Loading

Module 1: Physical & Operational Architecture Development

Notes d'étude
Study Reminders
Support
Text Version

Physical Architecture Development

Set your study reminders

We will email you at these times to remind you to study.
  • Monday

    -

    7am

    +

    Tuesday

    -

    7am

    +

    Wednesday

    -

    7am

    +

    Thursday

    -

    7am

    +

    Friday

    -

    7am

    +

    Saturday

    -

    7am

    +

    Sunday

    -

    7am

    +

We found that using the requirement analysis, and the originatingrequirements documents, it is possible for us to identify the functions required in asystem, or from the top-level function we will be able to decompose these functions intosmall functions or the smallest functions, and then prepare a functional hierarchy for thesystem.The next logical step is basically to go for the design of physical system. So, this isbasically developing a physical architecture for the system, where we try to convert thesefunctional blocks into physical elements. We will try to see how we can identifycorresponding physical elements for the function, or how do we do a mapping of thefunction to physical elements which will satisfy the functional requirement of the system.So, in this chapter of physical architecture development, we will try to look at thefunctional decomposition and then the conversion of this function decomposition into aphysical architecture of the system.
(Refer Slide Time: 01:31)
As we shown in the previous lectures, the out of a 6 functions of the design process weare completed the system level design problem and the system functional architecturedevelopments, and the next task is basically the system physical architecturedevelopment.(Refer Slide Time: 01:50)
To explain the importance of physical architecture development, I will just explain casestudy basically of an real incident where the failures in the physical architecturedevelopment led to the failure.
So, using this case study I will emphasize the importance physical architecture. It is notonly the a customer requirement functions, but are, but other functions like, faulttolerance and identifying the physical component for fault tolerance plays a vital role inthe system development. And this case study is a perfect example to show that how thephysical architecture development over look the importance of having a fault tolerancesystem, and how this led to the failure of a important system especially in aircraft.So, this actually is a case study from an aircraft in Iowa. The United the aircraft 232;which was a 3-engine aircraft a crashed on 1989 while making an emergency landingafter losing one of the 3 engines. And in this case 110 people died and 185 survived.Most of you know most of the aircrafts are designed with lot of fault tolerant system, andin spite of all these aircraft failed.As I mentioned it was a 3-engine aircraft, and even if one engine fails or 2 engine fails itis possible to bring the aircraft to do a safe landing, and the pilots are being trained toovercome this failures in the system. And there are sufficient a fault tolerance in thesystem to overcome such emergencies. What actually happen here was the fan disk of theengine at the fuselage separated from the engine and crashed through the tail. So, thatwas the immediate cause for the failure. The fan disk of the engine separated and itcrashed through the tail of the aircraft.But that one engine failure was not problem because 2 engines were quite sufficient tomake a safe landing, but the aircraft stabilization system failed to control the descentrate. So, what actually led to the crash was the stabilization system failed. But how theengine failure caused stabilization failure was an important aspect of the system descent.Or and we look at that system descent, we will see that it is not only necessary to look atthe individual system we need to look at the overall system, and then see what are thepossibilities of failure and analyze it.So, here the even with the 2 engine the system stabilization system failed, how didhappen was; there were 3 redundant hydraulic systems, each powered by a uniqueengine. Which were available for aircraft stabilization? So, there were 3 hydraulicsystem. So, we can see this was a redundancy in the system, one hydraulic system issufficient to stabilization, and there were 3 hydraulic systems, each one powered byseparately by each engines. So, even if one engines fails and one hydraulic system fails
there were lot of redundancy is available, and it was possible to still control the aircraft.But the these 3 redundant hydraulic systems could not stabilize the system.(Refer Slide Time: 04:59)
Because the 3 hydraulic systems converge at the location near the tail where the fan diskripped out, the single point of failure for all the hydraulic system.So, the point where all the hydraulic system outputs where converging at one locationand the fan disk exactly ripped at that point and block the hydraulic system. That was thecause because even though all the 3 hydraulic systems are working perfectly, and therewere 2 engines powering the hydraulic system. The hydraulic power supply could notreach the control plane actuation points, because this was a single point where all the 3-hydraulic system converge to the give the supply.So, this was a single point failure in the system. And this is one of the important point tolook at when we provide the fault tolerant system we need to avoid the single pointfailures and then make enough redundancy in the physical architecture, to make sure thatsuch single point failures do not occur in the system.So, the physical architecture development is very important or it actually plays a vitalrole in identifying such situations and then eliminating the possible causes of failure. So,in this case there are pre-existing fracture on the surface of the fan disk was identified asthe main cause of engine failure. So, that was the main cause because the fan disk had
