Sunday, October 23, 2011

Manufactured Structural Products And Simplified Wind Classification

Manufactured products are typically classified into size ranges and/or performance grades. It is generally not practical or economical to have infinite variety of product offerings. Further more it is generally not economical to base the size ranges on an arithemetic series, and therefore size ranges are typically based on a geometric series usually a Renard series.

Buildings and other structures are typically designed and constructed one-off, rather than purchased off-the-shelf. However there is an increasing number of off-the-shelf buildings becoming available, the primary reason is that people want buildings not pictures of buildings. Unfortunately the industry is poorly served by consulting civil engineers, who are focused on one-off construction, and the need for consideration of site specific features. Whilst there are site specific features to consider it would be nice if the output of civil engineers actually reflected such custom consideration: rather than implementation of some routine solution and calculation of some point-value. Structural sections, bolts and a variety of other components all have standardised sizes and performance grades. All such components have critical characteristics for which minimum and/or maximum values can be determined for a specific generic application: there is no need to keep calculating from one project to the next: its a waste of time and paper. Simple selection criteria are required based on the controlling characteritics of the product. One such characteristic is the wind load.

To the wind loading code AS1170.2, a regional wind speed (VR) is selected: this is typically the wind speed experienced 10m of the ground, over terrain of category 2 (M[z,cat]=1): typically a local airport. This regional wind speed is then adjusted to match the site specific features, by the use of multipliers for terrain category (M[z,cat]), topography (Mt), shielding (Ms), and direction (Md), to give the site specific design wind speed (V[sit,beta])at a given reference height (z).

V[sit,beta] = VR Md (M[z,cat] Ms Mt)

The maximum from several different directions then becomes the design wind speed (V[des,theta])

This can then be converted into a reference pressure (qzu) as follows:

qzu = (0.5 rho[air]) V[des,theta]^2

The design pressure (p) on a given surface then obtained from:

p = qzu. Cfig Cdyn

{Example  Calculation sheet can be found at ExcelCalcs: schWindAssessment}

Where Cfig is a pressure coefficient dependent on the location of a given surface with in a given shape and configuration of building, and Cdyn is a dynamic response factor. From these formulae it can be seen that there are potentially an infinite number of pressures that a range of products need to be designed: not very practical. However for a generic application the values of Cfig and Cdyn are more or less lock in by AS1170.2, so the main variable is qzu, from one site to another. Whilst there are several factors which go into calculating qzu, the same value of qzu can be determined from a variety of differing inputs. Therefore wind load classes can be defined by the maximum value of qzu permitted in the class. And given that there is a minimum ultimate strength design wind speed of 30m/s, there is also a minimum value of qzu, which for the sake of argument could be called wind class N0. From the minimum then need a method of defining other classes. The basic principle adopted for AS4055, barring some historical anomalies and rounding, is that each wind class applies a pressure 1.5 times greater than the lower class. However due to the anomalies that doesn't quite hold true. Anycase 6 wind classes are defined N1 to N6, defined by ultimate strength design wind pressures (qzu), or otherwise by the associated design wind speed Vzu=V[des,theta]. Since tropical cyclones also impose fatigue issues, for the same wind speeds there are 4 cyclonic classifications C1 to C4.

Note that AS4055 is simplified wind loading for housing. The wind classification system itself does not have anything to do with houses, nor the dimensional constraints in AS4055 nor the pressure coefficients in AS4055: it would be far better if removed from AS4055 and placed in AS1170.2. For anything other than housing: AS1170.2 should be used to determine Vzu or qzu, and assign a wind class.

When defining a product, design to qzu determined from AS1170.2, then assign nearest lowest wind class. When selecting a product, assess the site to AS1170.2 and assign the nearest highest wind class. It should be noted that local government authorities (LGA's) produce wind speed maps for housing, a site classed as N1 for a house may not be classed as N1 for some other structure. A site can only be classified with respect to the reference height (z) of the structure. A three storey house is likely to fall outside the scope of the wind speed maps and the simplified tables in AS4055. Also the wind speed maps are roughly derived, so that it is beneficial to get a site specific wind load assessment.

For many N1 sites on the maps are at the lower end of wind class N2, and difficult to prove the full shielding which puts them into the lower class. However many of the N3 sites are near the upper end of wind class N2. In terms of timber framed housing to AS1684.2, this doesn't make much of a difference for there are only combined span tables for wind classes N1 and N2, member size largely controlled by live loading requirements rather than wind loading. However, the wind classification affects the requirements for lateral bracing and tie-down systems. The tie-down system mostly affects the connections, and that can be a matter of personal design philosophy. Not connecting the members to achieve the full capacity of the installed members, may be considered extremely wasteful. For eaxmple connections in a wind class N1 house have little to no reserve capacity for additional wind load, as may be imposed if a carport or verandah is attached to the house at some future date. {Contrary to popular opinion: you cannot attach any size carport or verandah you wish to a house structure in wind class N1. There is no reserve capacity, and chances are cannot attach one at all.}

Anycase the wind classification system permits simple selection, for a variety of manufactured structural products (sheds, carports, verandahs, houses, fences, windows, doors), however, determination of wind class should be by using AS1170.2 not AS4055. Alternatively additional design aids should be created to further simplify the assessment.

