[Pops]

So in the back of my mind is that little Solviva style home/shop. I have all sorts of passive solar ideas but I'm thinking here about using some active measures for storage and control. In my limited experience building passive, the big problem is overheating as well as uneven temps (both time and zone) so I'd like to incorporate (in this imaginary project) some active transfer, basically using infloor hydronic tubing to move and store heat from place to place and time to time.

Pretty straight forward. I realize passive solar is as simple as exposure and storage, but, never one to leave well enough alone, I'd like to be able to program some simple logic using inputs from various sensors, in combination with a calendar and clock to anticipate and smooth out the swings and improve efficiency and comfort.

So the question for you computer geeks out there is how would I set this up without going broke and learning to speak Machine?
I can get around in windows and Mac OS but not linux. I don't know any code aside from some basic basic from about 1980.
What hardware would I use?
How would I get the inputs input (various remote temps, solar strength, maybe PV output) and the output to controls (pumps, valves, fans, ???) This part especially I don't have any idea about...

Little help?

[autonomous]

The application you describe is more suited for an embedded system running linux instead of Mac or Windows pc. An embedded system provides more options for interfacing with sensors and control relays, it is more cost effective and uses very little power.

Here is a new mini-linux board for $35 where you can program in the language of your choice. 700mhz Arm cpu and 512mb ram
http://www.newark.com/raspberry-pi/raspbrry-modb-512m/model-b-assembled-board-only/dp/43W5302
Specs on gpio:
http://elinux.org/RPi_Low-level_peripherals#General_Purpose_Input.2FOutput_.28GPIO.29

Another option if you really want to avoid linux is a wireless network that interfaces to your PC:
http://en.wikipedia.org/wiki/ZigBee

[autonomous]

This is a really great subject, particularly when it comes to living off grid. I can provide more information about sensor and control arrays.

What would be useful here is to have a better idea of what you are designing, do you have more details on how it will operate?

[Quinny]

Have you looked at the Arduino, it seems pretty modular and there are pre-built modules for lots of interfaces. I'm a Windows guy, but I agree that Rasberry Pi and Arduino combination is probably a good place to start looking especially if low power consumption and cost are important.

I'm currently looking at designing a system with Raspberry Pi/Arduino control/sensor units with a central monitoring server on a shiva plug.

[Pops]

auto,
The floor would be concrete slab with a typical embedded PEX tubing and water/antifreeze mixture, zoned by living area and solar exposure and one or more types of backup heater.
There would be temp sensors in the floor: in the areas of insolation and in the areas without.
Possibly air temp sensors inside and out
Maybe some way to gauge insolation directly, I guess that could be simply PV array output.

Operation would be basically moving heat around to balance the slab temp over time and between zones. And importantly, anticipating what will happen in a few hours since a system with a lot of mass takes a while to respond.
It would incorporate requirements for additional heat or cooling (using air temp sensors and calendar rules perhaps) and maybe even incorporate some type of geothermal cooling via shallow ground loops.

Thanks for the links.

[autonomous]

So here are a few ideas for interfacing sensors and relay switches to the Raspberry system:

For $4.50 you could buy a Maxim high-precision digital thermometer and thermostat, wire it to the I2C interface and program it with basic on the Raspberry:
http://www.mouser.com/ProductDetail/Maxim-Integrated/DS1631+/?qs=sGAEpiMZZMucenltShoSnu2ltHLyJVLbMfnFv9EmYFg%3d

$20 for an 8 channel relay board, interfaced to microcontroller boards such as Arduino or the Raspberry via the I2C interface.
http://www.elecfreaks.com/store/8-chann ... p-268.html

Low cost pyranometer (PV array output could be affected by the charge controller or loads):
http://www.instesre.org/construction/pyranometer/pyranometer.htm

A breadboard kit and the above three items should cover the peripherals needed, so the whole project would cost under $100. You could then write some code in basic or whatever language you like to scan the temperature sensors and toggle relays connected to fans, pumps or valves. Monitoring the temperature of the concrete slab would require a little more research, but it would essentially involve some thermal paste and positioning the thermal sensor chip in direct contact with the concrete slab.

