Technical Report LPT-2008-28

Mathematical Models

J. Morbach, A. Yang, W. Marquardt

July 2008

Enquiries should be addressed to:

RWTH Aachen University
Aachener Verfahrenstechnik
Process Systems Engineering
52056 Aachen
Tel.: +49 (0) 241 80 - 94668
Fax: +49 (0) 241 80 - 92326
E-Mail: [email protected]
Table of Contents
Table of Contents .................................................................................................................................... 2
List of Figures ......................................................................................................................................... 4
1. Mathematical Models ...................................................................................................................... 5
1.1. Mathematical Model .............................................................................................................. 5
Usage............................................................................................................................................... 8
Concept Definitions......................................................................................................................... 9
Class Descriptions ........................................................................................................................... 9
Relation Descriptions .................................................................................................................... 12
Attribute Descriptions ................................................................................................................... 13
1.2. Equation System .................................................................................................................. 14
Usage............................................................................................................................................. 15
Concept Descriptions .................................................................................................................... 15
Class Descriptions ......................................................................................................................... 15
Relation Descriptions .................................................................................................................... 19
Attribute Descriptions ................................................................................................................... 20
Individual Descriptions ................................................................................................................. 20
1.3. Numerical Solution Strategy................................................................................................ 23
Concept Descriptions .................................................................................................................... 24
Class Descriptions ......................................................................................................................... 24
Relation Descriptions .................................................................................................................... 26
Attribute Descriptions ................................................................................................................... 27
Individual Descriptions ................................................................................................................. 27
1.4. Cost Model........................................................................................................................... 28
Concept Descriptions .................................................................................................................... 28
Class Descriptions ......................................................................................................................... 28
1.5. Process model ...................................................................................................................... 32
Concept Descriptions .................................................................................................................... 33
Class Descriptions ......................................................................................................................... 33
Relation Descriptions .................................................................................................................... 34
Individual Descriptions ................................................................................................................. 35
1.6. Laws..................................................................................................................................... 36
Concept Descriptions .................................................................................................................... 37
1.7. Property models ................................................................................................................... 44
 Concept Descriptions .................................................................................................................... 45
1.8. Process unit models.............................................................................................................. 50
References ............................................................................................................................................. 51
Index of Concepts.................................................................................................................................. 52

3
 List of Figures
Fig. 1: Overview on partial model mathematical_model ...................................................................... 5
Fig. 2: Mathematical model, model quantity, and model quantity specification ................................................ 6
Fig. 3: Assignment of particular model quantity specifications to model quantites....................................... 6
Fig. 4: Correspondence between a model quantity and a physical quantity of the modeled object ................ 6
Fig. 5: Variables, ports, couplings........................................................................................................... 7
Fig. 6: Exemplary decomposition of model M into submodels M1 and M2 ............................................. 7
Fig. 7: Specification of the overall model ............................................................................................... 8
Fig. 8: Specification of the repetitive submodel structure....................................................................... 9
Fig. 9: Equation system characteristics ................................................................................................. 14
Fig. 10: Numerical solution strategy. .................................................................................................... 23
Fig. 11: Types of algebraic model solution strategies ................................................................................ 23
Fig. 12: Further specification of ODE solution strategy ........................................................................... 24
Fig. 13: Models for estimating the fixed capital investment ................................................................. 28
Fig. 14: Overview on process_model.................................................................................................... 32
Fig. 15: Laws and property models ....................................................................................................... 32
Fig. 16: Exemplary law modeling thermal equilibrium ........................................................................ 33
Fig. 17: Exemplary property model ...................................................................................................... 33
Fig. 18: High-level classification of laws............................................................................................... 36
Fig. 19: Specialization of balance laws................................................................................................... 36
Fig. 20: Specialization of constitutive law ............................................................................................... 36
Fig. 21: Specialization of generalized flux laws ....................................................................................... 37
Fig. 22: Specialization of equilibrium constraints ..................................................................................... 37
Fig. 23: High-level classification of property models.............................................................................. 44
Fig. 24: Some phase interface transport property models........................................................................... 44
Fig. 25: Some chemical kinetics models .................................................................................................. 44
Fig. 26: Some thermodynamic property models........................................................................................ 44
Fig. 27: Some intensive thermodynamic state models................................................................................ 45
Fig. 28: Definition of the class density model ......................................................................................... 45
Fig. 29: High-level classification of process unit models ........................................................................ 50
Fig. 30: Definition of the class CSTR model .......................................................................................... 50

4
1. Mathematical Models
The partial model mathematical_model is concerned with the description of mathematical models.
Fig. 1 gives an overview of the ontology modules of mathematical_model and their interrelations.
The main module, mathematical_model, introduces the basic concepts for mathematical modeling,
including model quantities and items pertaining to sub-models and their connections; CapeML (von
Wedel, 2002) was taken as an important source. The ontology module equation_system further
specifies the characteristics of the model equations that constitute a mathematical model. Based on these
characteristics, an appropriate numerical solver can be selected, which is the concern of the ontology
module numerical_solution_strategy. The modules process_model and cost_model describe two
particular types of mathematical models: process models model the behavior of process units and
materials, while cost models predict the cost of chemical process systems.
The process_model module is extended on the Application-Oriented Layer: The ontology module laws
establishes models for a number of physical laws that are common in the context of chemical
engineering (e.g., the law of energy conservation); property_models provides correlations for
designated physical quantities, such as vapor pressure correlations or activity coefficient models; finally,
the module process_unit_model establishes customary mathematical models for process units, such as
ideal reactor models or tray-by-tray models for distillation columns.

system
Upper Layer

model

mathematical_ equation_
model system

numerical_
cost_model process_model
solution_strategy
Conceptual
Layer

process_unit_
property_models laws
model

Application-Oriented Layer

Fig. 1: Overview on partial model mathematical_model

1.1. Mathematical Model

A mathematical model is a special type of model, which uses mathematical language to describe the
behaviour of a system. A mathematical model has a number of properties, the most important of which
are model quantities. As indicated in Fig. 2, a model quantity is a subclass of physical quantity that is linked
to the model via the relation hasVariable (a specialization of hasProperty, cf. lower right corner of Fig.
2). Like any physical quantity, a model quantity has a particular physical dimension and can be either a
scalar quantity or a tensor quantity. The value of a model quantity is represented by the class model quantity
specification; each model quantity has exactly one model quantity specification.

5
 hasProperty
system

Model PhysicalQuantity Value

hasVariable

Mathematical hasVariable hasValue ModelQuantity

ModelQuantity
Model Specification

Parameter Constant StateVariable InputVariable

Fig. 2: Mathematical model, model quantity, and model quantity specification

A model quantity can be one of the following types: a constant, a parameter, a state variable or an input
variable, depending on the intended specification of its value: Constants and parameters constitute the
fixed set of specified variables. Input variables represent time of spation dependent inputs, which have
to be specified for dynamic and/or spatial distributed systems. Finally, state variables constitute the
fixed set of unknown variables which have to be computed by the model. The model quantity
specification indicates the numericalValue of the model quantities. Unlike constants, parameter and input
variables may have different model quantity specifications in different simulation runs. If the model quantity
is of type parameter or state variable, the model quantity specification may indicate their upper limit and
lower limit (cf. Fig. 4).

hasValue ModelQuantity numerical 1

ModelQuantity Value xsd:any
Specification

Parameter hasValue UpperLimit

StateVariable hasValue LowerLimit

Fig. 3: Assignment of particular model quantity specifications to model quantites

A system which is the target of a models relation is classified as a modeled object. The correspondences
between the model quantity and the physical quantity of the modeled object can be explicitly represented by
means of the relation corresponds-ToQuantity (cf. Fig. 4).

system
isPropertyOf
System PhysicalQuantity 1

Corresponds
isPropertyOf
ToQuantity
1
1..n models Mathematical hasVariable 1..n
ModeledObject ModelQuantity
Model

Fig. 4: Correspondence between a model quantity and a physical quantity of the modeled object

Like any system, a mathematical model can be decomposed into subsystems, which are called submodels.
The submodel models the same system as its superordinate mathematical model; consequently, there is no
need to specify the models relation between submodel and system explicitly. However, such a relation
may be indicated if the submodel models a designated subsystem of the overall system.

6
The different submodels of a mathematical model are coupled via their model quantities, as will be
explained in the following:
At first, the concept of a model port is introduced. A model port is a special type of property set, which
comprises model quantities that can participate in a connection with another model. Thus, a model port
has the function to identify and to bundle the ‘public’ variables of a mathematical model.
Next, the concept of a coupling is established. A coupling is a property of the overall mathematical model,
which defines a connection between two of its submodels by linking their respective model ports. The
coupling implicitly defines equality constraints between the model quantities in the two model ports and
must be treated as such (e.g., during degrees of freedom analysis of a complex model). It may be used
to connect mathematical models both ‘horizontally’ (i.e., on the same level of decomposition) and
‘vertically’ (i.e., across levels of decomposition).

determines index
xsd:
Mathematical hasVariable PositionOf Value
ModelQuantity PortIndex positive
Model 1
1..n Integer
0..n
isOrderedBy
hasPort
ModelPort system
2
PropertySet
hasCoupling
Coupling

Fig. 5: Variables, ports, couplings

The order of the model quantities within a model port can be specified by a port index, as shown in Fig. 5.
The port index is used to identify corresponding model quantities in a coupling: Two model quantities of
different model ports are coupled if and only if their port indices have the same indexValues. The
specification of a port index may be omitted if the correspondence between model quantities is evident
from the context (e.g., if each of the coupled model ports comprises only a single model quantity, or if
corresponding model quantities can be uniquely identified through their physical dimension.)

[MQ] [MQ]
a x
determines index index determines
PositionOf Value Value
[MM] hasVariable [MQ] [PF] [PF] PositionOf [MQ] hasVariable [MM]
1 1
M1 b i11 i21 y M2
determines index index determines
PositionOf Value Value PositionOf [MQ]
[MQ] [PF] [PF]
2 2
c i12 i22 z
hasPort isOrderedBy hasPort
[MP] [Coupling] [MP]
P1 C P2

hasCoupling

[MM]
M

Fig. 6: Exemplary decomposition of model M into submodels M1 and M2

Fig. 6 shows exemplarily the definition of a mathematical model M, which consist of two submodels, M1
and M2. M1 has the model quantities a, b, and c, while M2 has the model quantities x, y, and z. Model M1
owns the model port P1, which comprises the quantities b and c. Similarly, the model port P2 of model

7
M2 comprises the quantities y and z. P1 and P2 are coupled via the coupling C, which is a property of
the overall model M. The corresponding quantities of the coupling are identified via their port indices: b
and y have the same indexValue and are thus linked by an equality constraint. The same holds true for
quantities c and z.

