Generation and
Evaluation of Alternatives for Building Elements
Murat Aygun, PhD.
Department of
Architecture, Istanbul Technical university
aygunmur@itu.edu.tr fax: +90 212 251
4895
The
intention here is to provide a parametric global model for generating element
alternatives in conjunction with an evaluation procedure to ascertain the best
overall performer. Any information pertaining to an object may be encompassed
within a conceptual parametric model thereby allowing efficient manipulation
for all purposes. The model enshrines three subsystems.1. Performance
requirements comprising the notional entities of Life cycle phases,
Participants and Domains with their respective entities. 2. Design Constraints
as exogenous factors predicating the requirements indexed to instances. 3.
Building objects as elements and their components described by object oriented modeling. The
comparative performance evaluation procedure for element alternatives comprises
two consecutive parts called physical and statistical evaluation. They employ
the same procedure but physical and statistical criteria are substituted
respectively. Thus the distribution of the criteria values for each alternative
is accounted for and a rank-ordered
shortlist of satisfactory alternatives may be prepared. Generation of
alternatives entails the combinetorial steps of Arrangement of components,
Material of components, Form of components. An application is included for the
purpose of demonstrating the proposal.
Buildings consist of
functional elements that significantly influence the overall performance. These
elements are either configured from previous examples or conceived ad hoc by collective experience or intuition.
They are hitherto
referred to with conventional terms, e.g. floors, walls and roofs. These labels
remain to be sufficient as far as conveying information on accustomed solutions
are concerned. Beyond that scope they impede on progressive element development
and appraisal. Hence a comprehensive yet versatile representation of all the
entities involved in the building system is required, may these be notional or
physical entities. The intention here is to provide a parametric conceptual
model for generating element alternatives in conjunction with an evaluation
procedure to ascertain the best overall performer.
An analytical approach would allow the consequences of the viable decision options to be explored and is conducive to identifying the synthesized solutions where all variable values have been compromised to an acceptable extent so that the demand and supply profiles are in concurrence without much under- or over-specification. Therefore the final selection is intended to take into account the distribution of all the criteria values for each alternative rather than aiming for the highest overall performer.
A brief survey of the
relevant works is presented as follows. Cross reviews the morphological chart
method for generating alternatives and the weighted objectives tree method for
the evaluation [1]. Mahdavi describes an object-oriented building
representation environment where a class inheritance hierarchy is adopted with
which relationships between elements are established [2]. Müller
focuses on building elements and divides their main functions into
sub-functions, compares the performances of component alternatives and then
configures the element [3 ]. Rivard et
al elucidate a shared conceptual model for integration in the building envelope
design process [4]. Vanier et al propose a product model for representing user
requirements as a complete data structure [5]. Aygun describes an analytical approach for the appraisal
of comparative performance for design alternatives[6]. Aygun
further presents a model intended for describing functional building
elements as evaluated in terms of
their life cycle performance by means of multiple indicators [7].
Any information
pertaining to an object may be encompassed within a conceptual parametric model
thereby allowing efficient manipulation for all purposes. The model enshrines
three subsystems as below.
Performance requirements: The
expectations and stipulations of participants at any phase of the building
life-cycle in a particular domain are expressed by performance requirements.
The latter as notional parameters in conjunction with the embedded constraints
below comprise transitive entities in the order of which is immaterial as far
as the information contained within is concerned. Each instance of the
preceding entity relates to all instances of the succeeding entity, i.e. the
branching is cross-linked and the links are identical , thus all predecessors
pertain to the same successors. The instances of these entities are ordered as arrays, i.e. one
dimensional matrices. These entities are listed below.
1. Life cycle phases (manufacture,
construction, occupancy including repair and maintenance, refurbishment,
demolition and recycle),
2. Participants (investor, designer,
contractor, user, community),
3. Domains (safety, health, comfort,
utility, ecology, aesthetics, cost)
Design Constraints: The
requirements above are predicated by constraints , also notional quantities,
describing the conditions of the environment in which the building or
it’s parts is situated. The constraints are deemed as exogenous external
or internal factors. They may be global or local. The salient constraints
relate to the site, climate, surrounds and legislation.
