This project aims to create a web-based management system for the University of Surrey's new student run garden. It will be based around a wireless sensor network connecting distributed moisture sensors in each planting bed to a central gateway via Xbee. Water is supplied via water butts fed by a solar powered pump. The sensor data will be published via MQTT to the IBM Bluemix cloud platform where a Node-RED app will process it and use the information to control solenoid valves that apply water to each bed individually.
The garden layout will be managed via a web interface which allows users to configure the watering regime of different plants and assign them to the beds they have been planted in. It will also allow the user to look at trends and historic data form the various sensors.
Upgrades to the system includes motorized vent control for the garden's polytunnel and thermostatic heating for the winter months.
Components
1×
Solar Panels
To produce sustainable electricity
1×
'Leisure' Battery
Long cycle version of a car battery
For multiple on the same board we will have to multiplex the I2C connection, so we have order two and will test this with the ESP nodes when they arrive. The backpack PCB will contain the power regulator (the distributed power is 24V to get around the garden with minimal loss), the multiplexer and the FET drivers for the water valves.
We have now been donated a 240W solar panel from the Advanced Technology Institute at the University of Surrey, who also do a lot of research into organic and transparent solar cells, so stay tuned for future work with them! This panel will be installed in the next few weeks onto our shed roof, and we will then connect it to an MPPT charger and leisure battery. The controller/charger we are looking at allows us to monitor the voltage and current on the cell and so this will be published to a webpage which we will link here as soon as it's up!
Another job over the summer will be to install the water butt, guttering and pump system into the end of the polytunnel. This will provide us with a natural source of water for the irrigation system, and top it up when needed from the mains supply. Our latest thoughts for a pump is to use a 24V yacht deck-cleaning pump. This is both small and does not require an inverter, which is expensive for motors due to the high inrush current on starting.
Most simple automatic watering systems either use no sensing or just a rain sensor to control to the hose valve. While this will suit most simple cases, our garden has two requirements which need a more complicated system!
Manual watering. The garden society members must be able to manually water the garden if required. One of the benefits of the student garden is that it can be a mental health asset, and watering can be a tranquil and calming task so it is important to still allow students to pick up a watering can or hose and tend to the plants. If the automatic system did not detect this then there is a good chance that the plants could become waterlogged and damaged. Because of this, we will need soil moisture sensors. These sensors are the ideal solution anyway, as they more directly measure the water present at the plant root.
Technology expectations. This is an 'Internet of Things' garden and so must meet expectations of being pretty wow in the tech department! Also, as we hope the project will be handy for others, we want to try and implement a range of sensors which may be more useful in other gardens or crops. Therefore, we will try and incorporate thermometers, soil moisture sensors, wind sensors, rainfall sensors, light sensors, humidity sensors and probably CO2 sensors for the polytunnel.
Stay tuned for more information on our choice of soil moisture sensor coming soon!
Just a little progress update! The garden is going well and a good deal has been grown over this summer. We're really feeling the stress now as the hot weather is here and most of the students have gone home so it's up to a dedicated watering team to keep the plants wet over this period - a problem which the irrigation system will solve! However, another source of sponsorship is taking a while to work with, so in the mean time I have been building an alternative shopping list using the cheapest parts possible. This of course has the added benefit of making it accessible to hackers and makers ;)
One of the most expensive parts of an automatic irrigation system is the valves. These have not had their price driven down like MCUs and wifi modules, so are still pretty pricey. For a typical irrigation solenoid valve the cost is £20 - £40, which is a lot of money considering we need about 21 of them! The lowest cost solution I have found is instead to use cheap washing machine solenoid valves, typically priced at under ten pounds each. However, six of these in each box, which is the current plan, needs quite a large box to space them out with fittings. A neater solution using the same technology is to use washing machine triple valves. These cost about the same as three of the single cheap valves and provide three outlets for each single inlet, saving a huge amount of space! This means I only need two of these in each box.
The valves are available in both mains (240V) and 12VDC varieties, the latter of which will be supplied to each box for the controller, sensors and valves.
We just got approved a £1k funding application from the university annual alumni fund! This means we can start to purchase the weatherproof enclosures and power cabling for the project.
We are still awaiting on some sponsorship opportunities for hosing and fittings so we won't purchase any of those parts yet.
Stay tunes for many more details in the next few days.