Since today I’ve been using only time based irrigation dosing. That’s perfectly fine if I just want to feed ”some amount” of water for the plants. Things get more complicated if I want to dose exactly e.g. 50 ml or some other exact amount of water. In this case following field and installation related nonlinearties affect the irrigation dosing:

  • Water pump power
    • Pump power might change due to load changes
  • Pipe diameter
    • Narrow pipe needs less pump power
    • Many sizes of pipes in a row complicates things even more
  • Pipe length
    • Longer pipe needs obviously more water. (this is linear)
  • Irrigation height from pump to plant
    • Gravity has an big effect for the throughput.
  • Water ”kickback” effect in a pipe
    • Paddle wheel of the pump leaks and allows water to flow back when the pump is off. Paddle wheel leaks also during the active mode and especially with high irrigation heights
  • Unexpected air in the pipe
    • If the pipe is pended many times or if it has many up-down routes, unwanted air can lead to the pipe causing error on dosing accuracy

These all factors affects dramatically for the dosing and if not considered during the design, desert drought and very unhappy plants might be the result. I’ve tried to figure out how to deal all these factors and I have some preliminary ideas, but they are not yet ready for publishing. They need little bit more time and cooking. Here are some preliminary ideas:

  1. No pipe calibration at all (time based feeding)
  2. Irrigation with calibrations that uses some correlation factors for the points I listed above
  3. Irrigation with auto-calibration. In this option pumped water is measured and used for calibration factors.

In addition to field/installation related nonlinearties, temperature and humidity needs to be taken in account as well. For that, I’ve already installed sensors and created preliminary algorithms. Fine tuning will be done in the summer time.