Building objects: The second
subsystem conveys information on physical objects and has an hierarchical order
and encompasses intransitive entities each preceding instance of which as an
ancestor relates to a different set of succeeding instances as descendants,
i.e. order-dependency. These entities are
1. Building (e.g. residential, office,
hospital),
2. Space (e.g. living, working,
sleeping),
3. Elements
3.1 Functional elements (e.g. floor,
wall, roof),
3.2 Structural systems
3.3 Service systems (e.g. mechanical,
electrical)
4. Components (e.g. finish,
insulation, waterproofing, core).
The approach observes
the rules of inheritance and encapsulation as the precepts of object-oriented
modeling. The orders of hierarchy and inheritance along the entities concerned
are accomplished in opposite directions, i.e. hierarchical inheritance. While
hierarchy is deductive (top-down), inheritance is inductive (bottom-up).
Since each ancestor is
almost invariably connected to at least one other ancestor of the same or another descendant there
will be one or more ancestors shared by descendants. Each instance of any given
physical entity is associated
either closely or remotely with all the instances all performance
entities.
The requirements for a
descendant are interpreted as functions of
ancestors in the context of
the physical object model. Hence element requirements are designated to
discrete components as their functions. They are therefore synonymous with
functions in this respect and subsequently asserted as quantitative terms
called performance criteria amenable to objective evaluation.
An element consists of any number of components,
each of which serves one or more primary functions. Conversely each component is allocated to one or more
functions, i.e. function sharing. By definition an element must have at least
one component which can be connected to another of the same element and also
shared by an adjacent element. Consequently this subsystem is partially
cross-linked. The branching extends laterally to include all elements in the
building system. Components are described herein by one or more attributes that
are deemed as the congenial properties of the respective components sharing the
same idiosyncratic attributes. Hence a physical entity as part of a building is
defined as indexed to and also is a function of all its descendant instances,
i.e. higher-level entities as embedded objects, in an hierarchical order:
Element = f (Components)
The
object hierarchy allows any sub-types (descendants) derived from the main types
(ancestors) to inherit the accrued attributes while retaining their embedded
attributes. Instances of these
objects are obtained when actual values are assigned to these attributes as
independent design variables of the functional element concept. The synopsis of
the element model is presented below:
Element
Location:
External (Below -, Above -, Partly above ground), Internal, Semi-enclosed.
Inclination:
Horizontal, Vertical, Inclined.
Components
( order of layers in the context of
the building envelope): External finish or layer (Surface characteristics),
Cavity (vented or unvented), Protective Layers (e.g. fire, heat, water, sound),
Core, Carrier, Supplementary Layer (e.g. filter or drain sheet, shading device)
Component
Geometry
(Form, Dimension, Position)
Texture
and Colour
Material
(Chemical, Physical and Biological description)
Joints
(Intra- / Inter-component; Unifying / seperating)
Structural
(Self-supporting, Supporting other component of same element / other element)
The notation above is
self explanatory except the distinction drawn between inter- and
intra-component joints. The former refers to those between two different
components belonging to the same or different elements. The intra-component joints
refer to those within the same component consisting of small units or layers,
e.g. tiles or laminates.
Thus by superposition of
the performance requirement model on the element model a requirement for any
given physical entity, e.g. a building element, can be expressed as a function
of the relevant instance of all three entities of this subsystem in any order.
A physical entity may acquire a requirement either directly from a notional
entity or indirectly as inherited
from its’ higher-level descendant physical entities. In both cases of
acquisition the requirements become encapsulated in that entity. Furthermore
any one of the requirements may be in concurrence, in conflict with or
unrelated to another as can be displayed in a triangular matrix.
Hence all requirements
may be listed as indexed to instances of entities in sequential order by
exhaustive enumeration.
Requirement wxyz
= f (Physical Entityw, Life-cycle x, Participant y,
Domain z)
Performance requirements
in conjunction with the specified constraints are translated to quantitative
expressions acting as criteria for the measurement of performance. The
variables of these expressions are the component and consequently element
attributes and constraint parameters.
Any attribute as a
variable may be implicated in one or more performance criteria for assessing
requirements. Conversely a criterion may involve any number of variables each of which incurs a rise or
fall in performance. If a given variable raises the performances of some
criteria (positive effect), i.e. mutually exclusive, while diminishing others
(negative effect) then a compromise value for that variable is required to
resolve the ambivalence by means of optimization. That variable is then called incoherent.