Usage
The ontology module mathematical_model provides only the basic concepts for the description of
mathematical models. For practical applications, further concepts may be required, which would
typically be supplied by additional ontology modules located on the application-oriented layer. Some
possible extensions of the mathematical_model module are discussed below.
One possible extension would be the introduction of concepts suitable for representing the model
equations. Such an extension could be realized easily by reusing the concepts of the
mathematical_relation module. However, such an extension is not required in practice, since
specialized representation formats for mathematical equations are available, such as MathML (Ion &
Miner, 1999), CapeML (von Wedel, 2002), or CellML (Lloyd et al., 2004).
Another possible extension would be the definition of different types (i.e., subclasses) of model ports. A
particular model port type could, for example, prescribe the number of model quantities comprised in a
model port, their types (i.e., constant, parameter, or state variable), their physical dimensions, etc.
Moreover, a model port type could be further characterized through attributes (e.g., assigning a
direction to a model port, thus turning it into either an inlet port or an outlet port). That way, standardized
model interfaces can be defined – for instance, one may define a standard energy port, which contains a
single scalar model quantity with the physical dimension of an energy flow and must furthermore be
tagged as an inlet or outlet port. Such standardization facilitates checking the feasibility of a coupling: A
coupling of two mathematical models will be feasible if their model ports (a) are of the same type (e.g.,
energy port) and (b) have matching attributes (e.g., an inlet port can only be coupled to an outlet port).

[MM] [MM] [MM]

Reboiler Column Trays
Submodel Model Submodel
hasPort
hasPort hasCoupling

[MM] hasCoupling [MM]

[Part] [Coupling] [Part]
TrayModel_i TrayModel_i+1
ReboilerVapor Vapor TraysVapor
OutletPort Coupling InletPort hasPort hasPort

[Mpart] [Coupling] [MPart]

Vapor Vapor Vapor
InletPort_i Coupling_i OutletPort_i+1

Fig. 7: Specification of the overall model

In practice, a mathematical model often consists of several interconnected submodels of same the type –
for example, the model of a distillation column contains several models of distillation column trays.
An application-oriented extension of mathematical_model could apply the loop design pattern
introduced in the Meta Model (Morbach et al, 2007) to define such repetitive model structures. An
example is given in the above and below figures: Fig. 7 specifies the overall structure of a Column
Model. It consists of a Reboiler Submodel and a Trays Submodel, which are coupled via a Vapor
Coupling (to simplify matters, the liquid phase is not considered in this example). The Trays Model is
defined iteratively (see grey-shaded area in Fig. 7 and Fig. 8): It consists of several submodels of the
same type, which are represented by the individual TrayModel_i. Each TrayModel_i has a
VaporInletPort_i, which is coupled to the VaporOutletPort_i+1 of the next TrayModel_i+1. This statement
is included in a ForLoop that counts from 1 to 20, that way defining a structure of 20 interconnected
tray models. The vapor inlet port of the 20th tray model corresponds to the previously defined
TraysVaporInletPort.
8
 [MPart] [MPart]
TraysVapor Vapor
InletPort OutletPort_i

sameObject hasPort
finalStatement
[Part]
[MM]
Vapor hasPort TrayModel_i
InletPort_i

statement statement sameObject sameObject

For_i For_i

numberOf [ForLoop] statement [MM]

20 Iterations ForLoop For_iPlus1 TrayModel_i+1

statement hasPort
For_i
[MPart]
[Part]
Vapor
Vapor
OutletPort_i+1
Coupling_i

Fig. 8: Specification of the repetitive submodel structure

Concept Definitions

Class Descriptions
Constant
Description
A constant is a specified model quantity, the model quantity specification of which has a constant
numericalValue in all simulation runs.
Relations
- Constant is a subclass of model quantity.
- A constant has exactly one model quantity specification, which has exactly one numericalValue.

Input variable
Description
Input variables represent time of spation dependent inputs, which have to be specified for dynamic
and/or spatial distributed systems.
Relations
- Input variable is a subclass of model quantity.

Lower limit
Description
An lower limit is model quantity specification which defines an lower bound for the numericalValue of a
model quantity specification.
Relations
- Lower limit is a subclass of model quantity.

9
Coupling
Description
A coupling connects two model ports of different submodels, thereby defining equality constraints
between model quantities comprised in the two model ports.
Relations
- Coupling is a subclass of property set.
- A coupling comprises exactly two model ports.

Mathematical model
Description
A mathematical model is a model that uses mathematical language to describe the modeled system
reflecting the phenomena capture in the system’s behavior.
Relations
- Mathematical model is a subclass of model.
- A mathematical model may have a number of model quantities.
- A mathematical model may have a number of model ports.
- A mathematical model may have a number of couplings.
- A mathematical model may be decomposed into submodels.

Modeled object
Description
A system that is modeled by means of a model is denoted as a modeled object.
Definition
A modeled object is a system that is modeledBy a model.
Relations
- Modeled object is a subclass of system.

Model port
Description
A model port is a collection of model quantities that can participate in a connection with another
mathematical model. Thus, a model port has the function to identify and to bundle the ‘public’ variables
of a mathematical model. Optionally, a model port can be ordered by a port index.
Relations
- Model port is a subclass of property set.
- A model port comprises some model quantities.
- A model port comprises only model quantities.
- A model port can only be ordered by a port index.

Model quantity
Description
A model quantity represents a physical quantity involved in a mathematical model, the value of which can be
either supplied by the modeler or computed from an evaluation of the mathematical model.
Definition
A model quantity is either a state variable or parameter or a constant.
Relations

10
- Model quantity is a subclass of physical quantity.
- The value of a model quantity is given by its model quantity specification.

Model quantity specification

Description
A model quantity specification specifies a model quantity in terms of its numerical value (or limits of its
numeric value) and its unit of measurement.
Relations
- Model quantity specification is a subclass of value.
- A model quantity specification has exactly one unit.
- A model quantity specification either specifies the numericalValue of a model quantity or defines an
upperLimit and a lowerLimit for its numerical value.

Parameter
Description
A parameter is a specified model quantity (i.e., an input variable), the model quantity specification of which
may take different numericalValue in different simulation runs.
Relations
- Parameter is a subclass of model quantity.
- The values of a parameter are model quantity specifications which have exactly one numericalValue.

Port index
Description
A port index orders the model quantities comprised in a model port by assigning each of them an
indexValue. In a coupling, model quantities with the same indexValue are coupled to each other.
Relations
- Port index is derived from the meta class index.
- A port index is index of exactly one model port.
- A port index determines the position of exactly one model quantity.
- A port index has exactly one indexValue.

Submodel
Description
A mathematical model can be decomposed into submodels.
Definition
A submodel is a direct subsystem of a mathematical model.
Relations
- Submodel is a subclass of mathematical model.
- A submodel is a direct subsystem of some mathematical model.
- A submodel can only be a direct subsystem of a mathematical model.

State variable
Description
A state variable constitute the fixed set of unknown variables which have to be computed by the model.
Its model quantity specification either indicates the upperLimit and lowerLimit of the model quantity (before
solving the model) or its numericalValue (after solving the model).

11
Relations
- State variable is a subclass of model quantity.

Upper limit
Description
An upper limit is model quantity specification which defines an upper bound for the numericalValue of a
model quantity specification.
Relations
- Upper limit is a subclass of model quantity.

Relation Descriptions
correspondsToQuantity
Description
The relation denotes a one-to-one correspondence between a model quantity and a physical quantity of the
modeled object.
Characteristics
- Derived from the meta relation inter-objectRelation
- Domain: model quantity
- Range: A physical quantity that is a property of some modeled object
- Functional

hasCoupling
Description
The relation indicates a coupling between two submodels of a mathematical model.
Characteristics
- Specialization of hasProperty
- Domain: mathematical model
- Range: coupling

hasModelPort
Description
The relation identifies the model port of a mathematical model.
Characteristics
- Specialization of hasProperty
- Domain: mathematical model
- Range: model port

hasModelVariable
Description
The relation indicates the model quantities of a mathematical model.
Characteristics
- Specialization of hasProperty
- Domain: mathematical model
- Range: model quantity

12
determinesPositionOf
Description
The one-to-one relation between a port index and the corresponding model quantity.
Characteristics
- Derived from the meta relation determinesPositionOf
- Domain: Port index
- Range: model quantity
- Functional

isIndexOf
Description
The relation isIndexOf points form a port index to the associated model port.
Characteristics
- Derived from the meta relation isIndexOfArray
- Domain: Port index
- Range: Model port
- Inverse: isOrderedBy
- Functional

isOrderedBy
Description
The relation isOrderedBy points from a model port to its sorting port index.
Characteristics
- Derived from the meta relation isOrderedBy
- Domain: Model port
- Range: Port index
- Inverse: isIndexOf
- Inverse functional

Attribute Descriptions
indexValue
Description
The attribute indexValue indicates the numerical value of a port index.
Characteristics
- Specialization of the meta relation index
- Domain: Port index
- Datatype: positiveInteger (built-in XML Schema Datatype)
- Functional

13
1.2. Equation System
The ontology module equation_system provides concepts for the description of the model equations
that constitute a mathematical model. The model equations are not explicitly represented, only their
equation system characteristics are specified. Moreover, the scope of equation system characteristics is
confined to those characteristics that are of relevance for selecting an appropriate solver and/or
solution strategy for the mathematical model (cf. Sec. “numerical solution strategy”).

hasModelRepresentationForm
LinearAlgebraic
hasDAE_Explicitness SystemType
hasCharacteristic Algebraic
hasVariablesType Equation
hasNumericalStiffness System
NonlinearAlgebraic
SystemType

LinearityVT has
Linearity mathematical_model OrdinaryDifferential
Differential AlgebraicSystem
Mathematical Algebraic
ModelRepresentation Model Equation
Form System PartialDifferential
hasCharacteristic AlgebraicSystem
DAE_Type

EquationSystem OrdinaryDifferential
ODE_Type Characteristics EquationSystem
Differential
VariablesType Equation
System
PartialDifferential
NumericalStiffness EquationSystem

Fig. 9: Equation system characteristics

A mathematical model can be classified according to different criteria: equation system type, variables
type, model representation form, etc. Following the recommendations for ontology normalization 1 given
by Rector (2003), equation systems are classified along a single axis only, using the equation system
type as a differentiating criterion (cf. Fig. 9). The other possible criteria are explicitly modeled as
equation system characteristics, which are to be linked to a mathematical model via the relation
hasCharacteristic (or one of its specializations, cf. upper left corner of Fig. 9). Note that some of the
equation system characteristics can only be assigned to special types of equation systems; for instance,
DAE explicitness only applies to differential algebraic equation systems.
While the meaning of most concepts displayed in Fig. 9 should be evident from their names, the
concept of model representation form requires some explanation. A mathematical model may appear in
two representation forms, which are termed open-form and closed-form:
An open-form model does not provide an explicit solution for its model equations; instead, a numerical
solver needs to be applied to the model to obtain a solution. Hence, the open-form model must provide
all the information required by the external numerical algorithm to solve the model. For example, a
model representing a set of algebraic equations may provide equation residuals and derivatives to a
Newton solver. Before an open-form model can be successfully solved, it has to be “squared”, meaning
that the number of its unknown variables must be the same as that of its equations. Among all the
model quantities of an open-form model, those declared as constants or parameters have to be given