The Raspberry is an amazing value, since it comes with a 32bit CPU, large amount of memory for an embedded system, video display driver, ethernet, usb and so on. I would setup a simple web server on it to monitor and control the system over the network.

I actually need to build a system like this for my greenhouses, which also includes a pyranometer. Very similar application to what you describe.

[Quinny]

Thanks autonomous, you've saved me a hell of a lot of research You are boviously ahead of the game here.

Have you (anyone) done any research into different sensors eg temperature, chemical, humidity, water acidity.

I am hoping setup a 'self' managing Greenhouse/Aquaponics system

[AgentR11]

I'd vote more along the lines of having the sensors read and relays controlled from a windows or linux commodity machine (doesn't even need to be seperate); even without zombie apocalypse, specialized CPU boards get hard to replace a few years after their production ends; whereas, Ethernet, even ancient, 10BaseT is still very much with us. Keeping the controlling code on hardware that will always be easily replaceable lets you adjust the modular pieces of code that read sensors or toggle relays; without having to mess with the main body, logic, and ui of the application.

so you end up with:

double read_temp(int dSensorNumber)
{
return blah;
}
int read_relay(int dRelayNumber)
...
int set_relay(int dRelayNumber, BOOL MakeItSo)

type stuff; a discontinued temp sensor dies, you buy whatever is available, and tweak the read_temp(), same with *_relay(), main CPU dies, you're gonna get another windows/linux machine anyway to replace it, so why use a dedicated CPU that might not even be available ten years from now.

[autonomous]

I totally agree with the need for using readily available hardware components. The software system should be architected to use well-established programming languages and frameworks that can be adapted to run on a large number of hardware platforms.

For greenhouses and habitable spaces, I was thinking along the lines of using electronic sensor and control processes to test and develop low-tech climate control systems. After a low-tech setup has been verified with a microcontroller system to operate within the desired parameters, it would not require electronic controls to continue to operate.

An example would be a composting greenhouse that provides heat and C02. Performance could initially be monitored and controlled with a microcontroller system, and over time, lower-tech parts could be substituted to replace electronic components. A lower-tech substitution might be a re-calibrated automotive cooling system thermostat which replaces a solenoid controlled valve and thermometer sensor. Low-tech parts like this would be much easier to maintain over time.

My last project involved developing an alternative energy monitoring and control system from scratch, where the purpose was to operate everything needed with as few solar panels as possible. I setup a complete off-grid electronics prototyping and fabrication shop where one could produce hand-made printed circuits in a few hours. This is along the lines of Kunstler, World Made by Hand.

The system consisted of an Arm-based microcontroller board and additional sensor device boards. I switched everything that required AC to highly efficient DC devices and was able to run everything needed from a 180 watt solar array. This included a refrigerator, swamp cooler, computer, sound system, and lighting.

It would be great to collaborate on a climate control system in this forum. There are a lot of great ideas floating around. This project should be open source as well.

[culicomorpha]

Interesting project, Pops.

I've done many aspects of this project professionally as an electrical engineer, and have a good amount of experience interfacing both analog and digital sensors in various kinds of feedback control systems. I've done plenty of both embedded and PC-based designs, but for a relatively simple system like this, I just can't see the logic of using a PC. For a handful of sensors and control relays, something like an Arduino or Basic Stamp is just fine. (I'll have to check out that Raspberry, that looks like a lot of functionality for not very much money) The other thing I think about all the time is power consumption of control systems. Even a small netbook consumes upwards of 20W, and that's 24/7/365 if it's controlling your house temp.

In terms of the thermal issues, from what I understand about passive solar structures, the big problem is in determining the proper sizing of the thermal mass. Too small and the temps are all over the place, too large and it takes a long time to respond. The other thing is that most passive solar houses use too many windows on the South-facing side, creating larger losses at night which exacerbate indoor temperature swings.