Else the appropriate extreme value is ascribed signifying a coherent variable.
These interactions between the component attributes and criteria can be also be
tabulated in terms of plus, minus and zero signs.
The change in
significance of the actual criteria values is described by the utility function
ascribed to each criterion. The relationship between utility and criterion can
be directly or inversely proportional, linear or curvilinear. In the latter
case the rate of change increases or decreases gradually or rapidly.
The relative performance
of any one or more of the five entities, i.e. building elements, life-cycle
phases, domains, participants and criteria, can be investigated by means of an
enhanced aggregation procedure. Iteration of the same procedure is required for
each additional entity under consideration, thereby increasing the degree of
aggregation. This modus operandi
therefore allows the inter- and intra-entity effects of the alternative
attributes to be explored.
The proposal comprises
two consecutive parts called physical and statistical evaluation as elucidated
below. They employ the same procedure but physical and statistical criteria are
substituted respectively.
Physical Evaluation: Initially
criteria are expressed as functions of attributes and then absolute values of
these criteria functions are obtained for each generic design alternative which
is then tested for compliance with the permissible values of performance
criteria. The latter are either self-imposed by designer or mandatory
stipulated by legislation, such as building regulations, codes and standards.
Consequently some alternatives are pre-eliminated. As a second option, only
those alternatives attaining above-average values for all of the criteria are
selected as prospective bona fide
solutions, reducing the total number of eligible alternatives further for
subsequent appraisal.
Since criteria can have
different dimensions, these values must be put in a non-dimensional form,
i.e.standardized, for the purpose of comparing performances of alternatives as
well as individual criteria. Therefore these values are converted from absolute
to relative by interpolation between the maximum and minimum absolute utility
values for each criterion.
Corresponding absolute
utility values are calculated through the appropriate utility function in terms
of the criteria involved. These values are employed to obtain the statistical
performance indicator for each element alternative at each phase as explained
later.
Even though not recommended
due to their phenomenological implications, weighting factors may subsequently
be applied with reservation to the relative utility values for taking into
account the relative importance of each criterion with regard to their
contribution to achieve preset objectives. For ascertaining the weighting of
criteria, the design objectives can first be rank-ordered by paired comparisons
in a square matrix. Then, if there are more than one level of sub-objectives,
then an objectives tree can be drawn. This process enables to be determined the
weights relative to each other at the same branch level and consequently those
relative to the overall objective.
Statistical Evaluation: This
procedure is the same as that of the preceding normalisation except that
statistical criteria are substituted in lieu of physical criteria by iteration.
Multiple statistical criteria are used here simultaneously as criteria of
overall performance to enhance the conventional aggregation process. The
utility values are employed to calculate the mean utility value as well as the
standard deviation and coefficient of variation for each element alternative at
each phase. Hence the mean is augmented by supplementary parameters. Thus the
distribution of the utility values for each alternative is accounted for.
Finally, these values are converted from absolute to relative as explained
previously. The mean of these values yield a single overall score as the
statistical performance indicator for each entity under consideration for
preparing rank-ordered shortlist of alternatives deemed as satisfactory.
By means of this model
element alternatives may be generated through the combination of component
alternatives so that all viable options are reviewed and also their effects explored.
Initially the prospective primary components are identified to which element
requirements are allocated as component functions. The combinetorial process
entails these steps:
Arrangement
of components: The components
identified in the preceding step may be arranged as layers in different
sequential orders with the exception of outer and inner finishes which retain a
constant position.
Material of
components:
Viable alternatives are listed for each component material and also its form to
be included in the element.
Form of
components:
Viable alternatives of component forms are selected to be included in the
combination process.
The number of element
alternatives is directly proportional to the number of design variables and also of their prospective values,
i.e. instances. The total number of generic alternatives may be reduced by
three ways. Prior to the generation the incoherent variables are identified so
that the combination only includes those. 1.The coherent variables are assigned
the appropriate extreme instances as constants. 2.Then the preferred
arrangement, materials and form of components are selected. 3.On completion of
the generation process, those alternatives containing any incompatible adjacent pairs or more
components are excluded from further evaluation.