1 (1) A class should not have more than one primitive parent; (2) classification of primitive classes should be based on a single
criterion (which may remain implicit); (3) other criteria are to be indicated explicitly. More details on this issue can be found in the
informal specification of the Meta Model (Morbach et al., 2007).
14
values (i.e., they need to be assigned a model quantity specification with a definite numericalValue) before
the model can be evaluated. The other variables are state variables. If there are still more state variables
than equations in a model, it is necessary to assign values to some selected variables (i.e., turn them
into parameters or input variables). Generally, one can freely choose the set of model quantities of an
open-form model to be specified, as long as no numerical difficulties arise. The values of the
remaining variables can be obtained by solving the model.
A closed-form model includes an underlying numerical algorithm, which solves its model equations.
Thus, it does not require any external solver for obtaining the values of its unknown variables. The
‘execution’ of the closed-form model yields the values of the unknown variables (i.e., its outputs)
based on the given values of the specified variables (i.e., its inputs). In this process, the algorithm of a
closed-form model accepts only a fixed set of input variables, and consequently returns a fixed set of
output variables; no choice for specifying additional variables is available, as in the case of open-form
models. Reflected in model quantity types, constants and parameters constitute the fixed set of specified
variables (i.e. input variables), while the state variables constitute the fixed set of unknown variables
(i.e. output variables).

Usage
The ontology module equation_system provides the basic concepts for the identification of
mathematical models from a mathematical point of view, e.g. whether it is an ODE or DAE etc.. This
identification was primarily of concern in the COGents project (cf. Sect. Fehler! Verweisquelle
konnte nicht gefunden werden.) where this module was applied to specify the type of mathematical
model to search for in various libraries.
Typically mathematical models may be classified either by means of content, e.g. a mathematical
model for a polyethene reactor, or simply by mathematical features as it is done here. As an example
consider a process engineer who searches for a particular mathematical model which is supposed to be
applied for the calculations of a reactor. Depending on the software, e.g. he might have only a solver
for ODEs available, a classification with respect to the characteristics, e.g. ODE type, is extremely
helpful to identify the suitable mathematical model.

Concept Descriptions
Individual concepts of the module equation_system are defined below. For an extensive description of
the introduced individuals, we refer to Morbach et al (2008j).

Class Descriptions
Algebraic equation system
Description
An algebraic equation system is a mathematical model which solely consists of algebraic equations.
Definition
An algebraic equation system is either a linear algebraic system or a nonlinear algebraic system.
Relations
- Algebraic equation system is a subclass of mathematical model.
- An algebraic equation system has a characteristic of type linearity (i.e., it is either linear or nonlinear).

DAE explicitness
Description
Characterizes the explicitness of a differential algebraic equation system.

15
Definition
The class DAE explicitness is an exhaustive enumeration of the individuals fully_implicit and semi-
explicit.
Relations
- DAE explicitness is a subclass of equation system characteristics.

Differential algebraic equation system

Description
A differential algebraic equation system (DAE system) is a mathematical model that comprises both
algebraic and differential equations.
Definition
A differential algebraic equation system is either an ordinary differential algebraic system or a partial differential
algebraic system.
Relations
- Differential algebraic equation system is a subclass of mathematical model.
- A differential algebraic equation system has a characteristic of type linearity (i.e., it is either linear or
nonlinear).
- A differential algebraic equation system has a characteristic of type explicitness (i.e., it is either
(fully_implicit or semi-explicit).
- An differential algebraic system has exactly one differentialIndex.

Differential equation system

Description
A differential equation system is a mathematical model that solely consists of differential equations.
Definition
A differential equation system is either an ordinary differential equation system or a partial differential equation
system.
Relations
- Differential equation system is a subclass of mathematical model.
- A differential equation system is of a certain differentialOrder.

Equation system characteristics

Description
The equation system characteristics characterize the model equations of a mathematical model.
Relations
- Equation system characteristics is a subclass of fixed value set.

Linear algebraic system type

Description
A linear algebraic system type is an algebraic system which contains only linear equations.
Definition
A linear algebraic system is an algebraic system that is characterized as linear.
Relations
- Linear algebraic system is a subclass of algebraic equation system.

16
Linearity VT
Description
Linearity VT characterizes the linearity of a mathematical model.
Definition
Linearity is an exhaustive enumeration of the individuals linear and nonlinear.
Relations
- Linearity is a subclass of equation system characteristics.

Mathematical model (continued)

Relations
- A mathematical model is of a certain variables type.
- A mathematical model has a particular model representation form.

Model representation form

Description
A mathematical model may appear in two forms, as indicated by the model representation form:
- An open-form model is solved by an external algorithm. One can freely choose the inputs and
outputs of the open-form model.
- A closed-form model includes an underlying numerical algorithm that solves the model equations.
The algorithm accepts only a fixed set of input variables, and consequently returns only a fixed set
of output variables.
Definition
The class model representation form is an exhaustive enumeration of the individuals open-form and
closed-form.
Relations
- Model representation form is a subclass of equation system characteristics.

Nonlinear algebraic system type

Definition
A nonlinear algebraic system is an algebraic equation system that is characterized as nonlinear.
Relations
- Nonlinear Algebraic system is a subclass of algebraic equation system.

Numerical stiffness
Description
In mathematics, stiff equations are equations where certain implicit methods, in particular BDF,
perform better, usually tremendously better, than explicit ones (Hairer and Wanner 1996). The class
numerical stiffness is an exhaustive enumeration of the individuals stiff and nonstiff.
Definition
The class numerical stiffness is an exhaustive enumeration of the individuals stiff and nonstiff.
Relations
- Numerical stiffness is a subclass of equation system characteristics.

17
ODE_type
Description
Characterizes the explicitness of an ordinary differential equation system, which can be given in
implicit_formulation or explicit_formulation.
Definition
ODE_Explicitness is an exhaustive enumeration of the individuals implicit_formulation and
explicit_formulation.
Relations
- ODE_Explicitness is a subclass of equation system characteristics.

Ordinary differential algebraic system

Description
An ordinary differential algebraic system comprises algebraic equations as well as ordinary differential
equations, but no partial differential equations.
Relations
- Ordinary differential algebraic system is a subclass of differential algebraic equation system.

Ordinary differential equation system

Description
An ordinary differential equation system (ODE system) is a differential equation system which solely
consists of ordinary differential equations.
Relations
- Ordinary differential equation system is a subclass of differential equation system.
- An ordinary differential equation system has a characteristic of type numerical stiffness (i.e., it is either
stiff or nonstiff).

Partial differential algebraic system

Description
A partial differential algebraic system is a differential algebraic equation system which comprises both partial
differential equations and algebraic equations.
Relations
- Partial differential algebraic system is a subclass of differential algebraic equation system.

Partial differential equation system

Description
A partial differential equation system (PDE system) is a differential equation system which consists of partial
differential equations.
Relations
- Partial differential equation system is a subclass of differential equation system.

Variables type
Description
A variables type indicates whether the model quantities of a mathematical model are all continuous, all
discrete, or partly continuous and partly discrete:
18
Definition
The class variables type is an exhaustive enumeration of the individuals continuous_model,
discrete_model, and mixed_model.
Relations
- Variables type is a subclass of equation system characteristics.

Relation Descriptions
hasDAE_Explicitness
Description
Indicates an equation system characteristic of type DAE explicitness.
Characteristics
- Specialization of hasCharacteristic
- Domain: Differential algebraic equation system
- Range: DAE explicitness

hasModelRepresentationForm
Description
Indicates an equation system characteristic of type model representation form.
Characteristics
- Specialization of hasCharacteristic
- Domain: mathematical model
- Range: Model representation form

hasNumericalStiffness
Description
Indicates an equation system characteristic of type numerical stiffness.
Characteristics
- Specialization of hasCharacteristic
- Domain: Ordinary differential equation system
- Range: Numerical stiffness

hasODE_Type
Description
Indicates an equation system characteristic of type ODE explicitness.
Characteristics
- Specialization of hasCharacteristic
- Domain: Ordinary differential equation system
- Range: ODE explicitness

hasVariablesType
Description
Indicates an equation system characteristic of type variables type.
Characteristics

19
- Specialization of hasCharacteristic
- Domain: Mathematical model
- Range: Variables type

Attribute Descriptions
differentialIndex
Description
The attribute represents the differential index of a differential algebraic equation system, as defined by
Gear & Petzold (1984).
Characteristics
- Domain: Differential algebraic equation system
- Datatype: positiveInteger (built-in XML Schema Datatype)

differentialOrder
Description
The attribute differentialOrder denotes the order of a differential equation, which is defined as the order
of the highest derivative of a model quantity appearing in the differential equation.
Characteristics
- Domain: Differential equation system
- Datatype: positiveInteger (built-in XML Schema Datatype)

Individual Descriptions
closed-form
Description
A closed-form model includes an underlying numerical algorithm that solves the model equations. The
algorithm accepts only a fixed set of input variables, and consequently returns only a fixed set of
output variables.
Characteristics
- Instance of model representation form
- Different from open-form

continous_model
Description
A continuous_model denotes a mathematical model in which all the model quantities are continuous.
Characteristics
- Instance of variables type
- Different from discrete_model and mixed_model

discrete_model
Description
A discrete_model denotes a mathematical model in which all the model quantities are discrete. An
example of a discrete_model is an integer model.