As far as the code algorithm is concerned, there are a lot of possibilities from simple to esoteric. The simplest might just be a couple of digital temperature sensors (DS1621 is easy to use) and a few relays to open valves or turn on a pump. A more complicated system could use a pid loop based on an array of inputs. Another possibility is to use hot water tanks or rock beds for heat storage, which could be called-upon as necessary to make up for shortfalls.

Although you describe the system as "active", one question I have is whether it would really be necessary to do much pumping of the water around, or whether there would be enough of a gradient to move things around largely without it. It's an open question in my mind. I was actually working on a very similar system to what I think you're envisioning, but unfortunately, life intervened (as sometimes happens) and I didn't get to complete it. I'm still planning to do something similar down the road, so I might be into helping if folks were into doing a low-cost open-source design. I think there is a fair amount of low-hanging fruit here in that a relatively inexpensive control system in conjunction with an in-floor heating system could revolutionize passive solar designs.

[Graeme]

Pops, Just saw this. Not sure whether it will help but I'll share anyway:

http://cleantechnica.com/2012/12/29/usi ... h-barrier/

[culicomorpha]

Thanks autonomous, you've saved me a hell of a lot of research You are boviously ahead of the game here.

Have you (anyone) done any research into different sensors eg temperature, chemical, humidity, water acidity.

I am hoping setup a 'self' managing Greenhouse/Aquaponics system

I've looked at what would be required for a aquaponics control system. Things like temperature and humidity are readily measured with a variety of low cost sensors (low-cost humidity sensors are not particularly accurate, but probably good enough, 5-10%-ish). Chemical sensors are more expensive, mainly because they require special probes and I haven't really found a good source for lost-cost sensors. It's been a while since I read the aquaponics papers (New Alchemy did a bunch of work on this in the 70s and their papers are still available online), but I think a few special sensors were also required (ammonia?) and dissolved oxygen.

There's a bunch of different sensors available here (http://www.parallax.com/Store/Sensors/t ... fault.aspx), including a number of gas sensors. Most of these parts are fairly easy to use and are for the most part well documented. Some more at Sparkfun here (https://www.sparkfun.com/categories/23), including a dissolved oxygen sensor. These sites are geared towards hobbyists, although most of these parts are used in commercial sensors, albeit usually with some added functionality like temperature compensation.

One thing I never really figured out about aquaponics is how different "bad" conditions are fixed. pH I understand, but other things were less clear. In the manual management of these systems, operators can tell when, for example, fish waste products are too high, and they will dump some of the water and replace with fresh water. Probably those sorts of things could be automated, but I think it would involve some manual trial and error to get the basic process down first.

[davep]

I've got a couple of Raspberry Pis and they're great little machines. I tried one as a desktop replacement but it's better as a media centre, email server etc. I'm looking at using one for rotating photovoltaic panels to follow the sun to improve overall efficiency and another for controlling watering systems. Without using the onboard graphics they use about one watt-hour of energy, which is phenomenally low. The market for GPIO controllers is obviously immature at the moment but I expect to see great things soon. Here's a starting point with some good references http://elinux.org/RPi_Low-level_peripherals

Edit: Oops, now I've actually bothered more than skim-reading beyond the first post I see Autonomous is way ahead of me...

[Pops]

Thanks all.

Back in the '80 I designed and built a house in the California Sierras with good passive features. CAs Title 24 energy regs are (or were then) a really good checklist for design as the code is a points system, ie: 1 point for each insulation R, -1 point for each SF of glass oriented north, +1 for each sf south, + 1 point for each 100sf of 5/8 drywall (instead of 1/2"), etc... Those are just numbers from my ear but you get the idea. To meet the code, the building must score a certain number based on local climate.