some problem. But the design flaw of single point failure resulted in the aircraft crash.The main reason why the aircraft crash was the single point where the all the hydraulicsystems where converging, and from that point onwards it was a single point failuresituation.So, this was the reason why actually it resulted. So, in the architecture development wewill look at ah the fault tolerant systems, and then how do we provide the necessaryredundancies in the system to avoid this kind of failures of engineering systems. So, letus look at the physical architecture development from the logical steps forward from thefunctional architecture. So, as we progress we will look at the redundancies and thenhow do we provide the physical system to have a sufficient redundancies in the system,and avoid such single point failures.(Refer Slide Time: 07:14)
So, here let us discuss about the physical architecture, and what is physical architecture?So, physical architecture of a system is a hierarchical description of the resources thatcomprise the system. So, basically, we a look at the physical resources, which actuallyprovide the functions in the system. So, architecture is basically we a write down thehierarchical structure describing each physical resource or the physical element whichprovides the necessary functions in the system.The hierarchy begins with the system, and the systems top level components andprogresses towards down to the configuration items. So, we start with the top-level
component or the top-level assemblies or sub assemblies, and then we move towards thesub assemblies and components and finally, to the configuration items. Configurationitems are basically the software hardware or a combination of software and hardwarepeople facilities or documents. So, any of these could be a configuration item. So, wewill list down all these items in a hierarchical fashion, and then complete the physicalarchitecture.So, in the physical architecture by looking at the physical architecture, will be able to tell
what are the components being used there, what are the assemblies, what are the sub-assemblies, and what are the other items which actually comes as configuration items in
terms of people facilities as well as the documentation. And then we will be able toidentify which are these components provide the functions or which function is mappedto the particular component, or which are the component which provide multiplefunctions, or which are the multiple functions satisfied by a single component; all thesethings you will be able to identify from the physical architecture.So, it provides the resources for every function identified in the functional architecture.So, that is the basic idea we you have the functional architecture, and from the functionalarchitecture we will try to identify the resources for every function.(Refer Slide Time: 09:11)
Now, how do we develop these physical architecture? It actually starts with thefunctional architecture. As I mentioned it is a logical step next to the after developing thefunctional architecture.So, we develop a very generic physical architecture from the functional architecture. So,generic physical architecture is basically a conversion of the functions into component,but we do not identified the exact component. Here we just write the generic name or thegeneric way of representing that function in terms of a component. So, that is the genericphysical architecture, without giving any details of the components used and then fromthat generic and physical architecture in order to get the physical elements. We create amorphological box for alternative physical elements. We will discuss about these indetail how do we do the morphological box or how do we develop the alternatives.But morphological box is a tool for developing the alternatives for physical component.So, we identify one generic component in the generic physical architecture, and then wedevelop or we identify all the generic component, and then develop morphological boxfor these components; where we will have many alternatives for that component and thenchoose the component. So, that is the use of morphological box where we havealternative physical elements. And then using these alternative physical elements we willgenerate the alternative instantiated architecture.So, instantiated architecture is basically an architecture a physical architecture where weidentify the component, and then write down the component instead of the genericphysical name of the component. So, here it is more specific and that is why it is knownas instantiated architecture. And from these instantiated architecture we will be having amany choices for he here, and then based on the our requirement will select a suitablephysical architecture. So, we go through the these steps, we will start with a genericphysical architecture, and then we go for a morphological box to identified thealternatives, and using the morphological box we develop instantiated physicalarchitecture, and using the instantiated architecture will identify a decide physicalarchitecture for the system.So, that is the process. And the exit criterion or when we decide or how do we choose aphysical architecture is basically the provision of a single physical architecture; that is,satisfactory in terms of a detail quantity and quality for development. So, we look at the