20
Characteristics
- Instance of variables type
- Different from continous_model and mixed_model

explicit_formulation
Description
In an explicit_formulation, the ordinary differential algebraic system is explicitly solved for the highest-
order derivative, i.e., Fehler! Es ist nicht möglich, durch die Bearbeitung von Feldfunktionen
Objekte zu erstellen..
Characteristics
- Instance of ODE explicitness
- Different from implict_formulation

fully_implicit
Description
A differential algebraic equation system is fully_implicit if it has the form Fehler! Es ist nicht möglich,
durch die Bearbeitung von Feldfunktionen Objekte zu erstellen..
Characteristics
- Instance of DAE explicitness
- Different from semi-explicit

implict_formulation
Description
In an implicit_formulation, the ordinary differential algebraic system is not solved for the highest-order
derivative, i.e., Fehler! Es ist nicht möglich, durch die Bearbeitung von Feldfunktionen Objekte
zu erstellen..
Characteristics
- Instance of ODE explicitness
- Different from explicit_formulation

linear
Description
Characterizes a linear equation system.
Characteristics
- Instance of linearity
- Different from nonlinear

mixed_model
Description
A mixed_modeldenotes a mathematical model in which some of the model quantities are discrete while
the others are continuous. An example of a mixed_model is a mixed integer model.
Characteristics
- Instance of variables type
- Different from continous_model and discrete_model

21
nonlinear
Description
Characterizes a nonlinear equation system.
Characteristics
- Instance of linearity
- Different from linear

nonstiff
Description
Characterizes a nonstiff differential equation.
Characteristics
- Instance of numerical stiffness
- Different from stiff

open-form
Description
An open-form model is solved by an external algorithm. One can freely choose the inputs and outputs
of the open-form model.
Characteristics
- Instance of model representation form
- Different from closed-form

semi-explicit
Description
A differential algebraic equation system is semi-explicit if it has the form x& = f ( x, y ) , 0 = g ( x, y ) .
Characteristics
- Instance of DAE explicitness
- Different from fully_implicit

stiff
Description
Characterizes a stiff differential equation.
Characteristics
- Instance of numerical stiffness
- Different from nonstiff

22
1.3. Numerical Solution Strategy
In this ontology module, strategies for solving mathematical models are defined. At present, it is
confined to numerical solution strategies only. A classification of numerical solution techniques is
given, and the ability of a strategy to solve a particular type of mathematical model is explicitly
specified. The major concepts are shown in Fig. 10: A model solution strategy solves a mathematical
model; the subclasses of model solution strategy represent different types of algorithms, which are
specifically designed to solve a certain type of mathematical model with certain equation system
characteristics. To this end, a model solution strategy may apply some other, specialized model solution
strategy. So far, only numerical solution strategies have been considered in OntoCAPE, but
symbolic/analytical solution methods could be added in an analogous manner.

applies
mathematical_model
solves
MathematicalModel ModelSolutionStrategy
0..n

OrdinaryDifferential solves ODE_

EquationSystem SolutionStrategy

applies

DifferentialAlgebraic solves DAE_

EquationSystem SolutionStrategy
applies

Algebraic solves AlgebraicModel

EquationSystem SolutionStrategy

PartialDifferential
PartialDifferential solves
AlgebraicModel
AlgebraicSystem
SolutionStrategy

Fig. 10: Numerical solution strategy.

Fig. 11 shows the refinement of class algebraic model solution strategy. An exemplary linear algebraic
model solution strategy is Gauss-elimination, an example of a nonlinear algebraic model solution strategy is
Newton’s method.

mathematical_model

Algebraic solves AlgebraicModel

EquationSystem SolutionStrategy

Linear solves LinearAlgebraic

AlgebraicSystem ModelSolutionStrategy

Nonlinear solves NonlinearAlgebraic

AlgebraicSystem ModelSolutionStrategy

Fig. 11: Types of algebraic model solution strategies

23
An ODE solution strategy can be further characterized by indicating if the algorithm is one-step_method
(e.g., the classical Runge-Kutta methods) or a multi-step_method (e.g., the Adams-Bashforth methods).
Moreover, it can be specified whether the algorithm is a solution strategy for explicit ODEs or implicit ODEs
(cf. Fig. 12).

ODE_ hasTypeOfInvolvedSteps
TypeOfInvolvedSteps
SolutionStrategy

one-step_method
SolutionStrategyFor SolutionStrategyFor
ExplicitODEs ImplicitODEs multistep_method

Fig. 12: Further specification of ODE solution strategy

Concept Descriptions
Individual concepts of the module numerical_solution_method are defined below.

Class Descriptions
Algebraic model solution strategy
Description
An algebraic model solution strategy is a model solution strategy for solving algebraic equation systems.
Relations
- Algebraic model solution strategy is a subclass of model solution strategy.
- An algebraic model solution solves only algebraic equation systems.

DAE solution strategy

Description
A DAE solution strategy is a model solution strategy for solving differential algebraic equation systems.
Examples are implicit Runge-Kutta, BDF, etc.
Relations
- DAE solution strategy is a subclass of model solution strategy.
- A DAE solution strategy applies an algebraic model solution strategy (for initialization, solving
corrector equation, etc.).
- A DAE solution strategy applies an ODE solution strategy (for solving differential equations).
- A DAE solution strategy solves only differential algebraic equation system.
- A DAE solution strategy can only solve differential algebraic equation systems up to a certain
differentialIndex. This restriction is specified through the attribute handlesDifferentialIndexUpTo.

Linear algebraic model solution strategy

Description
A linear algebraic model solution strategy is a model solution strategy for solving linear algebraic systems. An
example is Gauss-elimination.
Relations
- Linear algebraic model solution strategy is a subclass of algebraic model solution strategy.
- A linear algebraic model solution strategy solves only linear algebraic systems.

24
Model solution strategy
Description
A model solution strategy is a (typically numerical) algorithm that can be used to solve mathematical
models.
Relations
- Model solution strategy is a subclass of system.
- A model solution strategy solves some mathematical models.
- A model solution strategy can only solve some mathematical models.
- A model solution strategy may apply some other model solution strategy.

Nonlinear algebraic model solution strategy

Description
A nonlinear algebraic model solution strategy is a model solution strategy for solving nonlinear algebraic
systems. An example is Newton’s method.
Relations
- Nonlinear algebraic model solution strategy is a subclass of algebraic model solution strategy.
- A non-linear algebraic model solution strategy solves only nonlinear algebraic systems.

ODE solution strategy

Description
An ODE solution strategy is a model solution strategy for solving ordinary differential equation systems.
Relations
- ODE solution strategy is a subclass of model solution strategy.
- An ODE solution strategy has exactly one type of involved steps.
- An ODE solution strategy solves only ordinary differential equation systems.

Partial differential algebraic model solution strategy

Description
A partial differential algebraic model solution strategy is a model solution strategy for solving partial differential
algebraic systems.
Relations
- Partial differential algebraic model solution strategy is a subclass of model solution strategy.
- A partial differential algebraic model solution strategy solves only partial differential algebraic systems.

Solution strategy for explicit ODEs

Description
A solution strategy for explicit ODEs is used to solve ordinary differential equation systems that are given in
an explicit_formulation. Examples are explicit Euler, explicit Runge-Kutta, etc.
Relations
- Solution strategy for explicit ODEs is a subclass of ODE solution strategy.
- A solution strategy for explicit ODEs solves only ordinary differential equation systems that have an ODE
explicitness of type explicit_formulation.

25
Solution strategy for implicit ODEs
Description
A solution strategy for implicit ODEs is used to solve ordinary differential equation systems that are given in
an implicit_formulation. Examples are implicit Euler, implicit Runge-Kutta, etc.
Relations
- Solution strategy for implicit ODEs is a subclass of ODE solution strategy.
- A solution strategy for implicit ODEs solves only ordinary differential equation systems that have an ODE
explicitness of type implicit_formulation.

Type of involved steps

Description
A type of involved step denotes whether an ODE solution strategy is a one-step_method or a multi-
step_method.
- A one-step_method characterizes an ODE solution strategy that uses information of one integration
step. Examples are various Runge-Kutta methods.
- A multi-step_method characterizes an ODE solution strategy that uses information of multiple
integration steps. Examples are Adams, BDF, etc.
Definition
Exhaustive enumeration of the individuals one-step_method and multi-step method.
Relations
- Type of involved step is a subclass of fixed value set.

Relation Descriptions
applies
Description
A model solution strategy may apply some other, specialized model solution strategy (e.g., for
initialization, solving corrector equation, solution of a subproblem, etc.).
Characteristics
- Specialization of containsDirectly
- Domain: model solution strategy
- Range: model solution strategy

hasTypeOfInvolvedSteps
Description
Indicates the type of involved steps of an ODE solution strategy.
Characteristics
- Specialization of hasCharacteristics
- Domain: ODE solution strategy
- Range: type of involved steps
- Functional

26
solves
Description
The relation indicates the type of mathematical model, for the solution of which a particular model
solution strategy is designated.
Characteristics
- Specialization of isDirectlyRelatedTo
- Domain: model solution strategy
- Range: mathematical model

Attribute Descriptions
handlesDifferentialIndexUpTo
Description
A DAE solution strategy can only solve differential algebraic equation systems up to a certain
differentialIndex. This restriction is specified through the attribute handlesDifferentialIndexUpTo.
Characteristics
- Domain: DAE solution strategy
- Datatype: positiveInteger (built-in XML Schema Datatype)

Individual Descriptions
multi-step_method
Description
A multi-step_method characterizes an ODE solution strategy that uses information of multiple integration
steps. Examples are Adams, BDF, etc.
Characteristics
- Instance of type of involved step
- Different from one-step_method

one-step_method
Description
A one-step_method characterizes an ODE solution strategy that uses information of one integration step.
Examples are various Runge-Kutta methods.
Characteristics
- Instance of type of involved step
- Different from mulit-step_method

27
1.4. Cost Model
The ontology module cost_model establishes some cost models for predicting the (investment) costs of
chemical plants. A cost model is a special type an economic performance model, which models the
economic performance of a chemical process system.

mathematical_model
MathematicalModel
CPS_performance

models
EconomicPerformance EconomicPerformanceModel

hasProperty
correspondsToQuantity
Costs CostModel

correspondsToQuantity FixedCapitalInvestment
FixedCapitalInvestment
Model

CapacityFCIModel DetailedItemFCIModel FactorialFCIModel

TurnoverRatio PowerFactor StepCounting GlobalFactorial DifferentialFactorial

Model Model Model Model Model

SixTenthsRuleModel UnitCostEstimateModel

Fig. 13: Models for estimating the fixed capital investment

At present, the module merely holds a number of models for the estimation of the fixed capital
investment; in the future, further types of cost models are to be added, and the existing ones are to be
specified in detail. Fig. 13 gives an overview on the cost models defined so far. For an explanation of
the individual classes, we refer to the concept definitions below.

Concept Descriptions
Individual concepts of the module cost_model are defined below.

Class Descriptions
Capacity FCI model
Description
Capacity FCI models are based on fixed capital investments of past design projects that are similar to the
current chemical process system. Besides, some relating factors (e.g., the turn-over ratio), exponential
power ratios or more complex relations are given.
Relations
- Capacity FCI model is a subclass of fixed capital investment model.

28
Cost model
Description
A mathematical model to estimate the investment costs of a chemical process system.
Definition
A cost model is an economic performance model that has a model quantity which corresponds to the
quantity of costs.
Relations
- Cost model subclass of economic performance model.
- A cost model has a model quantity that corresponds to the quantity of costs.

Detailed-item FCI model

Description
A detailed-item FCI model requires careful determination of all individual direct and indirect cost items.
For such models, extensive data and large amounts of engineering time are necessary. Therefore, this
type of estimate is almost exclusively prepared by contractors bidding on complete and all-inclusive
work from finished drawings and specifications.
Relations
- Detailed-item FCI model is a subclass of fixed capital investment model.