Anyway, that house did great, even though I ran out of money to complete the mechanicals as I'd planned. It had baseboard and in-floor hydronic heat - actually one of the first lightweight slab on frame infloor systems in the area, the contractor and I learned together. It had 1 thermostat and 1 pump and used tap water straight from the domestic tank water heater.

Not surprisingly with only one pump, even though it was a pretty tight house for the day and the climate mild (4k feet in the sierras doesn't get that cold) the temps were uneven, hot in the sunny areas and cooler in the back of the house. It had flow adjustment for different zones of course but still I needed a rug in front of the toilet to keep my feet from sweating during my morning constitutional.

Here is a snap:

lupin front.jpg

As you can see, it wasn't overkill "Hey look, I'm Solar!" In fact, the idea was for it to look old, Cali-Country. There wasn't all that much glass, those 4, 2-6 x 5-0 (?) windows on the right were the main "collectors", they were in the dining room/kitchen which was all tile over lightweight conc. over wood frame.

I won't be building that house again (on the left it wandered off on a 45º, the porch curved, there were 3 dormers, all vaulted ceilings upstairs, blah blah) it was quite the one man project! But the thing I learned is even a modest amount of exposure can produce really uneven heating.


But I might be biting off too much. I'm more of a Jobs than a Woz if you get me. If I'm building a house I'm not going to be able to learn how to solder IC on breadboards at the same time... never say never, it might happen down the road. I'm sorta thinking finding a wireless (or USB) device that can do I/O with a PC and whatever sensors and controls.

The relay board above is perfect but how do I get the signal from the old 2001 Dell to it without a soldering iron? Same with sensor inputs?

[autonomous]

But I might be biting off too much. I'm more of a Jobs than a Woz if you get me. If I'm building a house I'm not going to be able to learn how to solder IC on breadboards at the same time... never say never, it might happen down the road. I'm sorta thinking finding a wireless (or USB) device that can do I/O with a PC and whatever sensors and controls.

The relay board above is perfect but how do I get the signal from the old 2001 Dell to it without a soldering iron? Same with sensor inputs?

Pops, for a little more money you could connect relays and sensor devices to your PC with USB. You could hack a simple control and data acquisition process and get it to work without touching a soldering iron. These USB devices come with code examples:

USB 8 channel relay board:
http://www.sainsmart.com/sainsmart-8-channel-12-v-usb-relay-board-module-controller-for-automation-robotics.html?___store=en&___store=en

USB temperature sensor:
http://www.ebay.com/itm/PC-Laptop-USB-Thermometer-Temperature-Sensor-C893-/150476229072

There are a few open-source industrial control and data acquisition frameworks based on SCADA that would be worth looking into:
http://mango.serotoninsoftware.com/
http://oscada.org/

http://en.wikipedia.org/wiki/SCADA

[Pops]

You are the man autonomous (or woman, as the case may be)!

[culicomorpha]

Nice place. I think one of the difficulties of this sort of thing is that it's really not very easy to do what-if kinds of experiments. Once stuff gets built, the geometry, dimensions, materials, and a whole slew of minor design decisions all combine in ways that are hard to translate into general rules.

There is a good article on hydronic heating systems in the latest Home Power magazine and they have a lot of good tips/rules-of-thumb for system design. They also show a few circulator pumps and one of them, produced by Grundfos (us.grundfos.com/products/find-product/alpha2.html), has some sort of automatic system detection capabilities. Apparently it adjusts flow rate based on the characteristics of the system so as to minimize power consumption. One of the other things mentioned in the article is that newer hydronic systems are designed to work at lower inlet temperatures (<120 degrees F) and this greatly increases their efficiency compared to systems that utilize high temperature sources. It probably also helps reduce hot-spots.