many possible options in the instantiated physical architecture. And then see whether itactually satisfy the in detail in quantity and quality of the development requirementbecause this has to be developed and we need to make the system. So, we look at the adetails of the instantiated physical architecture, and then see whether the quality andquantity are actually matching with the requirement. And once it is satisfied I will go forthat particular architecture and choice the physical architecture for the system.(Refer Slide Time: 12:11)
So, that is how we developed the physical architecture of a system. So, this is explainedusing the IDEF 0 diagram for different process. So, as you can see here. So, this is thefirst level well we do the brainstorming and select a generic physical architecture. So, inthis generic physical architecture of course, we need to do little bit of brainstorming herefrom the functional architecture to convert the functional architecture to a genericphysical architecture is bit tricky, and requires lot of a discussion and understanding ofthe functions and then finding out the generic elements for the generic architecture.So, we need to do bit of brainstorming. So, the team members will sit together and thenlook at the functional architecture. And then see how do we convert that in to a physicalarchitecture. So, for these the input will be the system level functional architecture, andthen the system level operational concept will be used in identifying the generic physicalelements. So, these are the 2 inputs, and then will be using this inputs and that will beused in the brainstorming session, and will be getting a generic physical architecture
here. And this generic physical architecture along with the input from the system levelfunctional architecture will be used for generating the morphological box for alternateinstantiated physical architectures. As I mentioned the morphological box will give youthe alternative elements for the generic elements identified in the generic architecture,and using this morphological box we can actually have alternate instantiated arc physicalarchitecture.And from there we can actually use these inputs from the instantiated physicalarchitecture and of course, will be using the input requirements as I mean the operationalconcept as well as the functional architecture, and then the requirement will be usingselecting the alternate instantiated physical architecture which actually satisfy the otherrequirement of the system, and that actually gives you the system level physicalarchitecture. And there would be many candidate physical architectures based on theinstantiated physical architecture. And the choice as I mentioned it depends on thequantity quality and the practical the feasibility of implementation, and developing thesystem.So, this is the general process of developing the physical architecture. As you can seethere would be many interactions and input and output between these, and wheneverthere are some changes will be go back and then get changed in the physical architecturegeneric architecture as well as the instantiated physical architecture will go through fewiterations to make sure that we are actually reach in a stage where we can choose onearchitecture for development, that is the way how the physical architecture developmentis happening.
(Refer Slide Time: 14:54)
So, let us go into the details of how we actually develop this. As I mentioned the genericphysical architecture is a description of the partitioned elements of the physicalarchitecture. Without any specification of the performance characteristics of the physicalresources that comprise each element. So, here we actually partition the elements as fromthe functional architecture we identify the elements, and then partition them and writedown them as in a hierarchical way a similar to the functional hierarchy.So, using the functional hierarchy will develop a physical hierarchy of the elements. Butin this case, we will not give any performance characteristics of the physical resources.So, it will be more like very generic an names. So, no performance specifications will bementioned in the architecture. So, that is the generic physical architecture though itactually provides common designators for physical resources in a hierarchicaldecomposition. So, it is or more like very common or general designator for the elementsor no specific components will be identified. So, the no specific physical systems areidentified here in the generic physical architecture. So, this is basically as a starting stepfor the physical architecture.So, converting the functional architecture to a generic physical architecture by giving thegeneric name for this components without specifying their performance, or the physicalelement used in the architecture. So, that is the generic physical architecture.
(Refer Slide Time: 16:22)