Differential factorial model

Description
Within differential factorial models, different factors are used for estimating the costs of the fixed capital
investment. Examples are modular estimate models, where individual modules consisting of a group of
similar items are considered separately, and their costs are then summarized (Guthrie, 1969).
Relations
- Differential factorial model is a subclass of factorial FCI model.

Economic performance model

Description
An economic performance model models the economic performance of a chemical process system.
Definition
An economic performance model is a mathematical model that models some economic performance.
Relations
- Economic performance model is a subclass of mathematical model.
- An economic performance model models only economic performance.

Factorial FCI model

Description
Factorial FCI models rely on the fact that the percentages of the different costs within the fixed capital
investment are similar for different chemical process systems. Based on one or several known costs (for
example the equipment costs), the fixed capital investment is estimated using some factors that are derived
from cost records, published data, and experience.
Relations
- Factorial FCI model is a subclass of fixed capital investment model.
29
Fixed capital investment model
Description
Fixed capital investment models (FCI models) are mathematical models that are used to estimate the fixed
capital investment of a chemical process system.
Definition
A fixed capital investment model is a cost model which has a model quantity that corresponds to the quantity
of fixed capital investment.
Relations
- Fixed capital investment model is a subclass of cost model.
- A fixed capital investment model has model quantity that corresponds to the quantity of fixed capital
investment.

Global factorial model

Description
A global factorial model estimates the fixed capital investment by multiplying the basic equipment cost by
some factor. This factor depends, among other things, on the type of chemical process involved,
required materials of construction, and the location of the chemical process system realization. Examples
for global factors are the ones proposed by (Lang, 1947). This model can be extended to calculate the
total capital investment.
Relations
- Global factorial model is a subclass of factorial FCI model.

Power factor model

Description
The power factor model relates the fixed capital investment of a new chemical process system to the one of
similar, previously constructed systems by an exponential power ratio (cf. Peters & Timmerhaus,
1991).
Relations
- Power factor model is a subclass of capacity FCI model.

Six-tenths rule model

Description
The six-tenths rule model is a power factor model with x=0.6.
Relations
- Six-tenths rule model is a subclass of power factor model.

Step counting model

Description
Step counting models are based on the assumption that the fixed capital investment can be estimated from
the number of process steps (depending on the specific approach, composite process steps or unit
operations and reactions are used), multiplied with the costs per process step and some correcting
factors. The costs of the process steps are estimated from their capacity and some other factors (Vogt,
1996).
Relations
- Step counting model is a subclass of capacity FCI model.

30
Turnover ratio model
Description
The turnover ratio model is a fast evaluation method for order-of-magnitude estimates. The turnover
ratio is defined as the ratio of gross annual sales to fixed capital investment. Values of turnover ratios for
different types of chemical processes are for example given by Schembra (1991) and Vogt (1996).
Relations
- Turnover ratio model is a subclass of capacity FCI model.

Unit-cost estimate model

Description
Unit-cost estimate models are based on detailed estimates of the main purchase costs for system realization
(either obtained from quotations or from cost records and published data).
Relations
- Unit-cost estimate model is a subclass of differential factorial model.

31
1.5. Process model
As an extension to mathematical_model, the ontology module process_model enables the definition of
specialized mathematical models for the domain of chemical engineering. Such models, which model
either process units or materials or subsystems of these, are called process models (cf. Fig. 14). The
modeling principle based on which a process model is developed may also be indicated.

mathematical_model
MathematicalModel ModelQuantity

hasVariable correspondsToQuantity

system
Modeling 1 hasModelingPrinciple Process
Principle Model PhysicalQuantity
models

isPropertyOf
hybrid data_driven

first-principles Material ProcessUnit

material chemical_process_system

Fig. 14: Overview on process_model

A process model may contain other process models, particularly established laws and property models (cf.
Fig. 15). Neither laws nor property models are self-contained models, but form part of an overall process
model, where they represent mathematical correlation between designated model quantities.

1..n 1..n
ProcessModel

behavior 0..n 0..n

Law PropertyModel
Physicochemical isAssociatedWith
Phenomenon 1

Fig. 15: Laws and property models

A law constitutes the mathematical representation of a scientific law, such as the law of energy
conservation. Each law can be associated with a physicochemical phenomenon. The former gives a
quantitative, the latter a qualitative description of a certain physical behavior. The correspondence
between a law and a physicochemical phenomenon can be stated via the relation isAssociatedWith, as
indicated in Fig. 15. Moreover, the model quantities of the law correspond to the physical quantities that
are influenced by the physicochemical phenomenon, as exemplarily shown in Fig. 16.

32
 isInfluencedBy
behavior
[PhysicochemicalPhenomenon]
ThermalEquilibrium
mathematical_model
isAssociatedWith
[ModelQuantity] correspondsTo
T2 Quantity
[ThermalEquilibriumLaw]
hasVariable
myThermalEquilibriumLaw
[ModelQuantity] correspondsTo
T1 Quantity
models

[SinglePhase] hasProperty [Temperature]

L1 T_L1

isDirectlyConnectedTo

[MultiphaseSystem] [PhaseInterface]
L1+L2 PhaseBoundary_L1/L2

isDirectlyConnectedTo

[SinglePhase] hasProperty [Temperature]

phase_system L2 T_L2

Fig. 16: Exemplary law modeling thermal equilibrium

A property model represents a mathematical correlation for the computation of one designated model
quantity, which corresponds to one specific physical quantity. An example is given in Fig. 17: An activity
coefficient model constitutes a correlation for the computation of activity coefficients; consequently, an
activity coefficient model has one model quantity which corresponds to an activity coefficient.

PropertyModel

ActivityCoefficient hasVariable corresponds

ModelQuantity ToQuantity ActivityCoefficient
Model
property_model mathematical_model phase_system

Fig. 17: Exemplary property model

Concept Descriptions
Individual concepts of the module process_model are defined below. For an description of the
instances of modeling principle, we refer to Morbach et al (2008j).

Class Descriptions
Law
Description
A law constitutes the mathematical representation of a scientific law. It usually forms part of an overall
process model.

33
Relations
- Law is a subclass of process model.
- A law is associated with exactly one physicochemical phenomenon.
- A law can only be a direct subsystem of a process model.

Modeling principle
Description
A modeling principle represents the principle following which the process model is developed.
- Following the data_driven modeling principle, a process model is derived from the values of the
properties of a modeled object. Examples of this type of models are neural network models.
- Following the first-principles modeling principle, the process model is based on established physical
laws and mechanisms.
- A hybrid modeling principle applies both the first-principles and the data_driven approach.
Definition
Modeling principle is defined by an exhaustive enumeration of the individuals data_driven, first-
principles, and hybrid.
Relations
- Modeling principle is a subclass of fixed value set.

Process Model
Description
A process model is a mathematical model that models a process unit or material (or subsystems of these).
Relations
- Process model is a subclass of mathematical model.
- A process model may have only process models as direct subsystems.
- A process model is built following a certain modeling principle.
- A process model models a process unit or a material, or a subsystem of these.
- Some model quantities of a process model correspond to some physical quantities of a process unit
or a material or a subsystem of these.

Property model
Description
A property model forms part of an overall process model. It represents a mathematical correlation for the
computation of one designated model quantity, which corresponds to one specific physical quantity.
Examples are vapor pressure correlations or activity coefficient models.
Relations
- A property model is a subclass of a process model.
- A property model is a direct subsystem of a process model.
- A property model can only be a direct subsystem of a process model.

Relation Descriptions
hasModelingPrinciple
Description
Indicates the modeling principle on which a process model is based.

34
Characteristics
- Specialization of hasCharacteristic
- Domain: process model
- Range: modeling principle

isAssociatedWith
Description
The relation denotes a correspondence between a law and a physicochemical phenomenon. The former
gives a quantitative, the latter a qualitative description of a certain physical behavior.
Characteristics
- Specialization of the meta relation inter-objectRelation
- Domain: law
- Range: physicochemical phenomenon

Individual Descriptions
data_driven
Description
Following the data_driven modeling principle, a process model is derived from the values of the properties
of a modeled object. Examples of this type of models are neural network models.
Characteristics
- Instance of modeling principle
- Different from first-principles and hybrid

first-principles
Description
Following the first-principles modeling principle, the process model is based on established physical laws
and mechanisms.
Characteristics
- Instance of modeling principle
- Different from data_driven and hybrid

hybrid
Description
A hybrid modeling principle applies both the first-principles and the data_driven approach.
Characteristics
- Instance of modeling principle
- Different from first-principles and hybrid

35
1.6. Laws
The ontology module laws, located on the Application-Oriented Layer of OntoCAPE, introduces a
hierarchical collection of laws that are frequently used in process modeling. The law hierarchy shown
was originally presented by Marquardt (1995). A selection of the taxonomy related to physicochemical
laws is given in Fig. 18: High-level classification of lawsFig. 18 - Fig. 22. The high-level concepts
include balance laws, constitutive laws, and constraints as shown in Fig. 18.

Law

Constraint BalanceLaw ConstitutiveLaw

Fig. 18: High-level classification of laws

Balance laws generally represent the change of an extensive quantity in process models. This typically
include balances for total mass and mass of species in a mixture (mass balance law), for momentum
(momentum balance law), for total or any other kind of energy (energy balance law), and for the particle
number in case of a particulate system (population balance laws) as depicted in Fig. 19.

BalanceLaw

Mass Energy Momentum Population

BalanceLaw BalanceLaw BalanceLaw BalanceLaw

Fig. 19: Specialization of balance laws

However, balance equations do not suffice to describe the behavior related to a process model. Thus,
constitutive laws have to be added in order to determine the process model completely. Three types of
constitutive laws may be distinguished (cf. Fig. 20), including generalized flux laws, phenomenological
coefficient law, and thermodynamic state function law. Generalized flux laws describe the contribution to any
kind of balance law. These laws typically constitute of phenomenological coefficient and a driving
force determined by some thermodynamic state function which are modeled by phenomenological
coefficient laws and thermodynamic state function laws.

ConstitutiveLaw

GeneralizedFluxLaw Phenomenological ThermodynamicState

CoefficientLaw FunctionLaw

Fig. 20: Specialization of constitutive law

In Fig. 21 the specializations of generalized flux law are presented including some further specialization
associated to transport and exchange phenomena. These specific laws have to be considered before a
concrete process model can be generated.

36
 GeneralizedFluxLaw

TransportLaw AccumulationLaw SourceLaw ExchangeLaw

IntraphaseHeat HeatRadiation
ConductionLaw Law

SurfaceMass
ConvectiveMaterial DiffusionLaw
FlowLaw
InterfaceHeat
ConductionLaw
IntraphaseMass
DiffusionLaw InterfaceMass
DiffusionLaw

IntraphaseViscous InterfaceViscous
MomentumTransportLaw MomentumTransportLaw

Fig. 21: Specialization of generalized flux laws

Finally, constraints describe all kinds of (algebraic) relations between process quantities which – literally
or by assumption – have to hold at any time. Typical examples are volume constraints or equilibrium
constraints, which specialize the class constraint in Fig. 22.