Regarding a controller, there are different possible levels of implementations, from on-the-cheap surplus parts to professional sensors and data collection. The former requires more time and knowledge, the latter requires more money. I think what autonomous suggested is fine if one has some hacking skills and you don't need electrical inspections where you live. If you do require an inspection, you will have to buy UL listed parts/assemblies only, and this generally means that hobbyist electronics are not going to cut it. There are some exceptions for low-voltage electronics, but no inspector is going to pass a homebrew thermostatic system without some serious persuasion. It is certainly possible to design a proper system meeting all the UL and NEC requirements, but this requires knowledge of all sorts of arcane rules regarding a whole slew of electrical codes.

I don't mean to be discouraging, but it's pretty well-known in the home RE power circuit that if you're going to do a home-brew system, inspectors aren't going to like it because they've never seen it. What commonly happens is that the rules are rigidly interpreted and the costs of remediation often erase the cost benefits of homebrew. Even in Home Power, there is example after example of funny-looking electrical panels where the authors describe an additional (unplanned) requirement from the inspector for a disconnect box, or circuit breaker box, whatever. Of course, if your system was all planned in advance and you got buy-in from an inspector, then I'd say go for it.

[autonomous]

Home Power Mag is a great resource. They suggest using a free tool for modeling the hydronic system before construction to determine sizing of collector and storage:

The RETScreen Clean Energy Project Analysis Software (usually shortened to RETScreen) is a free software package used to determine the feasibility of clean energy projects, which includes renewable energy installations and the means to assess a wide range of energy efficiency options. The software provides the user with a broad range of options for assessing the technical, financial and environmental suitability for an investment in a 'clean energy' project, which includes energy efficiency, renewable energy, and cogeneration (combined heat and power). It integrates a number of databases to assist the site assessor, including a global database of climatic conditions obtained from 4,700 ground-based stations and NASA's satellite data.

http://en.wikipedia.org/wiki/RETScreen
http://www.retscreen.net/ang/home.php

The Department of Energy also has a modeling system called EnergyPlus:
http://apps1.eere.energy.gov/buildings/energyplus/

Modeling Radiant Systems in an Integrated Heat Balance Based Energy Simulation Program:
http://gundog.lbl.gov/dirpubs/ash01_2002.pdf

[culicomorpha]

I've got a couple of Raspberry Pis and they're great little machines. I tried one as a desktop replacement but it's better as a media centre, email server etc. I'm looking at using one for rotating photovoltaic panels to follow the sun to improve overall efficiency and another for controlling watering systems. Without using the onboard graphics they use about one watt-hour of energy, which is phenomenally low. The market for GPIO controllers is obviously immature at the moment but I expect to see great things soon.

OT, but I was wondering if the Raspberry has enough RAM to run a web browser? I like the fact there's no fan and no moving parts, hence a silent computer. I have a Dell netbook now that is also silent, but it uses 20W typically, and I'd like to get that down to below 5W. It seems this is very close to what I'm looking for, but it's not clear if it has enough resources to enable web browsing.

A few weeks ago I actually threw together a simple automatic watering system using an Arduino. The design is a modified version of the one posted at http://www.botanicalls.com/archived_kits/twitter/. It has a homebrew soil moisture sensor and a simple common base amplifier that goes to an analog input. When the moisture level drops below a threshold, a relay turns on a windshield washer pump hooked up to some 1/4" tubing and mister nozzles. It works amazingly well, although the pump is a power hog. My next idea is to use a water reservoir and a solenoid valve and let gravity do the pumping. The Arduino might actually be a better choice in this application due to the onboard A/Ds; it looks like an additional multiplexed A/D using I2C or SPI would be required for the Raspberry PI.

[autonomous]

Creating a model of the hydronic heating system is unavoidable. Even if you setup a sensor array with control algorithms, you will still wind up creating a model through reverse-engineering. To really nail it down it would be best to develop a full simulation beforehand.

Experimental Validation of the EnergyPlus Low-Temperature Radiant Simulation
http://simulationresearch.lbl.gov/dirpubs/fishe03.pdf
http://apps1.eere.energy.gov/buildings/energyplus/pdfs/tips_and_tricks_using_energyplus.pdf