An ontology model to support the automated design of aquaponic grow beds

Abstract

Aquaponics is a promising sustainable farming method that combines aquaculture and hydroponics. It allows the growth of crops without soil, pesticides, or fertilizers, and with a minimum amount of water. In aquaponic systems, the design of the growing area is directly linked to the type of crop about to be planted. The type of crop directly determines, for example, the spacing between plants and between channels, which is critical to determine the footprint required and estimate the system productivity. This paper proposes a knowledge modeling approach to support the design of aquaponic systems by automatically determining the required characteristics of the aquaponic system based on crop selection.The knowledge modeling is outlined as an ontology model that formally describes the existent links between the aquaponic grow bed characteristics and its design parameters. This study gives practitioners the capacity to visualize the impact of the desired crop selection on the aquaponic system design, as well as supporting clearer decision-making regarding production facility layout and system design in aquaponic farms.

Full text

ScienceDirect
Available online at www.sciencedirect.com
Available online at www.sciencedirect.com
ScienceDirect
Procedia CIRP 00 (2017) 000–000
www.elsevier.com/locate/procedia
2212-8271 © 2017 The Authors. Published by Elsevier B.V.
Peer-review under responsibility of the scientific committee of the 28th C IRP Design Conference 2018.
28th CIRP Design Conference, May 2018, Nantes, France
A new methodology to analyze the functional and physical architecture of
existing products for an assembly oriented product family identification
Paul Stief *, Jean-Yves Dantan, Alain Etienne, Ali Siadat
École Nationale Supérieure d’Arts et Métiers, Arts et Métiers ParisTech, LCFC EA 4495, 4 Rue Augustin Fresnel, Metz 57078, France
* Corresponding author. Tel.: +33 3 87 37 54 30; E-mail address: paul.stief@ensam.eu
Abstract
In today’s business environment, the trend towards more product variety and customization is unbroken. Due to this development, the need of
agile and reconfigurable production systems emerged to cope with various products and product families. To design and optimize production
systems as well as to choose the optimal product matches, product analysis methods are needed. Indeed, most of the known methods aim to
analyze a product or one product family on the physical level. Different product families, however, may differ largely in terms of the number and
nature of components. This fact impedes an efficient comparison and choice of appropriate product family combinations for the production
system. A new methodology is proposed to analyze existing products in view of their functional and physical architecture. The aim is to cluster
these products in new assembly oriented product families for the optimization of existing assembly lines and the creation of future reconfigurable
assembly systems. Based on Datum Flow Chain, the physical structure of the products is analyzed. Functional subassemblies are identified, and
a functional analysis is performed. Moreover, a hybrid functional and physical architecture graph (HyFPAG) is the output which depicts the
similarity between product families by providing design support to both, production system planners and product designers. An illustrative
example of a nail-clipper is used to explain the proposed methodology. An industrial case study on two product families of steering columns of
thyssenkrupp Presta France is then carried out to give a first industrial evaluation of the proposed approach.
© 2017 The Authors. Published by Elsevier B.V.
Peer-review under responsibility of the scientific committee of the 28th CIRP Design Conference 2018.
Keywords: Assembly; Design method; Family identification
1. Introduction
Due to the fast development in the domain of
communication and an ongoing trend of digitization and
digitalization, manufacturing enterprises are facing important
challenges in today’s market environments: a continuing
tendency towards reduction of product development times and
shortened product lifecycles. In addition, there is an increasing
demand of customization, being at the same time in a global
competition with competitors all over the world. This trend,
which is inducing the development from macro to micro
markets, results in diminished lot sizes due to augmenting
product varieties (high-volume to low-volume production) [1].
To cope with this augmenting variety as well as to be able to
identify possible optimization potentials in the existing
production system, it is important to have a precise knowledge
of the product range and characteristics manufactured and/or
assembled in this system. In this context, the main challenge in
modelling and analysis is now not only to cope with single
products, a limited product range or existing product families,
but also to be able to analyze and to compare products to define
new product families. It can be observed that classical existing
product families are regrouped in function of clients or features.
However, assembly oriented product families are hardly to find.
On the product family level, products differ mainly in two
main characteristics: (i) the number of components and (ii) the
type of components (e.g. mechanical, electrical, electronical).
Classical methodologies considering mainly single products
or solitary, already existing product families analyze the
product structure on a physical level (components level) which
causes difficulties regarding an efficient definition and
comparison of different product families. Addressing this
Procedia CIRP 100 (2021) 55–60
2212-8271 © 2021 The Authors. Published by Elsevier Ltd.
This is an open access article under the CC BY-NC-ND license (https://creativecommons.org/licenses/by-nc-nd/4.0)
Peer-review under responsibility of the scientific committee of the 31st CIRP Design Conference 2021.
10.1016/j.procir.2021.05.009
© 2021 The Authors. Published by Elsevier Ltd.
This is an open access article under the CC BY-NC-ND license (https://creativecommons.org/licenses/by-nc-nd/4.0)
Peer-review under responsibility of the scientic committee of the 31st CIRP Design Conference 2021.
Available online at www.sciencedirect.com
ScienceDirect
Procedia CIRP 00 (2019) 000–000
www.elsevier.com/locate/procedia
2212-8271 © 2021 The Authors. Published by Elsevier Ltd.
This is an open access article under the CC BY-NC-ND license (https://creativecommons.org/licenses/by-nc-nd/4.0)
Peer-review under responsibility of the scientific committee of the 31st CIRP Design Conference 2021
31st CIRP Design Conference 2021 (CIRP Design 2021)
An ontology model to support the automated design of
aquaponic grow beds
Rabiya Abbasia, Pablo Martineza, Rafiq Ahmada,*
aLaboratory of Intelligent Manufacturing, Design and Automation (LIMDA),Department of Mechanical Engineering,
University of Alberta, 9211 116 St., Edmonton, AB, T6G 2G8, Canada
* Corresponding author. Tel.: +1 (780) 492 7180;E-mail address: rafiq.ahmad@ualberta.ca
Abstract
Aquaponics is a promising sustainable farming method that combines aquaculture and hydroponics. It allows the growth of crops without soil,
pesticides, or fertilizers, and with a minimum amount of water. In aquaponic systems, the design of the growing area is directly linked to the type
of crop about to be planted. The type of crop directly determines, for example, the spacing between plants and between channels, which is critical
to determine the footprint required and estimate the system productivity. This paper proposes a knowledge modeling approach to support the
design of aquaponic systems by automatically determining the required characteristics of the aquaponic system based on crop selection. The
knowledge modeling is outlined as an ontology model that formally describes the existent links between the aquaponic grow bed characteristics
and its design parameters. This study gives practitioners the capacity to visualize the impact of the desired crop selection on the aquaponic system
design, as well as supporting clearer decision-making regarding production facility layout and system design in aquaponic farms.