EquilibriumConstraint

Thermal Chemical Mechanical ChemicalReaction

Equilibrium Equilibrium Equilibrium Equilibrium
Constraint Constraint Constraint Constraint

Phase
Equilibrium
Constraint

Fig. 22: Specialization of equilibrium constraints

In Fig. 22 equilibrium constraints are specialized into thermal equilibrium, chemical equilibrium, and
mechanical equilibrium on the one hand, which refer to equal temperature, pressure or chemical potential
in adjacent phases. Phase equilibrium and chemical reaction equilibrium are considered on the other hand.
It shows that phase equilibrium is just an aggregation of thermal, chemical, and mechanical equilibrium.
Chemical reaction equilibrium refers to a network of chemical reactions residing in a phase where all
forward reaction rates equal the backward reaction rates. The constraint phase equilibrium can also be
formulated in various alternative ways which are fully equivalent 2 .
Currently, only the hierarchy of laws and the associated physicochemical phenomenon are modeled. In
future extensions of this ontology module, one may add further definitions and constraints in order to
specify a law’s model quantities and their corresponding physical quantities.

Concept Descriptions
Adsorption equilibrium law
Description
A law that is associated with the phenomenon of equilibrium_adsorption.

2
For example, a number of alternative formulations exist for chemical equilibrium, such as the equality of chemical potentials
between two phases and the equality of fugacities between two phases. The equality of fugacities can further be written in different
forms depending on what property models are to be used in conjunction with the law for the chemical equilibrium.

37
Relations
- Adsorption equilibrium law is a subclass of equilibrium law.
- An adsorption equilibrium law is associated with the phenomenon of equilibrium_adsorption.

Adsorption kinetics law

Description
A law that is associated with the phenomenon of non-equilibrium_adsorption.
Relations
- Adsorption kinetics law is a subclass of non-equilibrium law.
- An adsorption kinetics law is associated with the phenomenon of non-equilibrium_adsorption.

Agglomeration law
Description
A law that is associated with the phenomenon of agglomeration (of particles).
Relations
- Agglomeration law is a subclass of particle kinetics law.
- An agglomeration law is associated with the phenomenon of agglomeration.

Breakage law
Description
A law that is associated with the phenomenon of breakage (of particles).
Relations
- Breakage law is a subclass of particle kinetics law.
- A breakage law is associated with the phenomenon of breakage.

Chemical kinetics law

Description
A law that is associated with some kinetic phenomenon, such as non-equilibrium reaction or
adsorption.
Relations
- Chemical kinetics law is a subclass of non equilibrium law.

Chemical reaction equilibrium law

Description
A law that is associated with the phenomenon of chemical_reaction_equilibrium.

Relations
- Chemical reactions equilibrium law is a subclass of equilibrium law.
- A chemical reactions equilibrium law is associated with the phenomenon of chemical_reaction_
equilibrium.

Balance law
Description
A law that is associated with an accumulation phenomenon.

38
Relations
- Conversation law is a subclass of law.
- A conversation law is associated with an accumulation phenomenon.
- A conversation law can only be associated with an accumulation phenomenon.

Constitutive law
Description
Constitutive laws have to be added to balance laws in order to determine the behavior process model
completely.
Relations
- Constitutive law is a subclass of law.

Constraint
Description
Constraints describe all kinds of (algebraic) relations between process quantities which – literally or by
assumption – have to hold at any time.
Relations
- Constraint is a subclass of law.

Convective material flow law

Description
A law that is associated with the phenomenon of convective_material_flow.
Relations
- Convective material flow law is a subclass of inter-phases transport phenomenon law.
- A convective material flow law is associated with the phenomenon of convective_material_flow.

Energy balance law

Description
A law that is associated with the phenomenon of energy_accumulation.
Relations
- Energy conservation law is a subclass of conversation law.
- An energy conservation law is associated with the phenomenon of energy_accumulation.

Equilibrium constraint
Description
A law that is associated with a physical equilibrium phenomenon.
Relations
- Equilibrium constraint is a subclass of law.
- An equilibrium constraint is associated with a phase equilibrium phenomenon.
- An equilibrium constraint can only be associated with a phase equilibrium phenomenon.

Generalized flux law

Description
Generalized flux laws describe the contribution to any kind of balance law.

39
Relations
- Generalized flux law is a subclass of constitutive law.

Growth law
Description
A law that is associated with the phenomenon of growth (of particles).
Relations
- Growth law is a subclass of particle kinetics law.
- A growth law can only be associated with the phenomenon of growth.

Heat radiation law

Description
A law that is associated with the phenomenon of heat_radiation.
Relations
- Heat radiation law is a subclass of inter-phases transport phenomenon law.
- A heat radiation law can only be associated with the phenomenon of heat_radiation.

Inter-phases transport phenomenon law

Description
A law that is associated with a cross-phase transport phenomenon (i.e., describing the exchange of
mass, momentum, or energy between phases and/or across a phase interface).
Relations
- Inter-phases transport phenomenon law is a subclass of transport phenomenon law.
- An inter-phases transport phenomenon law is associated with one of the following phenomena:
interface molecular transport phenomenon or heat_radiation or convective_material_flow or surface_
mass_diffusion.

Interface heat conduction law

Description
A law that is associated with the phenomenon of interface_heat_conduction.
Relations
- Interface heat conduction law is a subclass of inter-phases transport phenomenon law.
- An interface heat conduction law is associated with the phenomenon of interface_heat_conduction.

Interface mass diffusion law

Description
A law that is associated with the phenomenon of interface_mass_diffusion.
Relations
- Interface mass diffusion law is a subclass of inter-phases transport phenomenon law.
- An interface mass diffusion law is associated with the phenomenon of interface_mass_diffusion.

Interface viscous momentum transport law

Description
A law that is associated with the phenomenon of phase_interface_viscous_momentum_transport.

40
Relations
- Interface viscous momentum transport law is a subclass of inter-phases transport phenomenon law.
- An inteface viscous momentum transport law is associated with the phenomenon of phase_interface_
viscous_momentum_transport.

Intraphase heat conduction law

Description
A law that models the phenomenon of heat_conduction within a single phase.
Relations
- Intraphase heat conduction law is a subclass of intraphase transport phenomenon law.
- An intraphase heat conduction law is associated with the phenomenon of heat_conduction.

Intraphase mass diffusion law

Description
A law that is associated with the phenomenon of mass_diffusion within a single phase.
Relations
- Intraphase mass diffusion law is a subclass of intra phase transport phenomenon law.
- An intraphase mass diffusion law is associated with the phenomenon of mass_diffusion.

Intraphase viscous momentum transport law

Description
A law that is associated with the phenomenon of viscous_momentum_transport within a single phase.
Relations
- Intraphase viscous momentum transport law is a subclass of intraphase transport phenomenon law.
- An intraphase viscous momentum transport law is associated with the phenomenon of viscous_
momentum_transport.

Intraphase transport phenomenon law

Description
A law that models some type of molecular transport within a single phase.
Relations
- Intraphase transport phenomenon law is a subclass of transport phenomenon law.
- An intraphase transport phenomenon law is associated with a molecular transport phenomenon.
- An intraphase transport phenomenon law can only be associated with a molecular transport
phenomenon.

Mechanical equilibrium law

Description
A law that is associated with the phenomenon of mechanical_equilibrium.
Relations
- Mechanical equilibrium law is a subclass of equilibrium law.
- A mechanical equilibrium law is associated with the phenomenon of mechanical_equilibrium.

41
Mass balance law
Description
A law that is associated with the phenomenon of mass_accumulation within a material amount.
Relations
- Mass conservation law is a subclass of conservation law.
- A mass conservation law is associated with the phenomenon of mass_accumulation.

Momentum balance law

Description
A law that is associated with the phenomenon of momentum_accumulation within a material amount.
Relations
- Momentum conservation law is a subclass of conservation law.
- A momentum conservation law is associated with the phenomenon of momentum_accumulation.

Nucleation law
Description
A law that is associated with the phenomenon of nucleation (of particles).
Relations
- Nucleation law is a subclass of particle kinetics law.
- A nucleation law is associated with the phenomenon of nucleation.
- A nucleation law can only be associated with the phenomenon of nucleation.

Particle kinetics law

Description
A law that is associated with a particle phenomenon.
Relations
- Particle kinetics law is a subclass of non-equilibrium law.
- A particle kinetics law is associated with some particle phenomenon.
- A particle kinetics law can only be associated with a particle phenomenon.

Phase equilibrium law

Description
A law that is associated with the phenomenon of phase_equilibrium.
Relations
- Phase equilibrium law is a subclass of equilibrium law.
- A phase equilibrium law is associated with the phenomenon of phase_equilibrium.

Phenomenological coefficient law

Description
This laws typically constitute of phenomenological coefficient which are modeled by phenomenological
coefficient laws.
Relations
- Phenomenological coefficient law is a subclass of constitutive law.

42
Population conservation law
Description
A law that is associated with the phenomenon of particle_population_accumulation.
Relations
- Population conversation law is a subclass of conservation law.
- A population conversation law is associated with the phenomenon of particle_population_
accumulation.

Reaction kinetics law

Description
A law that is associated with some of reaction kinetics phenomenon.
Relations
- Reaction kinetics law is a subclass of chemical kinetics law.
- A reaction kinetics law is associated with the phenomenon of non-equilibrium_chemical_reaction or
non-equilibrium_surface_reaction.

Surface mass diffusion law

Description
A law that is associated with the phenomenon of surface_mass_diffusion.
Relations
- Surface mass diffusion law is a subclass of inter-phases transport phenomenon law.
- A surface mass diffusion law is associated with the phenomenon of surface_mass_diffusion.

Thermodynamic state function law

Description
This law typically constitutes of a driving force determined by some thermodynamic state function.
Relations
- Thermodynamic state function law is a subclass of constitutive law.

Thermal equilibrium law

Description
A law that is associated with the phenomenon of thermal_equilibrium.
Relations
- Thermal equilibrium law is a subclass of equilibrium law.
- A thermal equilibrium law is associated with the phenomenon of thermal_equilibrium.

Transport phenomenon law

Description
A law that is associated with a phenomenon of inter- or intra-phase transport.
Relations
- Transport phenomenon law is a subclass of non-equilibrium law.

