Paper Title : A TDS Prediction System For Hydroponic Vertical Farm Based On ARIMA

Model

Presented by:Gaganjot Kaur

Full Affiliation: PhD scholar and

Assistant Professor
Chandigarh University, Mohali

Presentation Date:3 -12-2022

Table of Content

 Introduction
 Experimental Setup
 ARIMA model
 Results and Discussion
 Conclusion
 Introduction
● The continuous increase in the world population has great impact on rising the food
demand.
● The global population is expected to grow to around 9.7 billion in 2050, and 10.9
billion in 2100 when compared with the projected 7.7 billion people worldwide in
2019.
● Also, is an alarming situation for India as the India’s population is expected to reach
1.7 billion which will make India the most populated country
● This tremendous increase in the population has created the biggest challenge in front
of the country to achieve Sustainable Development Goal (SDG) two i.e., SDG-2,
which talks about zero hunger. Also, as per the report, food production needs to be
doubled by 2050 in order to meet the demand of the growing population
● Agriculture resources such as clean water and arable land are declining [4, 5]. The
variation in the climate conditions and reduction in bio-diversity are other major
concerns in the agriculture field. Further, the lack of rainwater often results in a
declined supply of groundwater, lakes, or rivers, leading to a decrease in groundwater
levels and the scarcity of natural water resources.
Introduction

● Therefore, to sustainably feed the growing population modern methods of

farming such as soil-less farming techniques viz. hydroponics, aeroponics,
aquaponics, and vertical farming are the latest innovation techniques that may
provide some cure to the food supply demand, especially in the urban areas
helping to deal with such alarming situation.
● In general, soil-less farming involves growing plants in nutrient-rich solution
without the use of soil preferably in an indoor environment whereas vertical
farming is all about stacking the various plant layers in order to increase the
yield per unit area.
● As these systems do not need soil to be effective, they require far less space and
water in comparison to conventional agriculture.
Experimental Setup
• The experimental system prototype consists
of a 3 tier vertical farming setup with 3 rows
on each tier having capacity of 9*3= 27
plants with total of 81 plants for complete
setup as shown in Fig.

• The complete dimensions of system are 4 ft

x 1.5 ft x 5 ft. The system is based on Deep
Flow technique which is a combination of
NFT and DWC hydroponic techniques.

• LED grow lights are used for providing light

for photosynthesis. This system is integrated
with various sensors like pH, TDS and DHT 11
sensor and data is collected for a duration of
plant growth i.e. 30 days on ThingSpeak
cloud.
Recorded TDS and additive Seasonal
Decompose
(a) Data collected from
Thingspeak for one-
month
(b) trend followed by
data
(c) seasonality of time
series
(d)remainder obtained
by subtracting
trend and seasonal
component from
original series
Seasonal
decompose graphs
for TDS parameter
 ARIMA MODEL
 The ARIMA model that uses time series data is used to predict future trends on
collected data.
 (ARIMA) model refers to “Auto Regressive Integrated Moving” average model.
This model defines a given time series on the basis of its own previous values,
its lags and forecast [Link] model type is identified by three terms p, d, q,
where p denotes the order of autoregressive term, q refers to order of moving
average terms and d denotes the order of differencing needed in order to
make the time series stationary. p, d, q are all non-negative integers.
 The plain ARIMA model does not support seasonality, So in an extension to the
non-seasonal model, the Seasonal-ARIMA is used if the time series has
seasonality. It is represented as SARIMA(p, d, q)(P, D, Q) M, where P and Q
refers to SAR and SMA terms respectively and D is order of seasonal
differencing. M is the number of periods per season.
Autocorrelation and Partial Autocorrelation
Plots
RESULTS
AND
DISCUSSSION
 SARIMA

 In order to check stationarity of series, Augmented Dickey Fuller (ADF)

test is applied and value came out to be 0.683. As the value is more
than 0.05 so null hypothesis is rejected and series is concluded to be
non-stationary
 Trying to find out the best model we applied auto SARIMA with d=None
and D=1. Best model came out to be (1,1,2)(0,1,1)_{12} with the
minimum AIC value of 967.286. Other models (0,1,2)(0,1,1)_{12},(1,1,2)
(1,1,1)_{12} and (0,1,2)(0,1,2)_{12} are also chosen along with
manually found (2,1,1)(0,1,1)_{12} model to compete with (1,1,2)
(0,1,1)_{12} model in terms of AIC values and RMSE score.
Parameter estimate for SARIMAX
models
Forecasting vs Test Data for different
models
Test , Train and Predicted Data
Conclusion and Future Scope

 In this paper, TDS forecasting is reported using a time series ARIMA model for new-age agriculture
technique i.e., hydroponic vertical farming, in which proper maintenance of nutrient solutions (i.e.,
TDS and pH values) are required to maintain healthy plant growth. Therefore, to include fail-safe, it's
better to predict the parameters beforehand from the data collected over the ThingSpeak cloud for
Hydroponic Vertical farms.
 Since the data shows trend and seasonality, the SARIMA model is applied for prediction and the
best model with the lowest AIC was obtained at (1,1,2)(0,1,1)_{12}. However, the lowest RMSE was
obtained for (0,1,1)(0,1,1)_{12}. Thus, based on RMSE score, coefficient value and P value, overall
(0,1,2)(0,1,1)_{12} is selected as best model. Forecast values are plotted against test values and
RMSE of 18.88 is obtained for (0,1,2)(0,1,1)_{12}.
 For the future generation of agriculture techniques, these prediction algorithm finds great uses in
upcoming automated farming setups. Further, other prediction models like Prophet and Holt's winter,
etc. could be tested and compared with SARIMA in order to figure out the best prediction algorithm
techniques in the field of hydroponic vertical farming which is an upcoming agriculture technique.
 References
 [1] J. Cleland, “World population growth; past, present and future,” Environmental and Resource Economics, vol. 55, no. 4, pp. 543–554, Aug 2013. [Online].
Available: [Link]