©2021 The Authors. Published by Elsevier Ltd.
This is an open access article under the CC BY-NC-ND license (https://creativecommons.org/licenses/by-nc-nd/4.0)
Peer-review under responsibility of the scientific committee of the 31st CIRP Design Conference 2021
Keywords: knowledge modeling; aquaponics; precision farming; parametric design; design automation
1. Introduction
Food security and sustainability have become a major
concern over the years due to substantial urbanization,
destruction of the ecological environment, farmlands scarcity,
and increasing growth of the human population [1]. Traditional
agriculture methods employed for crop production require vast
amounts of land, time, and manpower and hence are not very
efficient to meet the current food demands. Consequently, the
current paradigm poses a need to explore new farming methods.
Aquaponics, a subset of integrated agri-aquaculture systems,is
expected to address these problems due to its ability to develop
and achieve economically viable and environmentally
sustainable food production practices [2]. The rationale of this
soilless recirculating growing system involves sharing the
mutual benefit of the available resources,such as water and
nutrients, between aquaculture and plant production.
An aquaponic system is comprised of two integrated units:
1) a hydroponic unit which consists of grow beds for plant
growth; and 2) an aquaculture unit that involves water tanks for
fish habitat and biofilters for the breakdown of ammonia [3].
These units work together in a symbiotic environment to enable
plant and fish growth. Primarily, depending on the structure of
the plants’ grow bed and crop type and size, there are three
different types of aquaponics system designs: nutrient film
technique (NFT), media bed, and deep water culture (DWC)
[4].In this paper, the NFT based aquaponic system is
considered because it is the most popular type of aquaponic
setup used. Moreover, it uses less water and is suitable to grow
leafy green crops. In NFT systems, a very thin film of nutrient-
rich water is pumped to enclosed channels. The top cover of the
channel consists of circular or squared shaped pockets known
as plant sites where plants sit in small plastic cups allowing their
roots to access the water and absorb the nutrients [5].
The design and management of an NFT-based indoor
aquaponic system present several challenges when scaling it to
a commercial level [6].These challenges are mainly
attributable to the design of grow channels based on crop
Available online at www.sciencedirect.com
ScienceDirect
Procedia CIRP 00 (2019) 000–000
www.elsevier.com/locate/procedia
2212-8271 © 2021 The Authors. Published by Elsevier Ltd.
This is an open access article under the CC BY-NC-ND license (https://creativecommons.org/licenses/by-nc-nd/4.0)
Peer-review under responsibility of the scientific committee of the 31st CIRP Design Conference 2021
31st CIRP Design Conference 2021 (CIRP Design 2021)
An ontology model to support the automated design of
aquaponic grow beds
Rabiya Abbasia, Pablo Martineza, Rafiq Ahmada,*
aLaboratory of Intelligent Manufacturing, Design and Automation (LIMDA),Department of Mechanical Engineering,
University of Alberta, 9211 116 St., Edmonton, AB, T6G 2G8, Canada
* Corresponding author. Tel.: +1 (780) 492 7180;E-mail address: rafiq.ahmad@ualberta.ca
Abstract
Aquaponics is a promising sustainable farming method that combines aquaculture and hydroponics. It allows the growth of crops without soil,
pesticides, or fertilizers, and with a minimum amount of water. In aquaponic systems, the design of the growing area is directly linked to the type
of crop about to be planted. The type of crop directly determines, for example, the spacing between plants and between channels, which is critical
to determine the footprint required and estimate the system productivity. This paper proposes a knowledge modeling approach to support the
design of aquaponic systems by automatically determining the required characteristics of the aquaponic system based on crop selection. The
knowledge modeling is outlined as an ontology model that formally describes the existent links between the aquaponic grow bed characteristics
and its design parameters. This study gives practitioners the capacity to visualize the impact of the desired crop selection on the aquaponic system
design, as well as supporting clearer decision-making regarding production facility layout and system design in aquaponic farms.
©2021 The Authors. Published by Elsevier Ltd.
This is an open access article under the CC BY-NC-ND license (https://creativecommons.org/licenses/by-nc-nd/4.0)
Peer-review under responsibility of the scientific committee of the 31st CIRP Design Conference 2021
Keywords: knowledge modeling; aquaponics; precision farming; parametric design; design automation
1. Introduction
Food security and sustainability have become a major
concern over the years due to substantial urbanization,
destruction of the ecological environment, farmlands scarcity,
and increasing growth of the human population [1]. Traditional
agriculture methods employed for crop production require vast
amounts of land, time, and manpower and hence are not very
efficient to meet the current food demands. Consequently, the
current paradigm poses a need to explore new farming methods.
Aquaponics, a subset of integrated agri-aquaculture systems,is
expected to address these problems due to its ability to develop
and achieve economically viable and environmentally
sustainable food production practices [2]. The rationale of this
soilless recirculating growing system involves sharing the
mutual benefit of the available resources,such as water and
nutrients, between aquaculture and plant production.
An aquaponic system is comprised of two integrated units:
1) a hydroponic unit which consists of grow beds for plant
growth; and 2) an aquaculture unit that involves water tanks for
fish habitat and biofilters for the breakdown of ammonia [3].
These units work together in a symbiotic environment to enable
plant and fish growth. Primarily, depending on the structure of
the plants’ grow bed and crop type and size, there are three
different types of aquaponics system designs: nutrient film
technique (NFT), media bed, and deep water culture (DWC)
[4].In this paper, the NFT based aquaponic system is
considered because it is the most popular type of aquaponic
setup used. Moreover, it uses less water and is suitable to grow
leafy green crops. In NFT systems, a very thin film of nutrient-
rich water is pumped to enclosed channels. The top cover of the
channel consists of circular or squared shaped pockets known
as plant sites where plants sit in small plastic cups allowing their
roots to access the water and absorb the nutrients [5].
The design and management of an NFT-based indoor
aquaponic system present several challenges when scaling it to
a commercial level [6].These challenges are mainly
attributable to the design of grow channels based on crop

56 Rabiya Abbasi et al. / Procedia CIRP 100 (2021) 55–60
2 Rabiya Abbasi et al. / Procedia CIRP 00 (2021) 000–000
selection. Each crop has a certain width and height at optimal
environmental conditions that impact the design infrastructure
of the aquaponics system in terms of plant site spacing and
distance between grow channels [7]. This in turn affects the
system productivity which involves crop yields and product
quality. Hence, to ensure high system productivity, the proper
design and placement of grow channels are significant. To
achieve this, the complex and heterogeneous existing links
between grow bed design and crop characteristics need to be
formally described by appropriately capturing the data and
managing the knowledge related to these entities. In this
essence, ontology is regarded as one of the normative
knowledge modeling tools that provide semantic
interoperability and a general understanding of specialized
multidimensional knowledge in various domains that is
cognitively transparent and accessible to human experts and
software agents [8–10]. The ontology models, in combination
with rule systems, act as strong candidates to construct a
decision support platform for the representation of different
knowledge sources and facilitation of knowledge-driven
decisions in a reusable and modular manner [11].
