1.1, the five main objectives that need to be accomplished in pursuit of the goal of the research in this handbook are: • the development of appropriate theory on the integrity of engine
Trang 1Fig 4.38 Example of defined computer automated complexity (Tang et al 2001)
An indicative example of defined computer automated complexity is given in Fig 4.38 (Tang et al 2001)
b) Complicatedness as a Function of Complexity
Complicatedness is the degree to which any control over the system is able to manage the level of complexity presented by the system The means of control can be another system or a person Complicatedness is a function of complexity,
K = K(C) Clearly, at C = 0, K = 0, the properties of a complicatedness function
are essentially the same as those of complexity but they are definitely not
identi-cal For example, consider K when C →∞ Inevitably, there is a level of complex-ity at which any means of system control simply cannot cope with the system as
a whole The system then becomes unmanageable through diminished or lack of control
It is relatively easy to visualise a graph for g = g(x,y) with C g = O(p2) (i.e
two-dimensional), and less easy to visualise a graph for h = h(x,y,z) with C h = O(p3) (i.e three-dimensional) However, a surface with four variables is indeed difficult to
visualise, although complexity has only reached O (p4) Consider the
incomprehen-sible systems A and B where C a = O(p100) and C b = O(p1,000) The
complicated-ness functions are virtually the same in this case, K a ≈ K b , although O (p1,000 )
O (p100) Therefore, when C = 0,K = 0 and C →∞, then K→ Kmax
Systems are designed to operate and be controllable at an optimal point of
com-plexity, i.e C ∗ Where C < C ∗, although complexity increases, it is well within the
interval of controllability Where C = C ∗, the system complexity is optimal with
re-spect to its ability to be controlled and, where C > C ∗, complexity is increasing, and
the system can be controlled only with decelerating (i.e exponentially diminishing)
Trang 2484 4 Availability and Maintainability in Engineering Design
Fig 4.39 Logistic function of complexity vs complicatedness (Tang et al 2001)
effectiveness This can be expressed mathematically as:
dK
dC = {0,∞} (i.e in the open interval between 0,∞)
d2K
d2C > 0 at C < C ∗ where complexity is increasing faster than
complicatedness
d2K
d2C = 0 at C = C ∗ where complicatedness has reached an
inflection point
d2K
d2C < 0 at C > C ∗ where complicatedness has reached
saturation
For C < C ∗, d2K /d2C > 0, complexity is increasing faster than complicatedness For C > C ∗, d2K /d2C < 0, the ability to manage complexity has reached
dimin-ishing returns
Because the logistic function is one of the simplest mathematical expressions that has all the properties considered previously, it is adopted to express complicat-edness as indicated in the following expression and illustrated in Fig 4.39 (Tang
et al 2001):
K (C) = Kmax
where:
e is the transcendental number e= 3.27182818284
α is a constant specific to the measure of control
C is the complexity of the system
Kmax= 1 indicates absolute complicatedness
Trang 3There are other means of expressing complicatedness, such as using the Weibull distribution The major differences, though, are the location of the inflection point,
the growth pattern before and after the inflection point, and the symmetry around the inflection point
c) Designing for Complex but Uncomplicated Systems
The complexity of engineering designs increases relative to the integration of their input vectors The integration of system elements resulting in new interactions and changes in bandwidth (due to volume or capacity constraints) increases the initial
design’s complexity However, engineered complexity can reduce intractably
com-plicated input vectors to a minimum number of output vectors that renders the
sys-tem controllable—and syssys-tem complexity manageable The application of neural networks is increasingly being considered for process control of complex integrated
systems, in situations where there are intractable numbers of data points to analyse This approach has proven effective for engineering designs in which the process is controlled in real time by adaptive and distributed artificial neural networks (ANN) embedded in distributed control systems The application of ANN is considered in detail in Sect 5.3.3
Earlier, the vehicle transmission was presented as a complex system that is
uncomplicated The automatic transmission presents the system image of A=
{P,R,N,D1,D2,D3}, λi j = 24, the number of linkages between the transmission interactions (four per ratio), and the bandwidth of linkages (capacity) between the
interactionsβi j = 1; thus, C a= (62)(24)(1) = 864 (where P = park, R = reverse,
N = neutral and D1to D3= drive transmission ratios) However, the manual
trans-mission presents the system image of M = {P,R,N,D1,D2,D3,C} where C = clutch.
This needs to be engaged and disengaged, so C’s interaction bandwidth is 2 Thus,
λi j= 10 (two per ratio) withβi j= 1, andλmn= 14 withβmn= 2 The complexity
of the manual transmission is:
C m=72
[10 + (14) · 2]2= 38,416 Suppose, for a novice driver, C ∗ ≈C a = 864 and, at C ≈ 40,000, Kmax= 1, indicating absolute complicatedness The analytic form of the complicatedness function for
engineering design can now be determined for a system with complexity C and complicatedness K:
• Determine optimal complexity, C ∗, which can be optimally controlled.
• At the optimal complexity C ∗ , set K ∗ = 1/2.
