Travelling Through Time

in aid of Sustainability:

Building Design &

Maintenance SystemProfessor Farzad Khosrowshahi27 October 2010

• Maintenance on failure

• Preventative planned maintenance

• Time-based maintenance

Building Maintenance

Generations

• Too soon: uneconomical

• Too late: Poor service

Promise

. Building Maintenance Scheduling

. IDEAL (2) - Alternative Design Solutions

• Life cycle evaluation

• sustainability

Ideal (1): - Just-in-time maintenance

- live through it twice - virtual

Visualisation of Building.2D, 3D, animation, VR

4D Visualisation.In time

Visualisation through time.

Visualisation of the degradation of an object like a building:

application in maintenance

Building Visualisation Generations

Traditionally

Architect

Produce

building outline design

Client briefing:

Define functional and strategic needs

Building

Component Selection for

lifecycle,

sustainability,

energy.

Define Project

Environment(climate, usage,

locality)

Long Term &

Periodical

maintenance scheduling

Construction

Phase

Use to design for lifecycle and generate repair & maintenance program

Lifecycle engine

(knowledge-based

OR visual-based)

Proposed

Visual Building

Design

Information

Maintenance

Scheduler

Items Gen Properties

Building Design

Components AnalysisProperties of

Items in Building

Design

Maintenance

Scheduler Analysis

Visual User Interface

Eg. VRML Browser

Cad output

(convert to XML)

contains all building

details

’Items’ attributes

and properties.

Maintenance time period

and type of maintenance..

Default properties of

each standard ‘item’.

Produce a map of building ‘items’

and assign the suitable properties

to each one based on project initials.

Calculate building maintenance

procedure for the given time and

for the given scheduler.

Add building visual data

and calculated data for the given

time and maintenance scheduler

and produce output VUI

Architectural

DesignBuilding

Components

Maintenance

Schedule

Generation

Visualisation

Simulator

Maintenance

Scheduler

Simulator

Maintenance

Program

Virtual

Reality

Output

Architectural

Design

Building

Components

Maintenance

Scheduler

Generation

Visualisation

Simulator

Maintenance

Scheduler

Simulator

Maintenance

Program

Virtual

Reality

Output

Program from visualisation

Visualise the program

Development Process

Event-Effect relationship model

Events: list, agentsComponent: list, elements, time related variables

Event / Component interactionAuxiliary events

Maintenance scheduling expert system

Data generation: collect, generate, knowledge conceptualisation, mobilisation

Data Validation: case based , multi-modal

Event-effect modelling / database design

Product Development (software); interface design, web-based

Visualisation; VUI

Automate the maintenance decisions

LIFE CYCLE SIMULATIONLIFE CYCLE SIMULATIONLIFE CYCLE SIMULATIONLIFE CYCLE SIMULATIONLIFE CYCLE SIMULATIONLIFE CYCLE SIMULATIONLIFE CYCLE SIMULATIONLIFE CYCLE SIMULATION

Generalised life span modelGeneralised life span modelGeneralised life span modelGeneralised life span modelGeneralised life span modelGeneralised life span modelGeneralised life span modelGeneralised life span model

Time

Dead

Components

T1T2

A β PDF represent the lifetime

If pi(tj)<=RND(tj) then “Item i is broken at the tj”

LifeAge

Dirt

&

Dust

Usage

rate

Room

Temperature

Voltage

Generalised status algorithm.Generalised status algorithm.Generalised status algorithm.Generalised status algorithm.Generalised status algorithm.Generalised status algorithm.Generalised status algorithm.Generalised status algorithm.

Color of Component I at the time T

Red =Initial_Red_Colour*PDF_FNr(I,T)

*MaintenanceFactor;

Green=Initial_Green_Colour*PDF_FNg(I,T)

* MaintenanceFactor;

Blue =Initial_Blue_Colour*PDF_FNb(I,T)

* MaintenanceFactor;

End of Colour calculation.

Example life span algorithm for Example life span algorithm for Example life span algorithm for Example life span algorithm for Example life span algorithm for Example life span algorithm for Example life span algorithm for Example life span algorithm for

general light source.general light source.general light source.general light source.general light source.general light source.general light source.general light source.

Lumens of lamp I at the time T

Red_Lumens = Initial_Red_Lumen *

PDF_FNr(I,T) * Dust_Factor;

Green_Lumen= Initial_Green_Lumen *

PDF_FNg(I,T) * Dust_Factor;

Blue_Lumen = Initial_Blue_Lumen *

PDF_FNb(I,T) * Dust_Factor;

The_Lamp_Lit_Colour = RGB(Red_Lumens,

Green_Lumens, Blue_Lumens);

End of Lumens calculation.

Death status of lamp I at the time T

Probability_of_Death = PDF_FNd(I,T) *

Environment function based on

Maintenance factor (I);

if Probability_of_Death >= RND(I,T) then

Bulb is in broken state;

Else

Bulb is not in Broken State;

End of lamp death status.

Example of a bedroom before light Example of a bedroom before light Example of a bedroom before light Example of a bedroom before light Example of a bedroom before light Example of a bedroom before light Example of a bedroom before light Example of a bedroom before light

manipulation.manipulation.manipulation.manipulation.manipulation.manipulation.manipulation.manipulation.

