@dantman16: IoT-Enabled Planter System For Urban Space

IoT-Enabled Planter System For Urban Space
By

Daniel Lee

Professor, College of IT Convergence

INTRODUCTION

Despite our best effort to be active and spend more time outdoors, a majority of our time is spent indoors, whether it’s a home or working in an office environment. The modern building design favors energy-efficient climate-control resulting in reduced fresh air exchange. Indoor air pollution is a health risk either as a concentrated trace chemical or psychological ‘sick building syndrome’.

In addition to indoor air pollution, inhalable particles (micro dust) pollution is an even bigger issue in Korea. The fine particles of 10 and 2.5 micrometers blankets the sky during the winter months.

In 1989, NASA has conducted a Clean Air Study to find the most effective common indoor plants for removing toxins and pollution in the air. Here we present NASA’s recommended common indoor plants in conjunction with my inventions, microsiphon, Veloster, s-trap, and snorkel used in indoor aquaponics planter.

It has successfully gone through alpha and best test stage, and currently being sold at amazon.com, and etsy.com under ‘justponics’ brand. The 3D printing lean manufacturing is performed by lefab3d.com, a member company of PeopleSpace incubator.

MISSION STATEMENT

Through these improvements, we wish to make immediate impacts on people, planet, and profit by offering;

1) small-scale food production

2) tabletop garden and pond, and

3) plant-based indoor air purifier

INCORPORATION

Justponics LLC was formed on 7/29/2019 as a California limited liability corporation

HOW IT WORKS

INDOOR FOOD PRODUCTION

AIR PURIFICATION

NASA study suggested that in addition to absorbing carbon dioxide and releasing oxygen through photosynthesis certain common indoor plants may also provide a natural way of removing toxic agents from the air.

English ivy (Hedera helix)

Green Spider plant (Chlorophytum elatum)

Peace lily (Spathiphyllum 'Mauna Loa')

Chinese evergreen (Aglaonema modestum)

Bamboo palm (Chamaedorea seifrizii)

Variegated snake plant, mother-in-law's tongue (Sansevieria trifasciata 'Laurentii')

Heartleaf philodendron (Philodendron cordatum)

Selloum philodendron

(Philodendron bipinnatifidum)

Elephant ear philodendron (Philodendron domesticum)

Red-edged dracaena (Dracaena marginata)

Cornstalk dracaena (Dracaena fragrans 'Massangeana')

Weeping fig (Ficus benjamina)[5]

Barberton daisy (Gerbera jamesonii)

Florist's chrysanthemum (Chrysanthemum morifolium)

Aloe vera (Aloe vera)

Janet Craig (Dracaena deremensis "Janet Craig")

Warneckei (Dracaena deremensis "Warneckei")

Banana (Musa oriana)

We conducted air quality sampling with a Laser Particle Counter which can detect particles down to 0.5 microns. The large particle size range is calibrated to 2.5 microns and above. It has a PC Interface to capture the air quality data to a PC for graphing and analysis. A room with our Edu Kit was monitored overnight for the PM2.5 level. The result was full of potential (a.k.a. non-conclusive) The room’s PM2.5 level settled down after 6 hours. Due to the PM2.5’s Brownian mobility, the particles might have landed on the leaves of the plants. There is no evidence PM2.5 has true mobility of its own. Electrostatic or other forms of attraction (including gravity) might be at play. We were not able to measure plants’ gas exchangeability with our limited instruments at this time. It would seem plausible a plant with the most leaf surface area should capture the most PM2.5 out of the air. The plant should have assisted or self-cleaning ability to cleanse itself of PM2.5 attracted on its leaves. This needs further investigation.

Air quality monitor used for instant spot monitoring

DECORATION & COMPANION PLANT

Aside from small scale food production and air purifying qualities, Edu kit serves as a beautiful decorative companion plant planter.

RESEARCH

Since we were devising an aquaponics device, we looked into prior work done on hydrodynamics wave propagation but came to learn that the problem was a complicated one, and had not yet been rigorously solved analytically. Instead, a series of assumptions and simplifications were often used in practice. The fluid wave propagation was a three-dimensional problem, but it was simplified from three-dimensional to a one-dimensional problem by using Webster’s horn equation. We got useful information about the behavior with this type of approach. A large number of cases were solved this way with high satisfaction. We got a really lucky break when we made the further assumption that water is a type of high viscous air. It is indeed a lucky break for me because the math on aerodynamics was abundantly available.

(edu kit v2.0 in action. Self-contained eco system)

The following items are invented and manufactured by us.
MICROSIPHON

(This microsiphon is optimized for 1 to 10-gallon tank with an operating flow rate of 10 to 40 gallons per hour )
VELOSTER

Bell siphon Veloster is a specially shaped hydrodynamics device designed to facilitate failsafe starting of siphon by passing a large volume of water into a narrow riser pipe. Without the Veloster, water will attempt to drain simultaneously creating a traffic jam. The science here is a reversal of impedance resonance. The impedance resonance is why a trumpet can carry a tiny lip vibration at the trumpet mouthpiece over a great distance. The boundary layer development was critical in achieving a strong siphon because the high viscosity of water produces high shear stresses, and distorts the inner non-viscous flow. The device achieves compressed entry of water by transitioning slow water to fast water flow. Entering water is smoothed out, and the flow fills the entire pipe area.