The functional elements
in question are depicted by means of a conceptual parametric model comprising
performance requirements as
perceived by life cycle phases, participants and domains . Element alternatives
are generated by combinetorial manipulation of their attributes, i.e.
component, material and form. Subsequently these generic prospective solutions
are subjected to comparative evaluation in terms of multiple criteria
pertaining to the life cycle thereby accounting for the distribution of
criteria values for each solution. The proposal allows the most satisfactory
option to be ascertained under the specified circumstances. The phases of the
element life cycle are also appraised in terms of their overall performance.
Hence the consequences of various design decisions can be explored by
investigating the effect of variables on the overall performance of each
alternative.
The method is conceived
primarily for those involved in the research and development in building
performance and envisaged to act as a design or decision tool for professionals
in practice operating at various scales of the building production. There is
also a wide scope for application in diverse product fields. Benefits that may
be derived from this approach include improved information exchange and
repository facilitated by efficient retrieval and dissemination.
Subsequent applications
may include the generation of more comprehensive building assemblies entailing
many elements and components. Further research may be undertaken in
establishing priorities of performance requirements which remains to be an
ambiguous or vague subject.
The main difficulty is
encountered in sufficient real data related to the life cycle of products.
Therefore the aspect of measuring and recording such data becomes essential for
compiling performance knowledge bases.
The prescribed method is
illustrated by means of an application at the element level of the conceptual
model. The hypothetical element under investigation is the paved flat roof of a
sports hall in a suburban area with a temperate climate. The building structure
is a single span steel frame with trussed roof girders. The consecutive steps
followed are elucidated below.
1. Requirements and
Components:
The interaction between performance requirements and components are established
in conjunction with the associated life cycle phase, user and domains (Table
1). An example is this notation given below:
Thermal transmittance =
f ((Thermal insulator, Core) , Occupancy, User, (Ecology, Comfort, Cost))
2. Element alternatives: Generic
alternatives for the roof element are obtained by the combinetorial process of
changing the arrangement ( i.e. the relative position of components), material
and also form of components ( Tables 2 and 3). Arrangement 1 in Table 2 is
deemed as most appropriate and subjected to further evaluation, thus reducing
the number of alternatives. The capital letters in the adjacency matrices of
this table denote different types of joints between adjacent components: A
(external finish – air gap), B (air gap – thermal insulator), C
(thermal insulator – waterproofer), D (waterproofer – core), E
(thermal insulator – core), F (air gap – waterproofer). The
description of each type of joint varies according to the material and form of
the adjacent components. The heat transmittance coefficient of all the roof
alternatives is maintained as a constant value, ie. 0.4 W/kg m2.
3. Performance criteria: In this
case the alternatives are evaluated by means of 3 criteria, i.e. energy, waste
and cost shared by the consecutive phases of manufacture, construction and
occupancy.
For each generic roof
alternative a set of criteria values are obtained by instantiating variables.
While some of the criteria are in accordance with each other, e.g. energy and
waste, others can be in conflict, e.g. waste and cost.
4. Evaluation: The
implemented consecutive stages are described as follows. Table 4 displays the
actual criterion values of alternatives at each phase, providing the input data
for the subsequent evaluation process.
Then relative values are obtained through standardization. Utility functions and weights are
applied as linear and equal respectively. The statistical parameters for these
values, i.e. arithmetic mean, standard deviation and variation coefficient are
calculated for each phase as given in Table 5. They are then converted to
relative values and aggregated to obtain an overall statistical performance
indicator. Table 6 presents the final results that rank-order the alternatives
that attain above-average utility values at each phase and over the complete
life cycle respectively. Individual phases are also ranked.
1. CROSS, N., Engineering Design
Methods, John Wiley & Sons, Chichester, England, 1994.
2. MAHDAVI, A. Semper: A New
Computational Environment for Simulation-based Building Design Assistance,
International Symposium of CIB W67 on Energy and Mass Flow in the Life Cycle of
Buildings, Vienna, 1996, pp.467-472,
3. MÜLLER, H.,
Lehrbuch der Hochbaukonstruktionen: Methodik des Konstruierens, Cziesielski,
E., Ed., Teubner, Stuttgart, 1990, pp.9-24.