[2] M. Rahman, “Exploring the effects of economic growth, population density and international trade on energy consumption and environmental quality in
india,” International Journal of Energy Sector Management, vol. ahead-of-print, 04 2020.

[3] P. G. Pardey, J. M. Beddow, T. M. Hurley, T. K. Beatty, and V. R. Eidman, “A bounds analysis of world food futures: Global agriculturethrough to 2050,”
Australian Journal of Agricultural and ResourceEconomics, vol. 58, no. 4, pp. 571–589, 2014. [Online]. Available:
[Link]

[4] B. R. D ̈o ̈os, “Population growth and loss of arable land,” Global Environ-mental Change, vol. 12, no. 4, pp. 303–311, 2002. [Online].
Available:[Link]

 [5] R. Pr ̆av ̆alie, C. Patriche, P. Borrelli, P. Panagos, B. Ros, ca, M. Dumitras ̧cu, I.-A. Nita, I. S ̆avulescu, M.-V. Birsan, andG. Bandoc, “Arable lands under the
pressure of multipleland degradation processes. a global perspective,” EnvironmentalResearch, vol. 194, p. 110697, 2021. [Online]. Available:
[Link]

[6] S. Caputo, History, Techniques and Technologies of Soil-Less Cultivation. Cham: Springer International Publishing, 2022, pp. 45–86.

[7] M. SharathKumar, E. Heuvelink, and L. F. Marcelis, “Vertical farming:Moving from genetic to environmental modification,” Trends in PlantScience, vol. 25,
no. 8, pp. 724–727, 2020. [Online]. Available:[Link]

[8] C. E. Wong, Z. W. N. Teo, L. Shen, and H. Yu, “Seeing thelights for leafy greens in indoor vertical farming,” Trends in FoodScience Technology, vol. 106, pp.
48–63,2020. [Online]. Available:[Link]

[9] S. Ragaveena, A. Shirly Edward, and U. Surendran, “Smartcontrolled environment agriculture methods: a holistic review,” Reviewsin Environmental Science
and Bio/Technology, vol. 20, no. 4, pp.887–913, Dec 2021. [Online]. Available: [Link]
 References
[1-] H. K. Srinidhi, H. S. Shreenidhi, and G. S. Vishnu, “Smart hydroponics system integrating with iot and machine learning algorithm,” in 2020 International
Conference on Recent Trends on Electronics, Information, Communication Technology (RTEICT), 2020, pp. 261–264.

[11] S. Bhowmick, B. Biswas, M. Biswas, A. Dey, S. Roy, and S. K. Sarkar, “Application of iot-enabled smart agriculture in vertical farming,” in Advances in
Communication, Devices and Networking, R. Bera, S. K.
Sarkar, O. P. Singh, and H. Saikia, Eds. Singapore: Springer Singapore,2019, pp. 521–528.

[12] M. Mehra, S. Saxena, S. Sankaranarayanan, R. J. Tom, and M. Veeramanikandan, “Iot based hydroponics system using deep neural networks,” Computers
and Electronics in Agriculture, vol. 155, pp. 473–486, 2018. [Online]. Available:[Link]

[13] M. S. Verma and S. D. Gawade, “A machine learning approach forprediction system and analysis of nutrients uptake for better cropgrowth in the
hydroponics system,” in 2021 International Conferenceon Artificial Intelligence and Smart Systems (ICAIS), 2021, pp. 150–156.

[14] M. I. Alipio, A. E. M. Dela Cruz, J. D. A. Doria, and R. M. S. Fruto, “On the design of nutrient film technique hydroponics farmfor smart agriculture,”
Engineering in Agriculture, Environment andFood, vol. 12, no. 3, pp. 315–324, 2019. [Online]. Available:https
://[Link]/science/article/pii/S1881836617303294

[15] J. Torres-Tello and S.-B. Ko, “Interpretability of artificial intelligencemodels that use data fusion to predict yield in aeroponics,” Journal ofAmbient Intelligence
and Humanized Computing, Sep 2021. [Online].Available: [Link]

[16] G. Zhang, “Time series forecasting using a hy-brid arima and neural network model,” Neurocomput-ing, vol. 50, pp. 159–175, 2003. [Online]. Available:
[Link]

[17] D. Benvenuto, M. Giovanetti, L. Vassallo, S. Angeletti, and M. Ciccozzi,“Application of the arima model on the covid-2019 epidemic dataset,”Data in Brief,
vol. 29, p. 105340, 2020. [Online]. Available:[Link]

[18] S. C. Hillmer and G. C. Tiao, “An arima-model-based approachto seasonal adjustment,” Journal of the American StatisticalAssociation, vol. 77, no. 377, pp.
63–70, 1982. [Online]. Available:[Link]