43
1.7. Property models
The ontology module property_models, which is located on the Application-Oriented Layer of
OntoCAPE, provides a hierarchically ordered collection of frequently used property models. As
indicated in Fig. 23, a property model might be one of the following:
- A chemical kinetics property model, which specifies how to calculate the rate coefficient of a
homogenous or heterogeneous reaction.
- A phase interface transport property model, which provides a correlation for computing certain phase
interface transport properties.
- A thermodynamic property model, which indicates the correlation between certain intensive
thermodynamics state variables and intra-phase transport properties

PropertyModel

ChemicalKinetics PhaseInterfaceTransport Thermodynamic

Model PropertyModel PropertyModel

Fig. 23: High-level classification of property models

The classification of these specialized property models is given in Fig. 24 to Fig. 27:

PhaseInterfaceTransport
PropertyModel

HeatTransfer MassTransfer
CoefficientModel CoefficientModel

Fig. 24: Some phase interface transport property models

ChemicalKinetics
PropertyModel

AdsorptionRate HomogeneousReaction HeterogeneousReaction

CoefficientModel RateCoefficientModel RateCoefficientModel

Fig. 25: Some chemical kinetics models

ThermoModel

SurfaceTension IntraPhaseTransport IntensiveThermodynamic

Model PropertyModel StateModel

ViscosityModel MassDiffusion ThermalConductivity

CoefficientModel Model

Fig. 26: Some thermodynamic property models

44
 IntensiveThermodynamicStateModel

DensityModel HeatCapacityModel

PartialMolarEnthalpyModel PartialMolarVolumeModel

PhaseEquilibriumRatioModel VaporPressureModel

FugacityCoefficientModel SpecificEnthalpyModel

ReactionEquilibriumConstantModel ActivityCoefficientModel

AdsorptionEquilibriumConstantModel SpecificGibbsFreeEnergyModel

Fig. 27: Some intensive thermodynamic state models

Exemplarily, the definition of the class density model is shown in Fig. 28: A density model has some
model quantities, one of which corresponds to a physical quantity of type density. The other property models
are defined analogously.

mathematical_model
hasVariable
DensityModel ModelQuantity

correspondsToQuantity

Density

phase_system

Fig. 28: Definition of the class density model

Concept Descriptions
Activity coefficient model
Description
A property model that provides a correlation for an activity coefficient.
Relations
- Activity coefficient model is a subclass of intensive thermodynamic state model.
- An activity coefficient model has a model quantity which corresponds to an activity coefficient.

Adsorption equilibrium constant model

Description
A property model that provides a correlation for an adsorption equilibrium constant.
Relations

45
- Adsorption Equilibrium Constant model is a subclass of intensive thermodynamic state model.

Adsorption rate coefficient model

Description
A property model that provides a correlation for an adsorption rate coefficient.
Relations
- Adsorption rate coefficient model is a subclass of chemical kinetics property model.

Chemical kinetics property model

Description
A property model that provides a correlation for a chemical kinetics coefficient.
Relations
- Chemical kinetics property model is a subclass of property model.

Density model
Description
A property model that provides a correlation for the computation of density.
Relations
- Density model is a subclass of intensive thermodynamic state model.
- A density model has a model quantity which corresponds to a density.

Fugacity coefficient model

Description
A property model that provides a correlation a fugacity coefficient.
Relations
- Fugacity coefficient model is a subclass of intensive thermodynamic state model.
- A fugacity coefficient model has a model quantity which corresponds to a fugacity coefficient.

Specific Gibbs free energy model

Description
A property model that provides a correlation for the computation of the specific Gibbs free energy.
Relations
- Specific Gibbs free energy model is a subclass of intensive thermodynamic state model.
- A specific Gibbs free energy model has a model quantity which corresponds to a specific Gibbs free
energy.

Heat capacity model

Description
A property model that provides a correlation for the computation of the heat capacity.
Relations
- Heat capacity model is a subclass of intensive thermodynamic state model.

46
Heat transfer coefficient model
Description
A property model that provides a correlation for a heat transfer coefficient.
Relations
- Heat transfer coefficient model is a subclass of phase interface transport property model.

Heterogeneous reaction rate coefficient model

Description
A property model that provides a correlation for a surface reaction rate coefficient.
Relations
- Heterogeneous reaction rate coefficient model is a subclass of chemical kinetics property model.

Homogeneous reaction rate coefficient model

Description
A property model that provides a correlation for a chemical reaction rate coefficient.
Relations
- Homogeneous reaction rate coefficient model is a subclass of chemical kinetics property model.

Intensive thermodynamic State Model

Description
A property model that provides a correlation for an intensive thermodynamic state variable.
Relations
- Intensive thermodynamic state model is a subclass of thermo model.
- An intensive thermodynamic state model has a model quantity which corresponds to an thermodynamic
state variable.

Intra-phase transport property model

Description
A property model that provides a correlation for an intra-phase transport coefficient.
Relations
- Intra phase transport property model is a subclass of thermo model.
- An intra phase transport property model has a model quantity which corresponds to a intra-phase
transport coefficient.

Mass diffusion coefficient model

Description
A property model that provides a correlation for an intra-phase mass diffusion coefficient.
Relations
- Mass diffusion coefficient model is a subclass of intra phase transport property model.
- A mass diffusion coefficient model has a model quantity which corresponds to a diffusion coefficient.

Mass transfer coefficient model

Description
A property model that provides a correlation for a phase interface mass transfer coefficient.
47
Relations
- Mass transfer coefficient model is a subclass of phase interface transport property model.

Partial molar enthalpy model

Description
A property model that provides a correlation for the computation of the partial molar enthalpy.
Relations
- Partial molar enthalpy model is a subclass of intensive thermodynamic state model.
- A partial molar enthalpy model has a model quantity which corresponds to a partial molar enthalpy.

Partial molar volume model

Description
A property model that provides a correlation for the computation of the partial molar volume.
Relations
- Partial molar volume model is a subclass of intensive thermodynamic state model.
- A partial molar volume model has a model quantity which corresponds to a partial molar volume.

Phase equilibrium ratio model

Description
A property model that provides a correlation for a phase equilibrium ratio.
Relations
- Phase equilibrium ratio model is a subclass of intensive thermodynamic state model.
- A phase equilibrium ratio model has a model quantity which corresponds to a phase equilibrium ratio.

Phase interface transport property model

Description
A property model that provides a correlation for a phase interface transport coefficient.
Relations
- Phase interface transport property model is a subclass of property model.

Reaction equilibrium constant model

Description
A property model that provides a correlation for a reaction equilibrium constant.
Relations
- Reaction equilibrium constant model is a subclass of intensive thermodynamic state model.

Specific enthalpy model

Description
A property model that provides a correlation for the computation of the specific enthalpy.
Relations
- Specific enthalpy model is a subclass of intensive thermodynamic state model.
- A specific enthalpy model has a model quantity which corresponds to a specific enthalpy.

48
Surface tension model
Description
A property model that provides a correlation for the computation of the surface tension.
Relations
- Surface tension model is a subclass of intensive thermodynamic state model.
- A surface tension model has a model quantity which corresponds to the surface tension.

Thermal conductivity model

Description
A property model that provides a correlation for the computation of the thermal conductivity.
Relations
- Thermal conductivity model is a subclass of intra phase transport property model.
- A thermal conductivity model has a model quantity which corresponds to a thermal conductivity.

Thermodynamic property model

Description
A property model that provides a correlation for phase system properties not relevant to rates.
Relations
- Thermodynamic property model is a subclass of property model.
- A thermodynamic property model has a model quantity which corresponds to a phase system property.

Vapour pressure model

Description
A property model that provides a correlation for the computation of the vapor pressure.
Relations
- Vapour pressure model is a subclass of intensive thermodynamic state model.

Viscosity model
Description
A property model that provides a correlation for the computation of the dynamic viscosity.
Relations
- Viscosity model is a subclass of intra phase transport property model.
- A viscosity model has a model quantity which corresponds to a dynamic viscosity.

49
1.8. Process unit models
The ontology module process_unit_models, located on the application-oriented level of OntoCAPE,
provides a collection of mathematical models that model the behavioral aspect of process units.
Please note that this module is introduced not for the purpose of providing a full account on this topic,
but rather for suggesting a principle for defining various types of process unit models and illustrating
the principle by means of only a few examples.
These exemplary property unit models are classified according to the modeled process units (Morbach et
al. 2008): a chemical reactor model models a chemical reactor behavior, a flash unit model models a flash unit
behavior, etc. (cf. Fig. 29).

models
MaterialAmount ProcessUnitModel

models
ChemicalReactorBehavior ChemicalReactorModel

models
FlashUnitBehavior FlashUnitModel

models
HeatTransferUnitBehavior HeatTransferUnitModel

DistillationColumn models TrayByTrayDistillation

Behavior ColumnModel

Fig. 29: High-level classification of process unit models

Beyond this high-level classification, the ontology module comprises some special types of process unit
models. Fig. 30 exemplarily shows the definition of a CSTR model: A CSTR model is a chemical reactor
model that models a chemical reactor behavior with the physicochemical phenomenon of phenomenon
ideal_mixing. Furthermore, the CSTR model is a first-principles model and incorporates the following
laws: energy conservation law, mass conservation law, and reaction kinetics law.

ChemicalReactorModel
MassConservationLaw

models
ChemicalReactorBehavior CSTR_Model EnergyConservationLaw

hasModeling
hasPhenomenon
Principle
ReactionKineticsLaw
ideally_mixed first-principles

Fig. 30: Definition of the class CSTR model

Currently, the ontology module provides only a few of such specialized process unit models. In the
future, it is to be extended to offer a substantial library of process unit models. As for now, the module
merely provides the framework for establishing such a library.

50
References
Gear CW, Petzold L (1984) ODE methods for the solution of differential/algebraic systems. Trans.
Society Computer Simulation 1:27–31.
Guthrie K (1969) Data and techniques for preliminary capital cost estimation. Chem. Eng. (New York)
24 (3):114-142.
Hairer E, Wanner G (1996) Solving Ordinary Differential Equations II -- Stiff and Differential-
Algebraic Problems, Springer, Berlin.
Ion P, Miner R, eds. (1999) Mathematical Markup Language (MathML) 1.01 Specification. W3C
Recommendation, revision of 7 July 1999. Online available at http://www.w3.org/TR/REC-
MathML/. Accessed April 2007.
Jensen AK (1998) Generation of problem specific simulation models within an integrated computer
aided system. PhD Thesis, Department of Chemical Engineering, Technical University of
Denmark.
Lang HJ (1947) Engineering approach to preliminary cost estimates. Chem. Eng. (New York):130-133.
Lloyd CM, Halstead MDB, Nielsen PF (2004) CellML: its future, present and past. Progress in
Biophysics and Molecular Biology 85 (2-3):433-450.
Martinson WS, Barton PI (2000) A differentiation index for partial differential-algebraic equations.
SIAM J. Sci. Comput. 21:2295–2315.
Morbach J, Yang A, Marquardt W (2008j) OntoCAPE 2.0 – Mathematical Models. Technical Report
(LPT-2008-28), Lehrstuhl für Prozesstechnik, RWTH Aachen University.
Peters MS, Timmerhaus KD (1991) Plant Design and Economics for Chemical Engineers, McGraw-
Hill, New York.
Schembra M (1991) Daten und Methoden zur Vorkalkulation des Anlagekapitalbedarfs von
Chemieanlagen. PhD thesis, Technische Universität Berlin.
Vogt M (1996) Neuere Methoden der Investitionsrechnung in der Chemischen Industrie. Diploma
thesis, Technische Universität Berlin.
von Wedel L (2002) CapeML – A Model Exchange Language for Chemical Process Modeling.
Technical Report (LPT-2002-16), Lehrstuhl für Prozesstechnik, RWTH Aachen University.

