minutes is neglected, so it is not subtracted from what you consume. That is only possible with a digital meter, because it has separate counters for production and consumption. Again, those who still have an analog counter have nothing to fear, they pay an average distribution fee and can peak whatever they want. I am realizing now that one of the biggest mistakes of my life was to open the door for the guy who came to install the digital meter.
Just because I cannot control kitchen appliances, I cannot get the peak lower than 6,5 kW. You want your fries to be ready, together with your steak, so switching the oven on after the hob has finished is not an option. That's mainly why I need the battery to give some extra juice when there's nothing that can be switched off. The dishwasher, laundry drier, washing machine and the additional heating all have smart switches and can be switched off by Domoticz. The boiler and heat pump also are switchable by interacting with their controllers. The car charging cycle can be interrupted through an API call, so that is also switchable.
So, to answer your question: no power level adjustments, just switching on and off. It is working Ok, because those things can be on hold for some time without creating other problems. The machines are not programmable, but because we are at home most of the time we can start the dishwasher, laundry washer and drier by hand, when there is PV production, and Domoticz will switch them off temporarily when a peak is eminent.
Freezers and fridges are only small power consumers, so I think there is not much to gain there. It is mostly the electric heating element that makes your counter turn.
Edit: today I have not seen the sun once and it has been raining most of the time, but still we have produced more than 10 kWh and put 4 kWh into the grid. That would have gone into the battery, to be used later this evening. Our self consumption was 60% and would have been 100% with a battery.

[jake]

In the mean time the script is adapted.

- After reading the new values from the electricity meter and comparing them with the previous ones, the result is the amount of enegy that went pass the meter
- With the power of the inverter x (time previous script run from history in the persistent dzVents data / 3600) I calculate the maximum energy that can go in or out of the battery
- When the max value is smaller than the energy that went pass the meter, I will reduce that energy to the max. inverter envery. The left over portion energy is automatically 'lost energy'
- This 'lost energy' comes on top of the 'lost energy' from either a full or empty battery (the latter only when this is enabled in the script setting)

I will add some new devices and run the script in parallel to my original one to see if the 'inverter cap' does the job. I should see less response in the battery and I should also see 'lost energy' at times that the original script doesn't.