So far, no attempt has been made towards knowledge
modeling of the aquaponics system particularly for the
representation of the grow bed design knowledge based on crop
selection. Therefore, the purpose of this paper is, to provide a
knowledge model in form of an ontology model to support the
parametric design automation in an indoor aquaponics system
with the notion to automatically determine the design
parameters of grow bed based on the crop selection. This
ontology model stores knowledge gathered from farm, domain
experts and databases. The inferred knowledge is then extracted
and used to calculate grow bed design parameters for specific
crop. To streamline the decision-making process a graphical
user interface (GUI) is developed. This research study allows
aquaponics’ practitioners to visualize the impact of crop
selection on aquaponic system design, which eventually will
facilitate in better decision making regarding crop production
in aquaponic farms.
The paper is structured in 8 sections. Section 2 introduces a
knowledge-based decision support framework for the
parametric design automation of aquaponic grow beds based on
crop selection. Section 3 provides the overview of the main
classes of aquaponic ontology, namely, AquaONT and
relationships between them. Governing equations devised to
determine design features of grow bed are described in section
4. The user interface developed to visualize the behavior of
AquaONT is presented in section 5. Section 6 presents use-case
considering basil crop. The analysis of the results obtained in
section 6 are covered in section 7. Finally, section 8 concludes
the paper by addressing the efficacy of this study.
2. Decision support framework for automated design
of aquaponic grow beds.
The effective decision making related to design of grow
beds based on crop selection in aquaponic farms is contingent
upon the representation, extraction, and usage of available
________ ________
1 https://jena.apache.org/tutorials/rdf_api.html
knowledge about contributing entities. For this purpose, a
decision support framework is proposed in this paper, the
layout of which is shown in Fig.1.
Fig. 1. Decision support framework for automated design of grow beds.
The proposed framework consisting of three primary stages
depicts the complete lifecycle of decision making process
based on knowledge extracted from ontology model. To
represent the aquaponics knowledge, first, an ontology model
is developed by acquiring knowledge from farm and domain
experts and unify it as domain concepts. Then, the existing and
inferred knowledge about crop characteristics and grow bed
design features is extracted from the ontology model using
Apache Jena API 1 and SPARQL query language. Then,
MySQL Workbench2, is used to import and organize extracted
knowledge into a database. MATLAB database explorer
toolbox is employed to link this database with the MATLAB
app designer module which along with various mathematical
equations are utilized to develop a graphical user interface
(GUI). Finally, the results (design features) obtained from GUI
are exported to SOLIDWORKS for parametric modeling of the
final grow bed design.
3. AquaONT: an ontology model for aquaponic system
In this section, AquaONT is introduced, which is an OWL
ontology developed to represent and model the essential
knowledge of the aquaponics system. This ontology model is
created in Protégé 5.5, which is an open-source ontology editor
developed by Stanford University. First, the upper-level
ontological knowledge model known as base ontology is
presented that provides the domain-specific concepts related to
the aquaponic system. Then, product and production system
concepts are presented, that define the crop characteristics and
grow bed features, respectively.
2 https://www.mysql.com/products/workbench.
Rabiya Abbasi et al. / Procedia CIRP 00 (2021) 000–000 3
3.1. Upper level ontological knowledge model
An ontology model, O, represents the dimensions of
domain-specific knowledge in terms of four fundamental
elements referred to as a tuple: O = {C, I, OP, DP}, where the
concept (C) is a set of instances, the instances (I) are the objects
in the domain, the object property (OP) is the relationship
between two concepts or instances, and the datatype property
(DP) links instances with literal variables (integer or string)
[12]. Fig.2 shows upper level ontological model of AquaONT,
also known as the base ontology model. Six “classes” or
“concepts” are created to represent the six knowledge domains.
These concepts are related to each other through object
properties, which are given in Table 1.
Fig. 2. Upper level ontological model of AquaONT
Table 1. Relationships between classes/concepts.
Domain
Object property
Range
Ambient_Environment,
Production_System
have_Impact_on
Product_Quality
Production_Facility
is_Maintained_at
Ambient_Environment
Contextual_Data
is_Received_from
Production_Facility
Consumer_Product
is_Output_of
Production_System
Product_Quality
is_Characteristic_of
Consumer_Product
Production_System
is_Established_in
Production_Facility
The class Ambient_Environment specifies the optimal
ranges of environmental parameters that ensure the healthy
growth of crops and fishes in an indoor aquaponics system.
These parameters are classified into two categories: 1) indoor
environmental parameters which include water temperature,
pH, electroconductivity, ammonia, dissolved oxygen, nitrate
and nitrite level, water hardness, water level, water flowrate,
alkalinity, salinity, air temperature, light intensity, humidity,
and CO2; and 2) outdoor environmental parameters which
involve the daily weather conditions, routine climatic changes,
day-night times, and seasons.
The notion of product in any production system refers to the
outcome of the process [13]. In an aquaponics system, there are
two primary products: ready-to-harvest crops and fish.
Consumer_Product class represents the product knowledge in
terms of crop and fish type, crop and fish growth status, and
crop and fish optimal growth parameters. A wide variety of
crops can be grown in an aquaponic system, but each crop
needs distinct environment to thrive, and has its own standard
height and width at maturity stage or at the time of harvesting.
These aspects are significant in determining the design of
growbeds and therefore, are also represented under this class.
Besides biological components, an indoor aquaponics
system consists of various mechanical and electrical
components. Production_System class models the knowledge
about these components under the subclasses digital system
and mechanical system. The digital system is further
categorized to include sensors, controllers, and other electronic
or network devices. Whereas the mechanical system subclass
represents design features of grow beds, fish tanks, and
biofiltration tanks with respect to crop and fish type.
In an indoor aquaponic system, the idea is to control and
maintain the optimal environmental conditions to enhance the
crop yields, for which location of the system plays a significant
role. Production_Facility class, therefore, specifies the location
where the aquaponics system is located and managed. This
class also represents the workers that are responsible for
managing each part of the aquaponics system through a
centralized system.