51
Appendix A Documentation Format

Classes
Classes are characterized by the following categories:
Description: A lexical description of the class, for example “A chemical reactor is an apparatus for
holding substances that are undergoing a chemical reaction.” The description explains the meaning of
the class to the user.
Definition: Unlike a description, a definition can be transcribed into a formal ontology language,
where it establishes the set of necessary and sufficient conditions from which the membership of an
ontological concept (class or individual) to the class can be inferred. Classes for which such a
definition can not be indicated are called primitive classes.
Relations: The following characteristics are indicated, if existent:
- Specialization. A list of parent classes from which the current class is derived via
specialization.
- Disjointness. A list of classes which are disjoint with the present class. Disjointness between
classes means that an instance of the first class cannot simultaneously be an instance of the
second class.
- Restrictions. Restrictions of binary relations (or attributes) specify the existence of a relation
(or attribute) as well as its cardinality and value range with respect to the current class.
Usage: Some recommendations for the use of the class may be given if such advice is required.

Relations
Binary relations are characterized by the following categories:
Description: Similar to that of classes mentioned above.
Characteristics: The following characteristics are listed, if existent:
- Specialization. A listing of the relations from which the relation is derived via specialization.
- Domain. The domain of the relation.
- Range. The value range of the relation.
- Inverse. The inverse of a relation.
- Further characteristics, such as if the relation is transitive, symmetric, or (inverse) functional.
Usage: As above.

Attributes
Attributes are characterized by the following categories:
Description: As above.
Characteristics: The following characteristics are listed, if existent:
- Specialization. A listing of the attributes from which the attribute is derived via specialization.
- Domain. The domain of the attribute.
- Range or datatype. The value range of the attribute, which is usually indicated by referring to
a built-in XML Schema Datatype (Biron et al., 2004).
- Further characteristics, such as if the attribute is functional.
52
Usage: As above.

Individuals
Predefined individuals are characterized by the following categories:
Description: As above.
Characteristics: The following characteristics are indicated, if existent:
- Instance of. The classes from which the individual is instantiated.
- Different from. A list of individual which are explicitly declared to be different from the
present individual.
- Relations. Instances of binary relations the individual is involved in.
- Attributes. Attribute values of the individual.
Usage: As above.

Notation Conventions
Classes and relations of the Meta Model are named according to the CamelCase 3 naming convention:
UpperCamelCase notation is used to denote identifiers of classes, while relation identifiers are
represented in lowerCamelCase notation. No particular naming convention is followed for identifiers
of individuals (i.e., instances of classes).
In this document, class identifiers are highlighted by italicized sans-serif font; for better readability, the
UpperCamelCase notation is not applied in the text, but the individual words that constitute the class
identifiers are written separately and in lowercase (e.g., class identifier). If relations are explicitly
referred to in the text, they are written in lowerCamelCase notation and are additionally highlighted by
sans-serif font. Individuals are accentuated by bold sans-serif font. Partial models are denoted bold serif
font, italicized serif font refers to ontology modules.
In figures, a graphical notation in the style of UML class diagrams is used; the basic elements are
depicted in Fig. 31. Grey shaded boxes represent classes, white boxes represent individuals. Attributes
are denoted by grey shaded boxes with dashed boundary lines, attribute values by white boxes with
dashed boundary lines. Specialization is depicted through a solid line with a solid arrowhead that
points from the subclass to the superclass. A dashed line with an open arrowhead denotes
instantiation. Binary relations are depicted though solid lines. Three basic relation types are
distinguished: a line with one open arrowhead represents a unidirectional relation; a line with two
open arrowheads represents a symmetric relation; a line without any arrowheads represents a
bidirectional relation 4 . Finally, graphic elements for two special types of relation are introduced: an
aggregation relation is depicted through a line with a white diamond-shaped arrowhead pointing
towards the aggregate class. Similarly, a black diamond-shaped arrowhead indicates a composition
relation.

3
CamelCase is the practice of writing compound words joined without spaces; each word is
capitalized within the compound. While the UpperCamelCase notation also capitalizes the initial letter
of the compound, the lowerCamelCase notation leaves the first letter in lowercase.
4
In OWL, a bidirectional relation is modeled through a unidirectional relation and its inverse.
53
class attribute binary relation (unidirectional)

individual attribute value binary relation (bidirectional)

composition binary relation (symmetric)

specialization

aggregation
instantiation

Fig. 31: Basic elements of graphical notation

54
Index of Concepts
Activity coefficient model.............................45 Economic performance model...................... 29
Adsorption equilibrium constant model ........45 Energy balance law....................................... 39
Adsorption equilibrium law...........................37 Equation system............................................ 14
Adsorption kinetics law.................................38 Equation system characteristics .................... 16
Adsorption rate coefficient model .................46 Equilibrium constraint .................................. 39
Agglomeration law ........................................38 explicit_formulation ..................................... 21
Algebraic equation system ............................15 Factorial FCI model...................................... 29
Algebraic model solution strategy.................24 first-principles............................................... 35
applies............................................................26 Fixed capital investment model .................... 30
Balance law ...................................................38 Fugacity coefficient model ........................... 47
Breakage law .................................................38 fully_implicit ................................................ 21
Capacity FCI model.......................................28 Generalized flux law..................................... 39
Chemical kinetics law ...................................38 Global factorial model .................................. 30
Chemical kinetics property model.................46 Growth law ................................................... 40
Chemical reaction equilibrium law ...............38 handlesDifferentialIndexUpTo ..................... 27
closed-form....................................................20 hasCoupling .................................................. 12
Constant...........................................................9 hasDAE_Explicitness ................................... 19
Constitutive law.............................................39 hasModelingPrinciple ................................... 34
Constratint .....................................................39 hasModelPort................................................ 12
continous_model ...........................................20 hasModelRepresentationForm...................... 19
Convective material flow law........................39 hasModelVariable......................................... 12
correspondsToQuantity .................................12 hasNumericalStiffness .................................. 19
Cost model.....................................................29 hasODE_Explicitness ................................... 19
Coupling ........................................................10 hasTypeOfInvolvedSteps.............................. 26
DAE explicitness...........................................15 hasVariablesType ......................................... 19
DAE solution strategy ...................................24 Heat capacity model ..................................... 47
data_driven ....................................................35 Heat radiation law......................................... 40
Density model ...............................................47 Heat transfer coefficient model .................... 47
Detailed-item FCI model...............................29 Heterogeneous reaction rate coefficient model
.................................................................. 47
determinesPositionOf ....................................13
Homogeneous reaction rate coefficient model
Differential algebraic equation system..........16
.................................................................. 47
Differential equation system .........................16
hybrid............................................................ 35
Differential factorial model ...........................29
implicit_formulation ..................................... 21
differentialIndex ............................................20
indexValue.................................................... 13
differentialOrder............................................20
Input variable.................................................. 9
discrete_model ..............................................20
Intensive thermodynamic State Model ......... 48

55
Interface heat conduction law........................40 Nucleation law .............................................. 42
Interface mass diffusion law..........................40 Numerical stiffness ....................................... 17
Interface viscous momentum transport law...40 ODE solution strategy................................... 25
Inter-phases transport phenomenon law........40 ODE_type ..................................................... 18
Intraphase heat conduction law .....................41 one-step_method........................................... 27
Intraphase mass diffusion law .......................41 open-form ..................................................... 22
Intraphase transport phenomenon law...........41 Ordinary differential algebraic system ......... 18
Intra-phase transport property model ............48 Ordinary differential equation system .......... 18
Intraphase viscous momentum transport law 41 Parameter ...................................................... 11
isAssociatedWith...........................................35 Partial differential algebraic model solution
strategy...................................................... 25
isIndexOf.......................................................13
Partial differential algebraic system ............. 18
isOrderedBy ..................................................13
Partial differential equation system .............. 18
Law..........................................................33, 36
Partial molar enthalpy model........................ 48
linear..............................................................21
Partial molar volume model.......................... 48
Linear algebraic model solution strategy ......24
Particle kinetics law...................................... 42
Linear algebraic system.................................16
Phase equilibrium law .................................. 42
Linearity ........................................................16
Phase equilibrium ratio model ...................... 49
Lower limit ......................................................9
Phase interface transport property model ..... 49
Mass balance law ..........................................41
Phenomenological coefficient law................ 42
Mass diffusion coefficient model ..................48
Population conservation law......................... 42
Mass transfer coefficient model ....................48
Port index...................................................... 11
Mathematical model ......................................10
Power factor model....................................... 30
Mathematical model (continued)...................17
Process model ............................................... 34
Mechanical equilibrium law ..........................41
Property model ............................................. 34
mixed_model .................................................21
Reaction equilibrium constant model ........... 49
Model port .....................................................10
Reaction kinetics law.................................... 43
Model quantity ..............................................10
Reset variables .............................................. 15
Model quantity specification .........................11
semi-explicit ................................................. 22
Model representation form ............................17
Six-tenths rule model.................................... 30
Model solution strategy .................................25
Solution strategy for explicit ODEs.............. 25
Modeled object ..............................................10
Solution strategy for implicit ODEs ............. 26
Modeling principle ........................................34
solves ............................................................ 27
Momentum balance law ................................42
Specific enthalpy model ............................... 49
multi-step_method.........................................27
Specific Gibbs free energy model................. 47
nonlinear........................................................21
State variable .......................................... 11, 15
Nonlinear algebraic model solution strategy.25
Step counting model ..................................... 30
Nonlinear algebraic system ...........................17
stiff................................................................ 22
nonstiff ..........................................................22

56
Surface mass diffusion law............................43 Turnover ratio model .................................... 31
Surface tension model ...................................49 Type of involved steps.................................. 26
Thermal conductivity model .........................49 Unit-cost estimate model .............................. 31
Thermal equilibrium law ...............................43 Upper limit.................................................... 12
Thermo model ...............................................50 Vapour pressure model ................................. 50
Thermodynamic state function law ...............43 Variables type ............................................... 18
Transport phenomenon law ...........................43 Viscosity model ............................................ 50

57