4. RIVARD, H, BEDARD, C, FAZIO, P,
Shared Conceptual model for the building envelope design process, Building and
Environment. Vol.34, 1999, pp.175-187.
5. VANIER, D J, LACASSE, M A AND
PARSON, A. Modeling of User Requirements using product modeling. 3rd
International Symposium: Application of the Performance Concept in Building.
CIB-ASTM-ISO-RILEM. Vol.2, Tel Aviv, 1996, pp.6-73-6-82,
6. AYGUN, .M. Comparative Performance
Appraisal by Multiple Criteria for Design Alternatives, Architectural Science
Review, University of Sydney, Vol.43.1, s.31-36, March 2000.
7. AYGUN, M. Evaluation of life-cycle
performance for functional building elements, International Symposium on the
Integrated life-cycle Design of Materials and Structures, RILEM/CIB/ISO,
Helsinki, Finland, 2000.
Table
1: Interaction between Requirements and Components
|
Requirements |
Components |
Interactions |
||||||||
|
q1Thermal
transmittance |
c1 External
finish |
c1 |
|
|
l |
|
|
|
l |
|
|
q2 Thermal
capacity |
c2
Air gap |
c2 |
|
|
l |
|
|
|
l |
|
|
q3 Solar
heat factor |
c3 Thermal
insulator |
c3 |
l |
|
|
|
l |
|
l |
l |
|
q4 Water penetration |
c4
Waterproofer |
c4 |
|
|
|
l |
|
|
|
l |
|
q5 Sound
reduction index |
c5 Core |
c5 |
|
l |
|
|
l |
l |
l |
l |
|
q6
Structural safety |
|
|
q1 |
q2 |
q3 |
q4 |
q5 |
q6 |
q7 |
q8 |
|
q7 Thermal
movement |
|
|
|
|
|
|
|
|
|
|
|
q8 Fire
resistance |
|
|
|
|
|
|
|
|
|
|
Table
2: The first 4 of 24 combinetorial alternatives of component arrangements and
their
respective adjacency matrices with different joints denoted by capital letters.
|
Arrangement 1 |
Arrangement 2 |
Arrangement 3 |
Arrangement 4 |
||||||||||||||||||||||||
|
c1
External finish |
c1
External finish |
c1
External finish |
c1
External finish |
||||||||||||||||||||||||
|
c2 Air gap |
c2 Air
gap |
c2 Air
gap |
c2 Air gap |
||||||||||||||||||||||||
|
c3 Thermal
insulator |
c3 Thermal
insulator |
c4
Waterproofer |
c4
Waterproofer |
||||||||||||||||||||||||
|
c4
Waterproofer |
c5 Core |
c3 Thermal
insulator |
c5 Core |
||||||||||||||||||||||||
|
c5 Core |
c4
Waterproofer |
c5 Core |
c3 Thermal
insulator |
||||||||||||||||||||||||
|
|
|
|
|
||||||||||||||||||||||||
|
c2 |
A |
|
|
|
|
|
c2 |
A |
|
|
|
|
|
c2 |
A |
|
|
|
|
|
c2 |
A |
|
|
|
|
|
|
c3 |
|
B |
|
|
|
|
c3 |
|
B |
|
|
|
|
c3 |
|
|
|
|
|
|
c3 |
|
|
|
|
|
|
|
c4 |
|
|
C |
|
|
|
c4 |
|
|
|
|
|
|
c4 |
|
F |
C |
|
|
|
c4 |
|
F |
|
|
|
|
|
c5 |
|
|
|
D |
|
|
c5 |
|
|
E |
D |
|
|
c5 |
|
|
E |
|
|
|
c5 |
|
|
E |
D |
|
|
|
|
c1 |
c2 |
c3 |