So, as we can see this we have seen the previous lecture. So, this is the functionalarchitecture. The functional architecture we write down the top-level function, and thenthe sub functions and then we decompose them into many levels, depending on therequirement and the lower level functions will be identified. And most of the time thislower level functions are the configuration item which will be identified in the genericarchitecture. And from this functional architecture we convert that in to a genericarchitecture this actually shows the functional architecture which actually we saw in theprevious lectures also.So, this is the from this functional architecture we develop the next level of physicalarchitecture.
(Refer Slide Time: 17:06)
So, as an example I can show you this is the functional architecture for an elevatorsystem. And we saw that this is the top-level function where move passengers betweenfloors and then accept passenger requests and provide feedback and control elevator carsthese are the sub function using functional decomposition. We identified how to get thesefunctions and from there we can actually decompose them into further. Like, provideinput output interface process data provide control commands etcetera. So, this way wecan actually develop.(Refer Slide Time: 17:33)
Now, when we convert this into a physical architecture, then we can see that the physicalarchitecture can actually be or a generic physical architecture will be more like directconversion of the functions into generic components.(Refer Slide Time: 17:47)
So, here you can see that the top-level component in terms of the generic architecture isthe elevator system. So, provide elevator services is the function, and elevator systembecomes the generic component which will provide this one.Similarly, provide input output interface for the passengers is the one of the functions.And here the passenger interface component becomes the generic elements. Similarly,elevator car or shaft component becomes the generic elements for the next function.Similarly, control elevator cars the control component becomes the generic element andmaintenance and self-test component becomes the generic elements for maintenancefunctions.
(Refer Slide Time: 18:31)
So, again we are not specified any specific control component or the elevator car or thetype of interface here. So, we have just converted the function to a generic physicalelement. And same way we can actually divide this component or the these assembly tosub component like elevator call announcement component, car control component anddestination control door control emergency control, then car component these a cabcomponent interior door component ventilation and lighting component. Like this shaftstructure components exit component controls shaft switch component floor stopcomponent leveling component drive break then here itself we can again decomposedinto normal drive emergency breaking like this.
(Refer Slide Time: 19:13)
So, what we are trying to do here is to convert the functions or the functional block intothe generic names or generic components. So, the here without any specification of thecar component or the control component or the breaking component or ventilationcomponent, we are simply writing down this as a generic name like interior doorcomponent or the ventilation and lighting component. So, what should be this particularelement what should be the specification of this will be decided in the next level.So, the first level is basically identifying the generic names or generic components andthen writing down them in the functional in the same as functional hierarchy or thesimilar way of functional hierarchy, and getting the generic physical architecture. So, thisis the way how we get the generic physical architecture. So, starting with the functionalhierarchy, we convert the functional elements in the functional hierarchy to genericelements and write down them in the hierarchical way then you are getting the genericphysical architecture.So, from these onwards we will be going to the next level, where we try to identify thecomponent for each of these elements. So, here we are not specified the elements. So,next level we will try to identify the elements for these components.
(Refer Slide Time: 20:33)
.
So, these another way of representing the hierarchy generic architecture, here again thisis for the aircraft component here. Again, we do not specify the particular elements here,but we will give a generic name for this like crew command devices, crew commandsensors, central controller aircraft device sensors, actuator controller, then actuator, thenaircraft devices, actuator controller and actuator.So, this is the way we write down the generic physical architecture. We do not specifywhat kind of a sensor we are using, or what kind of an actuator we are using and whatkind of controller we are using we simply specify the generic name of the thoseelements. And from based on this then will go to the level where we can identify fewactuators. And then choose one of the actuator for these replace this with the actualactuator we are using or the actual component, then we are getting the next architecturewhich is the instantiated architecture.So, in order to get the components names we need to do a brain storming and then try tofind out what are the possible option for these elements. And this is done through the thisis actually the same figure previous figure, explain in detail here or more for betterclarity. So, I can see the control the generic names of this elements written over here.
(Refer Slide Time: 21:55)