For remote monitoring and control of the aquaponic system,
context information is obtained from sensors through proper
connectivity channels and is utilized to enable data-driven
decisions in the knowledge model. This context information is
related to real-time data of surrounding conditions in
aquaponics farms and is therefore represented under the class
Contextual_Data.
The Product_Quality concept models the product qualitative
aspects, quality control standards, and quality assessment
criteria and links these attributes with the knowledge
represented for a consumer product, production system, and
ambient environment covered in previous concepts.
To verify AquaONT, Protégé built-in reasoner, HermiT
was used. The computation was done successfully without
errors, showing the accuracy of ontology. Similarly, to validate
ontology, Sparql queries were developed and executed. Every
time, these queries produce the same results for the given
conditions, representing the consistency and coherence of
ontology.
3.2. Consumer product and production system concepts
Ontologies enable the interoperability of autonomous agents
and support the design of production systems [14]. In this
study, AquaONT is used to enable parametric design
automation – involving determination of design features of
aquaponic grow beds pertaining to each crop. To achieve this,
two concepts, namely, Consumer_Product and
Production_System are employed and extended to include
several sub-concepts which are then populated with the
knowledge of grow bed design features and crop characteristics
gathered from domain experts and farm. The detailed
hierarchical architecture of these two concepts along with
significant sub-concepts and instances is shown in Fig. 3.
The different types of crops are defined as instances (Icrop)
under the sub-concept Crop_Type. The crops considered in this
study are leafy green vegetables: basil, chard, lettuce, parsley,
and spinach. The characteristics of these crops involve standard
plant spacing (PS), width (Wi), and height (H) as recommended
by aquaponics professionals. PS is defined as the distance
between two consecutive plants on the same channel. These
characteristics are the attributes of respective crop represented
as literals and linked with instances through corresponding
datatype properties: “hasPlantSpacing”, “hasPlantWidth”, and
“hasPlantHeight” respectively.

Rabiya Abbasi et al. / Procedia CIRP 100 (2021) 55–60 57
2 Rabiya Abbasi et al. / Procedia CIRP 00 (2021) 000–000
selection. Each crop has a certain width and height at optimal
environmental conditions that impact the design infrastructure
of the aquaponics system in terms of plant site spacing and
distance between grow channels [7]. This in turn affects the
system productivity which involves crop yields and product
quality. Hence, to ensure high system productivity, the proper
design and placement of grow channels are significant. To
achieve this, the complex and heterogeneous existing links
between grow bed design and crop characteristics need to be
formally described by appropriately capturing the data and
managing the knowledge related to these entities. In this
essence, ontology is regarded as one of the normative
knowledge modeling tools that provide semantic
interoperability and a general understanding of specialized
multidimensional knowledge in various domains that is
cognitively transparent and accessible to human experts and
software agents [8–10]. The ontology models, in combination
with rule systems, act as strong candidates to construct a
decision support platform for the representation of different
knowledge sources and facilitation of knowledge-driven
decisions in a reusable and modular manner [11].
So far, no attempt has been made towards knowledge
modeling of the aquaponics system particularly for the
representation of the grow bed design knowledge based on crop
selection. Therefore, the purpose of this paper is, to provide a
knowledge model in form of an ontology model to support the
parametric design automation in an indoor aquaponics system
with the notion to automatically determine the design
parameters of grow bed based on the crop selection. This
ontology model stores knowledge gathered from farm, domain
experts and databases. The inferred knowledge is then extracted
and used to calculate grow bed design parameters for specific
crop. To streamline the decision-making process a graphical
user interface (GUI) is developed. This research study allows
aquaponics’ practitioners to visualize the impact of crop
selection on aquaponic system design, which eventually will
facilitate in better decision making regarding crop production
in aquaponic farms.
The paper is structured in 8 sections. Section 2 introduces a
knowledge-based decision support framework for the
parametric design automation of aquaponic grow beds based on
crop selection. Section 3 provides the overview of the main
classes of aquaponic ontology, namely, AquaONT and
relationships between them. Governing equations devised to
determine design features of grow bed are described in section
4. The user interface developed to visualize the behavior of
AquaONT is presented in section 5. Section 6 presents use-case
considering basil crop. The analysis of the results obtained in
section 6 are covered in section 7. Finally, section 8 concludes
the paper by addressing the efficacy of this study.
2. Decision support framework for automated design
of aquaponic grow beds.
The effective decision making related to design of grow
beds based on crop selection in aquaponic farms is contingent
upon the representation, extraction, and usage of available
________ ________
1 https://jena.apache.org/tutorials/rdf_api.html
knowledge about contributing entities. For this purpose, a
decision support framework is proposed in this paper, the
layout of which is shown in Fig.1.
Fig. 1. Decision support framework for automated design of grow beds.
The proposed framework consisting of three primary stages
depicts the complete lifecycle of decision making process
based on knowledge extracted from ontology model. To
represent the aquaponics knowledge, first, an ontology model
is developed by acquiring knowledge from farm and domain
experts and unify it as domain concepts. Then, the existing and
inferred knowledge about crop characteristics and grow bed
design features is extracted from the ontology model using
Apache Jena API 1 and SPARQL query language. Then,
MySQL Workbench2, is used to import and organize extracted
knowledge into a database. MATLAB database explorer
toolbox is employed to link this database with the MATLAB
app designer module which along with various mathematical
equations are utilized to develop a graphical user interface
(GUI). Finally, the results (design features) obtained from GUI
are exported to SOLIDWORKS for parametric modeling of the
final grow bed design.
3. AquaONT: an ontology model for aquaponic system
In this section, AquaONT is introduced, which is an OWL
ontology developed to represent and model the essential
knowledge of the aquaponics system. This ontology model is
created in Protégé 5.5, which is an open-source ontology editor
developed by Stanford University. First, the upper-level
ontological knowledge model known as base ontology is
presented that provides the domain-specific concepts related to
the aquaponic system. Then, product and production system
concepts are presented, that define the crop characteristics and
grow bed features, respectively.
2 https://www.mysql.com/products/workbench.
Rabiya Abbasi et al. / Procedia CIRP 00 (2021) 000–000 3
3.1. Upper level ontological knowledge model
An ontology model, O, represents the dimensions of
domain-specific knowledge in terms of four fundamental
elements referred to as a tuple: O = {C, I, OP, DP}, where the
concept (C) is a set of instances, the instances (I) are the objects
in the domain, the object property (OP) is the relationship
between two concepts or instances, and the datatype property
(DP) links instances with literal variables (integer or string)
[12]. Fig.2 shows upper level ontological model of AquaONT,
also known as the base ontology model. Six “classes” or
“concepts” are created to represent the six knowledge domains.