Example of a bedroom after Example of a bedroom after Example of a bedroom after Example of a bedroom after Example of a bedroom after Example of a bedroom after Example of a bedroom after Example of a bedroom after

supervised light manipulation.supervised light manipulation.supervised light manipulation.supervised light manipulation.supervised light manipulation.supervised light manipulation.supervised light manipulation.supervised light manipulation.

Changes due to

lamps’ death.

General reduction

in lumens due to

dust factor and

lamp’s age.

EVENT-EFFECT MODELING

- CHELLENGE

Events-Effects

Event

Rain

Component

Cladding

Act on Effect

corrosion

� EVENTS: many and diverse

� Soft / Hard

� COMPONENTS (internal / external / outside): many choices

� Identify all variables that inflict degradation

� Define time-related visual attributes

�

� EVENT & EFFECT MODEL – linear

� Identify list of effects and associate with components� Develop data structure for each component – effect of time

� Require conceptualisation

� INTERRELATIONS – chain events

� scenarios / cases

� EVENT SIMULATION

� Identification of event situations

� Heuristic rules for frequency and intensity of events

Maintenance & Repair Complexity

Events:Rain, Wear,

temperature, people,

air movement, long-

wave radiation,

chemical attack,

biological attack, fire,

man-made disaster,

natural disaster.

Effect:Colour rendition, surface texture, shape,

mortality, erosion, corrosion, loosing

reflective, absorption and transmission

properties.

LEVEL 2

. Wall finishes

. Floor finishes

. Ceiling finish

. Sanitary appliances

. Services equipment

. Disposal installation

. Water installation

. Heating, Air & Vent.

. Electrical installation

. Gas installation

. Lift/Convey instal'n

. Communication

instal'n. etc.

LEVEL 1: COMPONENTS.Substructure

.Superstructure

.Internal FinishFitting &

Furnishing.Services

.External Works

.Etc…

LEVEL 3. Wall cover

. Doors

. Windows

. etc.

LEVEL 4. fixture

. handle

. etc.

LEVEL n. etc.

Component Aggregation Levels

Events-Effects – modeling

Event

Rain

Component

Cladding

Act on Effect

corrosion

Agents water

rain

steel

cladding

Events

Corrosion

Effects

Elements

Components

Events-Effects Relational Model

Short-circuited?

PVC

popu

tionwater water

pollution

water

acid

Active AgentsAgents Auxiliary Agents

Example: Agent-water plus Agent-pollution = Active Agent-acid

Combined effect of two or more Agents

Combined effect of two or more Events

Example: Event-rain plus Event- vandalism-x = Active link to

Component-Cladding-Element-Steel

popu

tionrain cladding

plastic

steel

cladding

ComponentsEvents Auxiliary Events

Rain/vandel

Extended model of Events-Effects

water

rain

steel

cladding

Agents

Events

Corrosion

Effects

Elements

Components

auxiliary

auxiliary

Probes

Functions

Yes / nochainEFFEC

TCorrosion

EVENTEVENT

EVENT

Component

INTELIGENT OBJECTS

- construction & engineering

- facilities management

- environmental science

- materials science

- mechanical engineering

- meteorology

- physics

- chemistry

- management – risk, project, safety, etc..

- architecture

- maths & stats, simulation & modelling

- computer science

- visualisation

- maintenance programming & scheduling

- energy conservation

- geologist

- psychologists

- sociologist

- time geography

- anthropology

- etc.

OTHER CHELLENGES;

Holistic Nature - Several Expertise

- anthropology

BEHAVIOURAL PATTERNSeg. Degradation of Flooring Systems

[Visualisation of Impact of time on internal flooring]

1- Pattern of usage

2- Visualisation of degradation

There are two methods for analysing human spatial behaviour:

- Activity system (Chapin, S., 1968); is primarily

concerned with the organisation of the sequence of

activities taking place in a setting.

- Behaviour Settings (Barker, R., 1968); concerned

with the relationship between the setting and a

recurring pattern of behaviour. [anthropology]

CentreKeep distance

5 0 5 100

0.2

0.40.398947

9.13483e-012

a( )x

b( )x

c( )x

8-5 x

5 0 5 100

0.2

0.4

0.60.415847

8.64717e-010

g( )x

8-5 x

dyy

yxfxg

j

iy

∫=

=),(

)(

The presentation of probability of appearance of

n creatures in a long room (Corridor)

with one entry (Door).

The presentation of appearance of

n creatures in a long room (Corridor)

with two entries (Doors).

Unused Carpet Texture

Used Carpet Texture

ANTI-ANTHROPY

Clients

s

t

a

n

d

a

r

d

s

Input DataOutput Knowledge

(tailored)

Wider

Dissemination

Wider Exploitation

-Popularize ~music

FINAL PRODUCT: ‘universal’ Web-based Design & Maintenance knowledge Broker

Designers

Maintenance manager

Financiers/Insurers

Public authorities

Community

Pressure Groups

Web – HOLMES

Knowledge Broker

ManufacturersResearchersHOLMES

Phase1 : Initial population

Phase 2: Promotion&PublicityPhase 3: Data volunteers

General Public

Etc.

END