c4 |
|
|
|
c1 |
c2 |
c3 |
c4 |
|
|
|
c1 |
c2 |
c3 |
c4 |
|
|
|
c1 |
c2 |
c3 |
c4 |
|
|
Table 3: Material and form alternatives for
components
|
Component |
Alternative 1 |
Alternative 2 |
|
External
finish |
Concrete
slabs on pads |
Seramic
tiles on screed |
|
Insulation |
Extruded polystyrene
board |
Polyurehane board |
|
Waterproofer |
Bitumen sheet |
PVC sheet |
|
Core |
Profiled steel sheet |
Hallow concrete panels |
Table
4: Actual values of performance criteria for roof alternatives at each phase
|
Roof |
Manufacture |
Construction |
Occupancy |
||||||
|
|
Energy |
Waste |
Cost |
Energy |
Waste |
Cost |
Energy |
Waste |
Cost |
|
|
(kwh/m2) |
(kg/m2) |
(€/m2) |
(kwh/m2) |
(kg/m2) |
(€/m2) |
(kwh/m2) |
(kg/m2) |
(€/m2) |
|
R1 |
3.9 |
2.1 |
205 |
0.9 |
0.5 |
47 |
0.49 |
0.4 |
9 |
|
R2 |
3.2 |
3.3 |
170 |
1.3 |
0.3 |
63 |
0.14 |
0.6 |
11 |
|
R3 |
4.1 |
3.6 |
120 |
0.8 |
0.4 |
53 |
0.46 |
0.3 |
7 |
|
... |
... |
... |
... |
... |
... |
... |
... |
... |
... |
|
R16 |
3.5 |
2.9 |
175 |
1.3 |
0.2 |
60 |
0.34 |
0.4 |
10 |
|
R17 |
4.2 |
2.1 |
170 |
1.1 |
0.3 |
52 |
0.54 |
0.3 |
9 |
|
R18 |
4.6 |
1.8 |
145 |
0.7 |
0.4 |
41 |
0.37 |
0.2 |
8 |
Table 5 Actual values of statistical
performance indicators for alternatives at each phase
(Sta.D: Standard Deviation, Var.C: Variation Coefficient)
|
Roof |
Manufacture |
Construction |
Occupancy |
||||||
|
|
Mean |
Sta.D. |
Var.C. |
Mean |
Sta.D. |
Var.C. |
Mean |
Sta.D. |
Var.C. |
|
R1 |
5.35 |
3.74 |
0.70 |
3.76 |
2.84 |
0.76 |
5.63 |
3.45 |
0.61 |
|
R2 |
6.84 |
2.56 |
0.37 |
6.24 |
4.83 |
0.77 |
6.32 |
1.64 |
0.26 |
|
R3 |
5.45 |
3.14 |
0.58 |
4.52 |
3.08 |
0.68 |
4.69 |
2.27 |
0.48 |
|
... |
... |
... |
|
... |
... |
|
... |
... |
|
|
R16 |
5.22 |
3.47 |
0.66 |
5.88 |
3.85 |
0.65 |
5.47 |
2.83 |
0.52 |
|
R17 |
4.46 |
1.55 |
0.35 |
3.28 |
2.24 |
0.68 |
6.20 |
4.85 |
0.78 |
|
R18 |
5.06 |
2.28 |
0.45 |
5.43 |
4.25 |
0.78 |
7.44 |
1.27 |
0.17 |
Table 6: Ranked alternatives and phases with above-average
utility values
|
Rank |
Manuf. |
Constr. |
Occup. |
Life cycle |
Phase |
|||||
|
1 |
R7 |
9.16 |
R5 |
9.36 |
R6 |
9.74 |
R12 |
8.43 |
Manuf. |
7.75 |
|
2 |
R10 |
8.54 |
R13 |
8.75 |
R15 |
9.55 |
R6 |
7.69 |
Occup. |
5.52 |
|
3 |
R6 |
8.12 |
R9 |
7.43 |
R12 |
8.44 |
R16 |
6.47 |
Const. |
5.64 |
|
4 |
R3 |
7.74 |
R7 |
7.45 |
R9 |
7.56 |
R7 |
5.80 |
|
|
|
5 |
R16 |
6.58 |
R6 |
6.84 |
R17 |
5.88 |
R14 |
5.32 |
|
|
|
6 |
R12 |
5.66 |
|
|
R10 |
6.06 |
|
|
|
|
|
7 |
|
|
|
|
R14 |
5.90 |
|
|
|
|
|
8 |
|
|
|
|
R5 |
5.64 |
|
|
|
|