• Solve forα from K ∗ = 1/(1 + e −αC ∗
) where Kmax= 1
• Determine K(C) = 1/(1 + e −αC)
Trang 4486 4 Availability and Maintainability in Engineering Design
4.4 Application Modelling of Availability and Maintainability
in Engineering Design
In Sect 1.1, the five main objectives that need to be accomplished in pursuit of the goal of the research in this handbook are:
• the development of appropriate theory on the integrity of engineering design for
use in mathematical and computer models;
• determination of the validity of the developed theory by evaluating several case
studies of engineering designs that have been recently constructed, that are in the process of being constructed, or that have yet to be constructed;
• application of mathematical and computer modelling in engineering design
veri-fication;
• determination of the feasibility of a practical application of intelligent computer
automated methodology in engineering design reviews through the development
of the appropriate industrial, simulation and mathematical models
The following models have been developed, each for a specific purpose and with specific expected results, in partly achieving these objectives:
• RAMS analysis model, to validate the developed theory on the determination of
the integrity of engineering design
• Process equipment models (PEMs), for application in dynamic systems
simula-tion modelling to initially determine mass-flow balances for preliminary engi-neering designs of large integrated process systems, and to evaluate and verify process design integrity of complex integrations of systems
• Artificial intelligence-based (AIB) model, in which relatively new artificial intel-ligence (AI) modelling techniques, such as inclusion of knowledge-based expert systems within a blackboard model, have been applied in the development of
intelligent computer automated methodology for determining the integrity of en-gineering design
The process equipment models (PEMs) for application in dynamic systems simula-tion modelling will now be looked at in detail.
4.4.1 Process Equipment Models (PEMs)
As indicated previously, process equipment models (PEMs) have been developed for application in dynamic systems simulation modelling to initially determine
mass-flow balances for preliminary engineering designs of large integrated process sys-tems The dynamic systems simulation modelling was developed using the propri-etary OOP simulation shell, Extendc (Diamond 1997).
Extendc is a flexible simulation modelling system with a customisable interface
where system blocks can be modified or created using a built-in compiled language
It combines the most powerful features of object oriented programming (OOP) for
Trang 5advanced dynamic simulation with discrete event/continuous system/combined sim-ulation capability, top-down/bottom-up systems hierarchic reachability, animated graphics, advanced statistical and sensitivity analysis, and computer interface with drag-and-drop and point-and-click capabilities
The PEMs incorporate all the essential preliminaries of process analysis to de-termine mass-flow balances for preliminary engineering designs of large integrated process systems The simulation models also incorporate algorithms of process de-sign integrity for assessing reliability, availability, maintainability and safety re-quirements of process systems These are incorporated in specific probability distri-bution modifiers within each PEM The application of dynamic systems simulation modelling incorporating the PEMs is primarily intended to determine the applica-bility and capaapplica-bility of simulation modelling during the engineering design stage,
in accurately assessing the effect of complex integrations of systems in large engi-neered installations
The dynamic systems simulation modelling is based on classic methodology
of systems simulation, which is described in detail in the following presentation
of the application of computer modelling in engineering design verification The PEMs have been developed within the Extendc Performance Modelling program
(Extend 2001), integrated into a dynamic systems simulation blackboard model
for application in concurrent engineering design in an integrated collaborative de-sign environment in which automated continual dede-sign reviews may be conducted throughout the engineering design process by remotely located design groups com-municating via the internet
Design methodology and dynamic systems simulation The integration of
dy-namic systems simulation with blackboard design methodology allows for the devel-opment and integration of the basic building blocks of systems engineering design
that can be represented in a design knowledge base Support systems in the form
of general-purpose design knowledge sources are similarly developed to support the
design knowledge base The design knowledge base and design knowledge sources form the core of an integrated design support system The design objects in the de-sign knowledge base can be synthesised to generate conceptual dede-sign solutions, as illustrated in Fig 4.40
A dynamic systems simulation blackboard model (ICS 2002) is developed to
con-trol the design knowledge sources and integrate the knowledge-based design appli-cations such as the PEM blocks The design knowledge base contains design objects, relations, constraints in terms of intended function and interfaces, as well as detailed information in terms of geometry and sizing
The blackboard model The blackboard model is a paradigm that enables the
flex-ible integration of analytic methodology into a single problem-solving environ-ment In terms of the type of problems that it can solve, there is only one major assumption—that the problem-solving activity generates a set of intermediate re-sults This is evident throughout the dynamic systems simulation modelling inte-grated into the blackboard model, with systems selection in hierarchical structures
as illustrated in Fig 4.41
Trang 6488 4 Availability and Maintainability in Engineering Design
Fig 4.40 Blackboard model and the process simulation model
The blackboard model consists of a data structure (the blackboard) containing information (the context) that permits a set of modules (knowledge sources) to inter-act The blackboard can be seen as a global database or working memory in which distinct representations of knowledge and intermediate results are integrated uni-formly It is also a means of communication among design teams, and can be used
as a common display for review and performance evaluation
Blackboard architecture consists of three major components:
• The knowledge sources, which are software specialist modules Each knowledge
