The Garden Digital Twin - Part 1!

Intro

Spring is on the horizon! I'm gearing up for the planting season in my garden and this time I'm determined to give my fruits and vegetables the extra care they deserve!

But I've come to realize that watering all the garden beds is incredibly time-consuming, especially with the responsibilities of being a young dad running his own business. It often gets pushed down my priority list, resulting in sporadic watering patterns where I sometimes go days without watering and then end up flooding the garden beds to compensate, giving the same amount of water to all plants regardless of their needs. The frustration of not seeing any harvest is palpable. This year, I need to approach things differently!

Interestingly, at work, I'm constantly occupied with developing Digital Twins to optimize and automate client machines and processes, ranging from automotive test benches to wind turbines and even pancake dosing lines. So why not apply the same concept to my garden? It sounds like a solid plan!

Let's create a Digital Twin of My Garden!

Let's sketch things out!

To begin with, I studied the available data on my water supply system: the pressure of water entering my garden for irrigation is usually (2 - 3 bar) and a bucket of 10 liters is filled in an average of 30 seconds. The inner surface roughness of rubber hoses is around 0.025 mm.

The next step was to determine the future location of hoses and watering points. I took measurements of the garden and created a plan for the arrangement of the beds, giving insights in the required amount of water piping!

Time for system engineering!

Step 1: Start simple, learn fast!

As a first step I decided to validate the hose model. I have a 45-meter-long hose with a 1/2" diameter and a 10-liter bucket for watering flowers. According to simulation, I have a flow rate of 19.7 l/min, which almost corresponds to my "filling the bucket" experiment! ?Ready to continue!

??Step 2: Let’s build the entire watering system!.

After validating the hose model, I started modeling my garden irrigation system as a whole!

A closer look to the different bags of fruit and vegetable seeds gave me insights in what I was doing wrong... The water consumtion per bed needs to be completely different... summing up for a total of 362 liters!

Garlic - 16 L // Rosemary - 18 L // Lettuce - 25 L // Beans - 21 L

Pepper - 30 L // Strawberry - 60 L // Leek - 21 L // Tomato - 40 L

Strawberry - 60 L // Dill - 21 L // Cucumber - 50 L

∑ = 362?L

In my local garden store, I found hoses with an inner diameter of 1/2" and 1/4". For my first version I decided to stick to a 1/2" diameter hose and 1/2" sprinklers.

??Step 2: Let’s simulate!

As a result of calculations, I found out that such a system design entails several problems:

• Too high consumption: in 5 minutes, 486?L of water will flow to all the beds in total
• The first garden bed (with garlic) will receive 281?L, and the furthest one with cucumbers will receive only 1.3?L, which is too far from the required values and can damage plants.

Not there yet... see you next time!

In the next article I'll show you how we found a solution by combining simulation with optimization! All the best till then! Enjoy spring in the meanwhile!