These concepts are related to each other through object
properties, which are given in Table 1.
Fig. 2. Upper level ontological model of AquaONT
Table 1. Relationships between classes/concepts.
Domain
Object property
Range
Ambient_Environment,
Production_System
have_Impact_on
Product_Quality
Production_Facility
is_Maintained_at
Ambient_Environment
Contextual_Data
is_Received_from
Production_Facility
Consumer_Product
is_Output_of
Production_System
Product_Quality
is_Characteristic_of
Consumer_Product
Production_System
is_Established_in
Production_Facility
The class Ambient_Environment specifies the optimal
ranges of environmental parameters that ensure the healthy
growth of crops and fishes in an indoor aquaponics system.
These parameters are classified into two categories: 1) indoor
environmental parameters which include water temperature,
pH, electroconductivity, ammonia, dissolved oxygen, nitrate
and nitrite level, water hardness, water level, water flowrate,
alkalinity, salinity, air temperature, light intensity, humidity,
and CO2; and 2) outdoor environmental parameters which
involve the daily weather conditions, routine climatic changes,
day-night times, and seasons.
The notion of product in any production system refers to the
outcome of the process [13]. In an aquaponics system, there are
two primary products: ready-to-harvest crops and fish.
Consumer_Product class represents the product knowledge in
terms of crop and fish type, crop and fish growth status, and
crop and fish optimal growth parameters. A wide variety of
crops can be grown in an aquaponic system, but each crop
needs distinct environment to thrive, and has its own standard
height and width at maturity stage or at the time of harvesting.
These aspects are significant in determining the design of
growbeds and therefore, are also represented under this class.
Besides biological components, an indoor aquaponics
system consists of various mechanical and electrical
components. Production_System class models the knowledge
about these components under the subclasses digital system
and mechanical system. The digital system is further
categorized to include sensors, controllers, and other electronic
or network devices. Whereas the mechanical system subclass
represents design features of grow beds, fish tanks, and
biofiltration tanks with respect to crop and fish type.
In an indoor aquaponic system, the idea is to control and
maintain the optimal environmental conditions to enhance the
crop yields, for which location of the system plays a significant
role. Production_Facility class, therefore, specifies the location
where the aquaponics system is located and managed. This
class also represents the workers that are responsible for
managing each part of the aquaponics system through a
centralized system.
For remote monitoring and control of the aquaponic system,
context information is obtained from sensors through proper
connectivity channels and is utilized to enable data-driven
decisions in the knowledge model. This context information is
related to real-time data of surrounding conditions in
aquaponics farms and is therefore represented under the class
Contextual_Data.
The Product_Quality concept models the product qualitative
aspects, quality control standards, and quality assessment
criteria and links these attributes with the knowledge
represented for a consumer product, production system, and
ambient environment covered in previous concepts.
To verify AquaONT, Protégé built-in reasoner, HermiT
was used. The computation was done successfully without
errors, showing the accuracy of ontology. Similarly, to validate
ontology, Sparql queries were developed and executed. Every
time, these queries produce the same results for the given
conditions, representing the consistency and coherence of
ontology.
3.2. Consumer product and production system concepts
Ontologies enable the interoperability of autonomous agents
and support the design of production systems [14]. In this
study, AquaONT is used to enable parametric design
automation – involving determination of design features of
aquaponic grow beds pertaining to each crop. To achieve this,
two concepts, namely, Consumer_Product and
Production_System are employed and extended to include
several sub-concepts which are then populated with the
knowledge of grow bed design features and crop characteristics
gathered from domain experts and farm. The detailed
hierarchical architecture of these two concepts along with
significant sub-concepts and instances is shown in Fig. 3.
The different types of crops are defined as instances (Icrop)
under the sub-concept Crop_Type. The crops considered in this
study are leafy green vegetables: basil, chard, lettuce, parsley,
and spinach. The characteristics of these crops involve standard
plant spacing (PS), width (Wi), and height (H) as recommended
by aquaponics professionals. PS is defined as the distance
between two consecutive plants on the same channel. These
characteristics are the attributes of respective crop represented
as literals and linked with instances through corresponding
datatype properties: “hasPlantSpacing”, “hasPlantWidth”, and
“hasPlantHeight” respectively.

58 Rabiya Abbasi et al. / Procedia CIRP 100 (2021) 55–60
4 Rabiya Abbasi et al. / Procedia CIRP 00 (2021) 000–000
Fig. 3. Architecture of Consumer_Product and Production_System class.
Likewise, the design parameters of the grow channels are
modeled under the instances (Idesign) of the sub-concept
NFT_Grow_Channel. These instances represent different
design categories, and each category specifies a certain width
(W), length (L), depth (D), plant site spacing (S), plant site size
(SS), vertical channel spacing (VCS), and horizontal channel
spacing (HCS) of a NFT grow system. These parameters are
the attributes represented as literals and linked with the design
categories through datatype properties “hasWidth”,
“hasLength”, “hasDepth”, “hasPlantSiteSpacing”,
“hasPlantSiteSize”, “hasVerticalChannelSpacing”, and
“hasHorizontalChannelSpacing” respectively. Fig. 4 shows the
crops’ basic dimensional characteristics and generalized design
features of a NFT grow channel.
Fig. 4. a) Crop characteristics; b) Grow channel design features
4. Calculation of grow bed design parameters
Using the attributes specified for instances of sub-concepts-
crop type and NFT grow channel given in section 5 - equations
are developed to calculate design parameters of grow bed. For
instance, PS and L are used to determine the number of plant
sites per channel (NPSC). NPSC is defined as the capacity of
each channel to grow number of plants. In Fig. 4b, NPSC is 8,
which implies that in this particular channel, only 8 plants can
be grown. The S on the grow channel is directly related to PS
and is essentially important to ensure high crop yields. Other
yield parameters that are impacted by PS in the aquaponic
system are plant height, leaf area, and leaf number. The general
rule of thumb in this essence is to build plant sites on each
channel and keep the spacing of channels according to the
expected width of the plant at its maturity stage [15]. NPSC
along with the total number of channels (NC) needed to build
the complete hydroponic unit determines the production
capacity (PC) of the aquaponic system, which is equivalent to
the maximum possible crop yield. Equations (1) and (2) are
developed for computing NPSC and PC, respectively.
 = 
⁄
(1)
 =  ×  (2)
The grow channels can be stacked horizontally or vertically
or both by maintaining the recommended HCS and VCS.