Red: slow water

Blue: fast water
S-TRAP

S-trap is a curved hydrodynamics device designed to facilitate minimizing siphon as long as possible so that the maximum siphon is achieved through a sudden jump in the water velocity. It operates fundamentally different from the 90-degree elbow. The scientific method here is Stick and Slip phenomenon. Stick-slip can be described as surfaces alternating between sticking to each other and sliding over each other, with a corresponding change in the force of friction. If an applied force is large enough to overcome the static friction, then the reduction of the friction to the kinetic friction can cause a sudden jump in the velocity of the movement. You can see the water column inside a bell rise in stick and slip motion fashion.

(S-trap showing different water height of inside and outside the bell due to backpressure)

(blueprint drawing of microsiphon)

SNORKEL

A snorkel is a special internal snorkel piping designed to facilitate failsafe stopping of siphon. It passes a large volume of air into a narrow riser pipe without interruption. The science here is Stoke’s Law. A bubble starts out as compressed gas equal to the surrounding fluid pressure when it was first formed. As a bubble rises it loses initial acceleration and reaches terminal velocity while increasing in size causing more drag. This means a bubble at terminal velocity does not have enough kinetic energy to effectively stop the siphon. We want a fast bubble to reach the top quickly so the siphon can stop quickly. The change in the velocity and direction of movement are kept as gradual as possible to attain the highest economy by applying the principle of laminar flow. The specific calculation of the Reynolds number is satisfied by manipulating the geometry of the flow system and flow pattern. The bubble is given the straightest path of least resistance by a snorkel.

DRONE

Used in IoT enabled smart farming. This drone detects and controls pests. We designed two pieces of shells made out of carbon fiber 3D printed shells. The shells are lightweight to increase the payload and flight time of the drone. The drone will carry an AI-based vision system to detect and identify microscopic pests. The future wishlist is to include laser bug zapper for surgical elimination of the pests.

IoT

We developed our own custom board to monitor and collect data from the various custom soil, water, and air sensors.

MOBILE APP

LITERS PER HOUR FORMULA

We are proud to say all bell siphons are fully tested and calibrated before leaving our site. Perfect! But some fans write to us the bell siphon does not work upon receiving. How could this happen? It’s because the bell siphon is used in all kinds of different environments so the water flow rate, measured in LPH, liters per hour, is incorrectly calculated.

Note that calibration is an ongoing process that needs to be performed throughout the life of the growth kit. Without calibration, the bell siphon may not work as designed. If you want better bell siphon experience, we recommend maintaining the optimum water flow rate.

There is a definite science behind it, but we will keep things as plain as possible. Bell siphon itself is a quite simple contraption as far as mechanics go in the sense that there are no moving parts besides water level rising and falling.

The flow rate of a siphon is a function between the diameter of standpipe and S-trap. The equation is as described by Bradbury (1910) gives the average flow rate as

Q = K x A * ((2 x g x h)^(1/2))

where Q= flow rate (L min−1);

K= discharge coefficient;

A= inner area of the discharge pipe (m^2);

h= average head [= average of the maximum and minimum head (mm)], and

g= acceleration due to gravity (mm min−2).

Water inflow is achieved by a 300LPH (80GPH) submerged pump which can be purchased for ~$10.

Because the inflow rate is depended on the capacity of the pump and the pumping height, we determined 300LPH (80GPH) rated pump was achieving 150LPH (40GPH) actual pumping rate when the pumping height was at 280 mm. 280 mm was calculated by assuming stacking two 140 mm height bins. We kept the pump at the lowest part of the lower bin and water outlet hose at the highest part of the upper bin.

MICROSIPHON CALIBRATION

With a perfectly calibrated bell siphon, you should see a constant water level rising and falling. The water may start trickling down the tube, but water inflow should overtake the trickling loss. Then you should see rapid dumping of water to the reservoir signaling of successful siphoning. Once the water level is sufficiently lowered, the siphon should stop with a strong burp.

(nice strong siphon. the water column is way above the Veloster)

There are two main problems a bell siphon can have. The first is where the water flow rate is too low so the siphon does not start, and the second is where the water flow rate is too high so the siphon does not stop.

SIPHON DOES NOT START

1. Trickling sound

Solution: increase the flow rate

The relationship between the pump’s flow rate and the pumping height was a second-degree polynomial. In our calculation, a minimum of 40LPH (10GPH) was required for a successful siphon start. The siphon did not start if the pump inflow could not overtake the inevitable slow trickling down of water. We called this stage a “perma-trickle”.