So, to get the instantiated physical architecture, which actually is a architecture to whichcomplete definitions of the performance characteristics of the physical resources havebeen added. So, we will take the physical architecture or the generic physicalarchitecture, and then we add the names of the components to these elements, and thenthe that gives us the an instantiated physical architecture. So, this can be done bymapping the functions to components, and then doing the and checking where thecomponents actually satisfy these functions.So, we can do different kinds of mapping of functions, one is known as the one to onemapping or on to mapping. There are different ways of mapping, this one we will seedetail and the and we allocate the architecture. So, in this for the time being you canunderstand that. This is basically used to identify whether all the functions are being mapto components, or how these components are map to the functions whether there is a oneto one or on to mapping. Or there are functions which actually satisfy multiplecomponents which actually satisfy multiple functions, or there are some function whichhave been left out without any component.So, this things can be identified using this kind of mapping functions. So, instantiatedarchitecture what we try to do is to identify the components for the physical elements.And then write down this elements with the architecture or the generic physical
architecture. So, when we add the actual physical component to the generic physicalarchitecture.(Refer Slide Time: 23:31)
We are getting the instantiated the physical architecture. So, to do this we need to go forthe morphological box. Basically, we take the physical architecture, we do a mappingand then we get the instantiated architecture.So, this mapping in order to do the mapping we need to generate the alternatives. So, wegenerate the alternatives through the mechanism called morphological box. So, we use amorphological box to get the components, and then we do a mapping of this componentwith the generic elements, and then we get the instantiated physical architecture. We willsee how to get the mapping.
(Refer Slide Time: 24:00)
Basically, we will develop the morphological box initially. So, morphological box will bedeveloped to get the alternatives, and then they will be mapped.As you can see a morphological box is a matrix representing the components of thegeneric architecture, and the alternative choices for fulfilling that generic component. So,this is the basically a matrix of alternatives for generic components. In the genericphysical architecture, we develop the elements generic elements, and then based on thesegeneric elements we can identify all the alternatives possible, and write down them in amatrix format we are getting the morphological box.So, this actually divides a problem into segments and posits several solutions for asegment. So, you can actually divide the whole problem into small segments, and theneach segments can be unless separately and we can get alternatives for each of thesesegment using morphological box. So, this an example for a morphological box.
(Refer Slide Time: 24:55)
If you look at the how morphological box are developed. So, this is example for a verysimple product called a hammer. So, you should take a hammer as a product, and thenhow do we develop the morphological box for these products. We look at the genericcomponents and the generic requirements like handle. So, if you take the handle as aproduct then we can see that handle size and handle material, then striking elementsweight of hammerhead, and nail removal element as the generic components in this case.(Refer Slide Time: 25:24)
So, these are the generic elements and we need to get develop the alternatives for thesegeneric elements. So, we can actually write down the possibilities here. So, alternativesare we can have a handle size of 8 inches or 22 inches, and we can have different handlematerials, fiberglass with rubber grip graphite with rubber grip or steel with rubber gripsteel I beam encased in plastic with rubber grip or wood. So, these are the possibleoptions for the handle material.So, this is the alternatives for handle material and for striking element. We can havealternatives like a one-inch diameter flat steel one-inch diameter grooved steel 1.25-inchdiameter flat steel. Or a 1.25-inch diameter grooved steel. Similarly, if the weight of thehammerhead can be 12, 16, 20 or 24, and nail removal element could be a steel claw atnearly a straight angle steel claw at 60-degree angle with handle. So, these are thealternatives possible for a particular system and particular product.So, this is this is just to explain the morphological box, actual system design will behaving much more complex alternatives, much more complex structure because we willbe having many generic elements. And then we need to have many options, and here thisis a simple product that is why you to see only few options given here. But even withthese few options or few alternatives, it becomes very difficult to choose a particularproduct because particular configuration, because we you need to look at the alternativeoptions here and then the implications of them in actually selecting the alternatives.(Refer Slide Time: 27:03)
So, here you can see these are the options from the architecture. If you look at themorphological box, you can see that there are 2 options over here, there are 2 optionshere, and you have 5 options in handle material, and another 4 options in hammerheadand then the striking feature again 4 options. So, how do we actually if you want to makeone hammer. So, what should be the possible combinations here.