Moreover, the farm space must also be taken into consideration
while choosing NC and respective stacking setup. With
horizontally stacked NFT channels, the length of the fully
developed hydroponic unit is the same as the length of the grow
channel, L, whereas the width of (WHU) is equivalent to the
sum of widths of all channels and horizontal spacings between
channels. Equation (3) is formulated to determine WHU.
 = ( × ) + (( − 1) × ) (3)
Another significant agronomic factor that enhances the crop
yield is plant density or plant population (PD). PD measures
the number of plants per unit area and its optimum value varies
with the genotype and geographic location [16]. In aquaponic
systems, the number of plants to be grown refers to the
production capacity of the system, whereas the unit area is
related to the area of hydroponic component. To compute PD,
Equation 4 is devised.
 = ( × )
⁄ (4)
These equations use the existing and inferred knowledge
from AquaONT to determine mentioned design features and to
visualize this, GUI is developed which is explained in the next
section.
5. AquaONT application: Graphical User Interface
To visualize the behavior of AquaONT, a GUI is developed
using MATLAB app designer tool which is shown in Fig.5.
This GUI uses inferred knowledge from AquaONT, and
equations developed in section 4. It allows users to make a crop
and a channel length selection and observe the impact on design
parameters in terms of numerical value. For better visualization
of design variations in the grow channel as a 3D CAD model,
these numeric values are sent to SOLIDWORKS, where they
are applied to the already built design referred to as default
parametric design.
Five fields are created on the GUI to represent the
knowledge of the ontology model: 1) Crop Field, 2) Grow Bed
Design Field, 3) Environmental Parameters, 4) NFT Channel
Selection, and 5) NFT based Crop Production System. First
four fields are populated with existing and inferred knowledge
from AquaONT - acquired directly through SQL database,
whereas the last field is linked with the set of equations created
in section 4. The Crop Field describes the five leafy green crops
along with their characteristics such as H, Wi, and PS [17]. The
Grow Bed Design Field gives information about the grow bed
type, PS, HCS, and VCS of each crop. The Environmental
Parameters field specifies the optimal growth conditions for
these crops. The entries of field 2 and 3 are auto-populated once
the crop is selected. For the selection of the right NFT channel,
the NFT Channel Selection field is incorporated, where length
of channel is the deciding factor. The channel lengths
considered are 6 feet, 8 feet, 10 feet, and 12 feet.
Rabiya Abbasi et al. / Procedia CIRP 00 (2021) 000–000 5
The other parameters under this field such as width and
depth of the channel are kept constant for the sake of
simplifying the model. Moreover, the shape of plant site is
chosen to be circular with diameter of 2 inches. The plant site
can also be squared shape. The last field on the GUI is the NFT
based Crop Production System. This field uses entries of
previous fields and governing equations given in section 4 in
order to calculate parameters. This field is important as it gives
information about the production capacity of the system along
with the length and width of the complete hydroponic unit once
the user selects the number of channels. In addition, three
auxiliary fields are created at the lower side of the GUI
window, which displays the total area of the hydroponic unit,
total growing area, and plant density (plant population).
6. Use Case – Grow bed design for basil crop
The use-case presented here aims to illustrate the feasibility
of AquaONT and GUI. For this purpose, basil crop is
considered, which is one of the most common economically
viable products in aquaponic systems. The optimal
environmental conditions to grow basil in indoor farms,
standard height, and width under these conditions, and HCS,
VCS and PS are shown in Fig. 5. These values are extracted
from AquaONT. Assuming the user selects 6 feet long NFT
channel for its aquaponic system and his/her farm can
accommodate a maximum of four channels. After entering
these values in the relevant fields in GUI, the design parameters
under the fifth field are automatically calculated. For the given
inputs such as L = 6ft and NC = 4, the results show that only 7
basil plants per channel can be grown and these plants must be
placed 10 inches apart on each channel. In addition, each
channel must be placed at a distance of 6 inches from each
other. The application also calculates total area, the effective
growing area, and the PD of the hydroponic unit which in case
of basil are: 18.25 ft2, 0.61ft2, and 2/ft2, respectively, see Fig.5.
Finally, to visualize the CAD model of NFT grow system
for basil, the calculated design parameters from MATLAB are
imported in SOLIDWORKS. These parameters are saved in a
design table which enables parametric modeling. The idea is to
develop a default design of a grow system in CAD software and
automatically update it with a single click without designing
the entire part or assembly again by using the new design
details stored in design table. This process is showcased by
presenting the basic case of basil crop. The default and updated
grow channel design for basil is shown in Fig. 6. Before
implementing the parameters saved in design table, L = 96in
with NPSC = 8 for default design but after application, L
becomes 72in with NPSC reduced to 7 – showing the updated
design configuration for basil. The process is repeated for basil,
lettuce, and parsley for different input values. The results
obtained are explained in next section.
Fig. 6. a) Default grow bed design. b) Updated grow bed design for basil.
7. Results and discussion
The proposed system is simulated for all the crops mentioned
in section 3.2. Fig.7 shows the design configurations of
hydroponic unit for three crops with two different input sets –
including {L, NC} = {72in, 4} and {96in, 6}. The results show
that for same channel length, NPSC is different for each crop.
This is due to the distinct requirement of plant site spacing (S)
for each crop such as {Sbasil, Slettuce, Sparsley} = {10,8,12}.
Similarly, the production capacity of the hydroponic unit is also
different for each crop. For the same NC, it is observed that the
PC of the system for lettuce is 22.22% and 33.33% higher than
for basil and parsley, respectively. If L is increased from 72in
to 96in and NC is increased from 4 to 6, the resulting NPSC
and PC will also be increased. For instance, in Fig. 7 (e, f)
NPSC and PC for parsley are increased from 6 and 24 to 8 and
48, respectively. With these visualization results in place, crop
Fig. 5. Graphical User Interface for AquaONT design application developed by LIMDA, University of Alberta.

Rabiya Abbasi et al. / Procedia CIRP 100 (2021) 55–60 59
4 Rabiya Abbasi et al. / Procedia CIRP 00 (2021) 000–000
Fig. 3. Architecture of Consumer_Product and Production_System class.
Likewise, the design parameters of the grow channels are
modeled under the instances (Idesign) of the sub-concept
NFT_Grow_Channel. These instances represent different
design categories, and each category specifies a certain width
(W), length (L), depth (D), plant site spacing (S), plant site size
(SS), vertical channel spacing (VCS), and horizontal channel
spacing (HCS) of a NFT grow system. These parameters are
the attributes represented as literals and linked with the design
categories through datatype properties “hasWidth”,
“hasLength”, “hasDepth”, “hasPlantSiteSpacing”,
“hasPlantSiteSize”, “hasVerticalChannelSpacing”, and
“hasHorizontalChannelSpacing” respectively. Fig. 4 shows the
crops’ basic dimensional characteristics and generalized design
features of a NFT grow channel.