(water trickles down so siphon does not start)

The grow tank is filled with water so it seems to be ready to start siphoning, but it doesn’t because water is slowly leaking into the reservoir below. The water-in is less than water-out. The drain pipe sucks in the air instead of creating a vacuum to prime the pipe. You can prime the siphon in two ways. Increase the water pump rate or increase the pressure resistance in the drain pipe. The water pump has an adjustable valve. Rotate towards + to increase the flow rate. The water pump needs regular cleanup to remove sludge build-up.

SIPHON DOES NOT STOP

1. Gurgling sound

Solution: decrease the flow rate

The grow tank is drained, but it is not refilling back up. Water-out is greater than water-in. You can break the siphon in two ways. Decrease the water pump rate or decrease the pressure resistance in the drain pipe. The water pump has an adjustable valve. Rotate towards – to decrease the flow rate.

At an inflow rate above 300LPH (80GPH), the siphon did not stop reliably at the end of the discharge. At this high rate, the water inflow rate exceeded the minimal speed which air could enter the bell to break the siphon. We called this stage a “perma-drain”.

We were quite surprised, at least for us, how easy it was to achieve a perma-drain stage. We were achieving this without any calculation. We will revisit why the margin of error was so large to achieve an equilibrium when by definition a rather precise point where inflow and outflow are equal must be achieved.

2. The water pump is getting weaker

The water pump, over time, whether through age, contamination, algae build-up, etc, will slow down. This will result in decreased inflow rate so the siphon does not start properly. You should clean or replace the pump.

3. Overgrowth of roots

Overgrown vegetation root blockage will restrict the water outflow so the siphon does not stop. Clear away the blockage or replace the siphon.

BIN

The container used was SAMLA 1 gallon plastic bin found at Ikea.com.

Width: 280 mm

Depth: 190 mm

Height: 140 mm

Volume: 3.8 liter

Material: Polypropylene (C3H6)n

We designed the Edu Kit for indoor use only so we believe beneficial physical properties of PP such as good chemical resistance with good elasticity and good fatigue resistance outweigh PP’s susceptibility to UV degradation. The replacement bin can be purchased economically from Ikea.com. In any case, please do not use the Edu Kit outdoors or under direct sunlight.

Since it would have been more optimum if the pump was cut off during the discharge, we considered a float valve switch design to cut off the pump when the siphon started. This design was put on hold as a float valve switch would have taken up valuable grow bed space.

OPTIMUM PUMP PERFORMANCE

In our calculation, 150LPH (40GPH) was most reasonable for 3.8 liters (1 Gallon) IKEA samala bin at 280 mm pumping height. The siphon started surely and ended satisfyingly without fail and worry, every single time. We called this a “one-siphon-to-rule-them-all” stage.

As we have calculated and demonstrated, the success of siphon depends largely on correct calibration based on the tank’s volume (width x depth x height, 3.8 liters in our case), the surface area (width x depth), water flow rate (water inflow – water outflow). There was a strong linear relationship between the water flow rate and trigger/break action with a coefficient of determination of 0.99.

One more important dimension of consideration was how fast water rises. This is a derivative of the net water flow rate and bin dimension. For a given net water flow rate and a bin volume, a narrow bin would achieve faster water rise than a wide bin. A faster water rise from a narrow bin will aid in the starting of the siphon. A slower water rise from a wide bin may not be able to start the siphon. A calculation based on actual water inflow rate alone would only be correct if we assume the grow bed is always used empty. The water rise rate depended on how densely medium (sand, pebble, hydroton, etc) was packed in addition to actual water inflow rate which depended on actual pump rate and pump height and in addition to how wide or narrow the bin was shaped.

MACHINE LEARNING

We developed an IoT device to collect data and record pertinent information in managing the edu kit. We anticipate our system will evolve to full integration of automated data recording, data analysis, machine learning, and decision-making into an interconnected system.

ML uses a set of well-defined models that collect specific data and apply specific algorithms to achieve expected results. The recent development of deep learning has expanded the scope of Artificial Neural Network application in agriculture.

Neural Networks are inspired by the human brain functionality and represent a simplified model of the structure of the biological neural network emulating complex functions such as pattern generation, cognition, learning, and decision making. Such models are used for regression and classification tasks

SVMs are binary classifiers that construct a linear separating hyperplane to classify data instances. SVMs are used to predict the yield and quality of crops by classification, regression, and clustering.

Soil Management

Machine learning algorithms study evaporation processes, soil moisture, and temperature. We are interested in detecting balanced fertilizer properties in soil.

Water Management

Machine learning algorithms are used for effective irrigation systems by estimating evapotranspiration and evaporation. We are interested in detecting and maintaining ph level of 6 in water using auto doser.

Yield Prediction

ML can make accurate detection and classification of crop quality characteristics. This is tied to the yield prediction forecast resulting in precision agriculture.

Weeds & Insects Detection

Weed control requirement for indoor planter system is minimum. The spider mite is a huge problem in tomato farming. Computer vision and ML algorithms and robots improve detection and precise discrimination of weeds and insects minimizing the need for herbicides and insecticides.

3D PRINTER FARM

We design and produce our own parts in-house.