source provides specific expertise The ability to support interaction and
cooper-ation among diverse knowledge sources creates enormous flexibility in
engineer-ing design
Flexibility in this context is the ability to change the blackboard database imple-mentation, the insertion/retrieval strategies, and the representation of blackboard objects without modifying knowledge sources or base data such as design specifi-cations Flexibility in blackboard architecture for engineering design is important for two reasons First, understanding of the insertion/retrieval characteristics and the representation of blackboard objects may be uncertain and, therefore, sub-ject to change as the design is developed Second, even after a schematic model prototype of the design has been completed, the number and placement of
Trang 7black-Fig 4.41 Systems selection in the blackboard model
board objects may differ from those of the prototype This requires changes to the blackboard representation to achieve the desired level of performance (Corkill
et al 1987)
• The blackboard, which is a shared repository of problems, partial solutions,
sug-gestions, and contributed information The blackboard can be thought of as a dy-namic library of solutions to the design problem that have been contributed by other knowledge sources Thus, a blackboard in engineering design is an ap-proach that allows knowledge sources to cooperate in solving a design problem This is analogous to a group of designers standing around a blackboard The blackboard is a database that is used to hold shared information among the par-ticipants (or knowledge sources) It may be structured so as to represent different levels of abstraction as well as distinct and possibly overlapping concepts in the design solution The division of the blackboard into systems hierarchy levels (as with the PEMs) parallels the process of abstraction of the knowledge, allowing elements at each level to be described approximately as abstractions of elements
at the next lower level This partition of the knowledge is useful, in that a partial solution (i e group of hypotheses relating to design optimisation) at one level can be used to constrain the design at lower system levels
Trang 8490 4 Availability and Maintainability in Engineering Design
Fig 4.42 Design equipment list data in the blackboard model
• The control shell, which controls the flow of problem-solving activity in the
sys-tem Knowledge sources need a mechanism to organise their application in the most effective and coherent fashion In a blackboard system, this is provided by the control shell
Knowledge sources Each knowledge source is data-directed, in that the blackboard
is monitored for data matching-specific design preconditions Knowledge sources may be classified in a number of different ways depending on the characteristic that
is used to discriminate these For example, a generic knowledge source may be use-ful in a whole set of knowledge-based systems (e g design equipment list data for application in dynamic systems simulation modelling of a particular design solution,
as illustrated in Fig 4.42), or specific to one application (e g specific probability distribution modifiers within each PEM for assessing reliability, availability, main-tainability and safety requirements of process systems in a design)
The generic knowledge source in Fig 4.42 of design equipment list data, for application in dynamic systems simulation models of specific alumina processing stages, gives relevant data of the equipment such as equipment code, flow volumes, mass-flow volumes, liquid volumes and solids volumes
Trang 9Fig 4.43 Systems hierarchy in the blackboard model context
The context The context is a set of entries or context elements in the blackboard
that contain the information representing the state of the solution process For ex-ample, in the dynamic systems simulation blackboard model, PEMs are selected ac-cording to a systems hierarchy, as illustrated in Fig 4.43 Those entries may include perceptions, observations, hypotheses, decisions, goals, interpretations, judgements
or expectations Also, they may have relationships to one another In particular, one such organisation may combine a set of entries as the representation of a single ob-ject viewed from different levels of abstraction There can be obob-jects that represent goals, questions and information, knowledge sources, and other general concepts in the blackboard, as well as domain-specific objects
Figure 4.43 illustrates the selection of information representing the state of the alumina process by plant/facility (third train), operation/area (bauxite grinding) and section/building (also bauxite grinding)
The user interface The user interface permits the interaction of the user (designer)
with events inside the blackboard and indirectly with the rest of the knowledge sources comprising the system This interaction may occur in both directions—by the users modifying the flow of control of the system by means of commands and an-swers to questions, or by the system informing the user of important events, prompt-ing for answers, or explainprompt-ing decisions The user interface manages the question
Trang 10492 4 Availability and Maintainability in Engineering Design
Fig 4.44 User interface in the blackboard model
and answer protocols, and informs the user of important events during the program’s execution Among its most important capabilities are the following: it checks if an answer is valid, (based on pre-specified or dynamic menus or constraints), advises the user on valid or desirable answers, manages default values, and automatically completes queried answers
Figure 4.44 illustrates a process pre-commissioning user interface in the black-board model for information relating to a specific alumina process equipment: the bauxite grinding system, and ball mill assembly
Dynamic system simulation in engineering design Dynamic system simulation
in engineering design provides for typical virtual prototyping of engineering
pro-cesses, rather than experiments on the physical prototype Not only does virtual prototyping make design verification faster and less expensive but it also provides various design teams in a collaborative design environment with immediate feed-back on design decisions This, in turn, promises a more comprehensive exploration
of design alternatives and a better performing final design To fully exploit the ad-vantages of virtual prototyping, dynamic system simulation is the most efficient and effective However, these simulation models have to be easy to create Creating dy-namic simulation models is a complex activity that can be quite time-consuming