Fig. 4. a) Crop characteristics; b) Grow channel design features
4. Calculation of grow bed design parameters
Using the attributes specified for instances of sub-concepts-
crop type and NFT grow channel given in section 5 - equations
are developed to calculate design parameters of grow bed. For
instance, PS and L are used to determine the number of plant
sites per channel (NPSC). NPSC is defined as the capacity of
each channel to grow number of plants. In Fig. 4b, NPSC is 8,
which implies that in this particular channel, only 8 plants can
be grown. The S on the grow channel is directly related to PS
and is essentially important to ensure high crop yields. Other
yield parameters that are impacted by PS in the aquaponic
system are plant height, leaf area, and leaf number. The general
rule of thumb in this essence is to build plant sites on each
channel and keep the spacing of channels according to the
expected width of the plant at its maturity stage [15]. NPSC
along with the total number of channels (NC) needed to build
the complete hydroponic unit determines the production
capacity (PC) of the aquaponic system, which is equivalent to
the maximum possible crop yield. Equations (1) and (2) are
developed for computing NPSC and PC, respectively.
 = 
⁄ (1)
 =  ×  (2)
The grow channels can be stacked horizontally or vertically
or both by maintaining the recommended HCS and VCS.
Moreover, the farm space must also be taken into consideration
while choosing NC and respective stacking setup. With
horizontally stacked NFT channels, the length of the fully
developed hydroponic unit is the same as the length of the grow
channel, L, whereas the width of (WHU) is equivalent to the
sum of widths of all channels and horizontal spacings between
channels. Equation (3) is formulated to determine WHU.
 = ( × ) + (( − 1) × ) (3)
Another significant agronomic factor that enhances the crop
yield is plant density or plant population (PD). PD measures
the number of plants per unit area and its optimum value varies
with the genotype and geographic location [16]. In aquaponic
systems, the number of plants to be grown refers to the
production capacity of the system, whereas the unit area is
related to the area of hydroponic component. To compute PD,
Equation 4 is devised.
 = ( × )
⁄ (4)
These equations use the existing and inferred knowledge
from AquaONT to determine mentioned design features and to
visualize this, GUI is developed which is explained in the next
section.
5. AquaONT application: Graphical User Interface
To visualize the behavior of AquaONT, a GUI is developed
using MATLAB app designer tool which is shown in Fig.5.
This GUI uses inferred knowledge from AquaONT, and
equations developed in section 4. It allows users to make a crop
and a channel length selection and observe the impact on design
parameters in terms of numerical value. For better visualization
of design variations in the grow channel as a 3D CAD model,
these numeric values are sent to SOLIDWORKS, where they
are applied to the already built design referred to as default
parametric design.
Five fields are created on the GUI to represent the
knowledge of the ontology model: 1) Crop Field, 2) Grow Bed
Design Field, 3) Environmental Parameters, 4) NFT Channel
Selection, and 5) NFT based Crop Production System. First
four fields are populated with existing and inferred knowledge
from AquaONT - acquired directly through SQL database,
whereas the last field is linked with the set of equations created
in section 4. The Crop Field describes the five leafy green crops
along with their characteristics such as H, Wi, and PS [17]. The
Grow Bed Design Field gives information about the grow bed
type, PS, HCS, and VCS of each crop. The Environmental
Parameters field specifies the optimal growth conditions for
these crops. The entries of field 2 and 3 are auto-populated once
the crop is selected. For the selection of the right NFT channel,
the NFT Channel Selection field is incorporated, where length
of channel is the deciding factor. The channel lengths
considered are 6 feet, 8 feet, 10 feet, and 12 feet.
Rabiya Abbasi et al. / Procedia CIRP 00 (2021) 000–000 5
The other parameters under this field such as width and
depth of the channel are kept constant for the sake of
simplifying the model. Moreover, the shape of plant site is
chosen to be circular with diameter of 2 inches. The plant site
can also be squared shape. The last field on the GUI is the NFT
based Crop Production System. This field uses entries of
previous fields and governing equations given in section 4 in
order to calculate parameters. This field is important as it gives
information about the production capacity of the system along
with the length and width of the complete hydroponic unit once
the user selects the number of channels. In addition, three
auxiliary fields are created at the lower side of the GUI
window, which displays the total area of the hydroponic unit,
total growing area, and plant density (plant population).
6. Use Case – Grow bed design for basil crop
The use-case presented here aims to illustrate the feasibility
of AquaONT and GUI. For this purpose, basil crop is
considered, which is one of the most common economically
viable products in aquaponic systems. The optimal
environmental conditions to grow basil in indoor farms,
standard height, and width under these conditions, and HCS,
VCS and PS are shown in Fig. 5. These values are extracted
from AquaONT. Assuming the user selects 6 feet long NFT
channel for its aquaponic system and his/her farm can
accommodate a maximum of four channels. After entering
these values in the relevant fields in GUI, the design parameters
under the fifth field are automatically calculated. For the given
inputs such as L = 6ft and NC = 4, the results show that only 7
basil plants per channel can be grown and these plants must be
placed 10 inches apart on each channel. In addition, each
channel must be placed at a distance of 6 inches from each
other. The application also calculates total area, the effective
growing area, and the PD of the hydroponic unit which in case
of basil are: 18.25 ft2, 0.61ft2, and 2/ft2, respectively, see Fig.5.
Finally, to visualize the CAD model of NFT grow system
for basil, the calculated design parameters from MATLAB are
imported in SOLIDWORKS. These parameters are saved in a
design table which enables parametric modeling. The idea is to
develop a default design of a grow system in CAD software and
automatically update it with a single click without designing
the entire part or assembly again by using the new design
details stored in design table. This process is showcased by
presenting the basic case of basil crop. The default and updated
grow channel design for basil is shown in Fig. 6. Before
implementing the parameters saved in design table, L = 96in
with NPSC = 8 for default design but after application, L
becomes 72in with NPSC reduced to 7 – showing the updated
design configuration for basil. The process is repeated for basil,
lettuce, and parsley for different input values. The results
obtained are explained in next section.
Fig. 6. a) Default grow bed design. b) Updated grow bed design for basil.
7. Results and discussion
The proposed system is simulated for all the crops mentioned
in section 3.2. Fig.7 shows the design configurations of
hydroponic unit for three crops with two different input sets –
including {L, NC} = {72in, 4} and {96in, 6}. The results show
that for same channel length, NPSC is different for each crop.
This is due to the distinct requirement of plant site spacing (S)
for each crop such as {Sbasil, Slettuce, Sparsley} = {10,8,12}.
Similarly, the production capacity of the hydroponic unit is also
different for each crop. For the same NC, it is observed that the
PC of the system for lettuce is 22.22% and 33.33% higher than
for basil and parsley, respectively. If L is increased from 72in
to 96in and NC is increased from 4 to 6, the resulting NPSC
and PC will also be increased. For instance, in Fig. 7 (e, f)
NPSC and PC for parsley are increased from 6 and 24 to 8 and
48, respectively. With these visualization results in place, crop
Fig. 5. Graphical User Interface for AquaONT design application developed by LIMDA, University of Alberta.

60 Rabiya Abbasi et al. / Procedia CIRP 100 (2021) 55–60
6Rabiya Abbasi et al. / Procedia CIRP 00 (2021) 000–000
characteristics such as PS, Wi, and H significantly impact the
design parameters of grow channel in an aquaponic system.
Having a correct grow bed design in aquaponics system for
crop growth is crucial because it ensures high yields. Moreover,
it also ensures the right amount of water and nutrient
absorptions, that eventually leads to high crop quality with
right nutritional value. In this essence, a quick knowledge-
based virtual tool assists in decision making related to the
proper design of grow bed based on crop characteristics.
For future work, intelligent techniques such as machine
learning, deep learning, and computer vision will be
incorporated to make the system smart and autonomous.
Moreover, a cost model will also be integrated to optimize the
aquaponic grow beds based on market demand.
8. Conclusions
Aiming at providing a knowledge-based system for
automated decision-making regarding crop production and
respective grow bed design in aquaponics farms, this paper has
proposed a decision support framework. An ontology model,
AquaONT,is developed to assist in decision making process,
which can be extended to include other elements and tested
against robust case studies. GUI is developed that uses inferred
and existing knowledge from AquaONT and mathematical
equations to calculate design parameters. To visualize the
impact of crop selection on the design of grow beds, parametric
modeling is performed.The analysis of results shows that the
correct design of grow bed ensureshigh crop yield and quality.
Acknowledgments: The authors acknowledge the financial
support of this work by the Natural Sciences and Engineering
Research Council of Canada (NSERC) (Grant File No. ALLRP
545537-19 and RGPIN-2017-04516).
References
[1] T.Y. Kyaw, A.K. Ng, Smart Aquaponics System for Urban Farming, in:
Energy Procedia, 2017. https://doi.org/10.1016/j.egypro.2017.12.694.
[2] G.J. Gooley, F.M. Gavine, Integrated systems Agri-Aquaculture Systems
A resouce handbook for australian industry development., 2003.
[3] W. Lennard, S. Goddek, Aquaponics: The Basics, in: Aquaponics Food
Production Systems, 2019. https://doi.org/10.1007/978-3-030-15943-
6_5.
[4] L. Pérez-Urrestarazu, J. Lobillo-Eguíbar, R. Fernández-Cañero, V.M.
Fernández-Cabanás, Suitability and optimization of FAO’s small-scale
aquaponics systems for joint production of lettuce (Lactuca sativa) and
fish (Carassius auratus), Aquacultural Engineering. (2019).
https://doi.org/10.1016/j.aquaeng.2019.04.001.
[5] B.C. Mohapatra, N.K. Chandan, S.K. Panda, D. Majhi, B.R. Pillai,
Design, and development of a portable and streamlined nutrient film
technique (NFT) aquaponic system, Aquacultural Engineering. (2020).
https://doi.org/10.1016/j.aquaeng.2020.102100.
[6] A.R. Yanes, P. Martinez, R. Ahmad, Towards automated aquaponics: A
review on monitoring, IoT, and smart systems, Journal of Cleaner
Production. (2020). https://doi.org/10.1016/j.jclepro.2020.121571.
[7] M.M. Maboko, C.P. du Plooy, Effect of plant spacing on growth and yield
of lettuce (lactuca sativa L.) in a soilless production system, South
African Journal of Plant and Soil. (2009).
https://doi.org/10.1080/02571862.2009.10639954.
[8] D.L. McGuinness, F. van Harmelen, OWL Web Ontology Language
Overview, W3C Recommendation. (2004).
[9] S. el Kadiri, B. Grabot, K.D. Thoben, K. Hribernik, C. Emmanouilidis,
G. von Cieminski, D. Kiritsis, Current trends on ICT technologies for
enterprise information systems, Computers in Industry. (2016).
https://doi.org/10.1016/j.compind.2015.06.008.
[10] R. Ahmad, S. Tichadou, J.Y. Hascoet, A knowledge-based intelligent
decision system for production planning, International Journal of
Advanced Manufacturing Technology. (2017).
https://doi.org/10.1007/s00170-016-9214-z.
[11] C. Zhang, G. Zhou, Q. Lu, Decision support oriented ontological
modeling of product knowledge, in: Proceedings of the 2017 IEEE 2nd
Information Technology, Networking, Electronic and Automation
Control Conference, ITNEC 2017, 2018.
https://doi.org/10.1109/ITNEC.2017.8284816.
[12] F. Zhang, Z.M. Ma, W. Li, Storing OWL ontologies in object-oriented
databases, Knowledge-Based Systems. (2015).
https://doi.org/10.1016/j.knosys.2014.12.020.
[13] E.M. Sanfilippo, Ontological foundations for feature-based modeling, in:
Procedia CIRP, 2018. https://doi.org/10.1016/j.procir.2018.03.002.
[14] P. Martinez, R. Ahmad, M. Al-Hussein, Automatic Selection Tool of
Quality Control Specifications for Off-site Construction Manufacturing
Products: A BIM-based Ontology Model Approach, Modular and Offsite
Construction (MOC) Summit Proceedings. (2019).
https://doi.org/10.29173/mocs87.
[15]A. Reyes-Yanes, P. Martinez, R. Ahmad, Real-time growth rate and fresh
weight estimation for little gem romaine lettuce in aquaponic grow beds,
Computers and Electronics in Agriculture. (2020).
https://doi.org/10.1016/j.compag.2020.105827.
[16]R. Bulgari, A. Baldi, A. Ferrante, A. Lenzi, Yield and quality of basil,
Swiss chard, and rocket microgreens grown in a hydroponic system, New
Zealand Journal of Crop and Horticultural Science. (2017).
https://doi.org/10.1080/01140671.2016.1259642.
[17] C. Somerville, M. Cohen, E. Pantanella, A. Stankus, A. Lovatelli, Small-
scale aquaponic food production. Integrated fish and plant farming, 2014.
Fig. 7. NFT grow bed design configurations for different crops: (a, d) Basil; (b, e) Lettuce; (c, f) Parsley