More importantly manufacturers should produce full technical specifications. Saying to AS1170.2, to AS4055, or to AS1684 is meaningless. The specific's of these codes should be identified in technical specifications for the product. In particular the internal and external pressure coefficients, or nett pressure coefficient used should be identified in the specification. Pressure coefficients do not vary between cyclonic and non-cyclonic regions. Those hanging baskets swinging from the pergola can equally well be thrown through the window in a cyclonic or non-cyclonic region. Determinining internal pressure coefficients is a complicated matter of assessing the risks associated with various states-of-nature that a building may experience. Assesssing all the states-of-nature is time consuming, so the simplest approach is to assume the building envelope is breached and that high internal pressures are generated. This is not however always conservative, since for a framed structure, it results in zero loading to the wind ward face. On the otherhand if high internal pressures do produce the maximum stresses in the frame it is not necessarily economical. The purpose of manufactured structural products is to be economical and provide fast supply.

The building code of Australia (BCA) requires consideration of the hazard to life, as well as the loss of amenity. Not having the amenity in the first place, may be fast becoming the major concept of loss. The magnitude of load has very little to do with the hazard to life. The design load always has a probability of being exceeded. so considering Tropical Cyclone Tracy, steel roof cladding was ripped from building and its sharp edges became lethal. All the cyclone testing of cladding and the cyclone washers, and the increased design loads, have failed to remove the sharp edges from the cladding. When the design load is exceeded as it will be one day, the hazard remains. But the population is going to become complacent about us having cyclone proof buildings, which we don't have: they will believe they are safe when they are not.

Safety is not a quantitative issue, it is a qualitative issue. When and how the structure fails when the design load is exceeded is the primary issue of design which is currently neglected. The current magnitude of design loads are largely determined by insurance councils and government with respect to the cost of replacing structure, not the hazard to life. Wind borne debris is a problem, but as indicated when the design loads are exceeded, it is still going to be there. The exercise is partly one of balancing inconvenience against disaster. Disaster arises when the community is not able to recover without external assistance. It should not be necessary for all parts of a building to have the same resistance nor to survive the same event. Many buildings, and many rooms within buildings are non-essential, and loss of such is not a major hardship. There are certain core facilities within houses which contribute to the quality of life in a modern city, these are primarily kitchens, bathrooms and laundries. Most other rooms in a house can be lost. More over buildings can be designed with weatherlocks which control the internal environment.

Another important issue to understand is that the BCA structural provisions are largely based on ultimate strength. That is stresses are permitted to exceed yield strength, and enter into the plastic behaviour zone of the material. This can be identified in the codes by the change from elastic modulus(Z) and use of plastic modulus (S), and the change from yield strength (fy) and the use of ultimate strength or fracture strength (fu). Materials are not expected to under go elastic recovery when the load is removed, they will remain permantly deformed even fractured. Probabilistic design permits less than or equal to the breaking load, it does not have to be strictly less than. The breaking load itself is an uncertain quantity which may be greater than the value used. The BCA uses 5th percentile characteristic strengths. After the structure has experienced its ultimate strength load it may have collapsed and ceased to be serviceable. This is important, a building is not designed to provide safe shelter during a hurricane, unless it is a post-disaster facility which is to remain serviceable after the event. After a design level event a normal building is expected to be no longer serviceable and to need replacement. The primary concern for the design level event is keeping the building anchored to the site, preventing it from becoming airborne debris. It is not the intent to keep it in service and operating.

For many years now the shed and garage industry has had members complaining about not being able to compete because there are those using internal pressure coefficients lower than the standard AS1170.2 specifies. Problem is AS1170.2 doesn't specify a value, it specifies a methodology for determining an appropriate pressure coefficient. That individual businesses in the industry cannot compete is largely because they rely on external consulting civil engineers, and otherwise know very little about structures and manufacturing or industrial engineering. Put simply they are mainly poorly designed businesses with poorly designed and even more poorly specified products. The businesses are also over loaded by sales people who get paid commissions to sell a product they of which they have zero understanding. A properly designed business could use significantly larger structural sections than most are using and wipe the majority of the suppliers off the map. The size of structual section has little to do with whether can compete or not.

So the Australian Steel Institute (ASI) shed group publishing documentation opposing the use of the wind classification system merely compounds the problems in the industry. The wind classification system is specifically for these types of manufactured products, and is far better to refer to wind class N1 or N2, than to refer to TC3 or TC2. The latter only gives consideration to one of the site characteristics whilst the wind class considers all the wind critical characteristics. To reiterate, AS4055 is for housing the wind classification system it defines is not limited to housing: but AS4055 can only be used to determine the wind class for housing, it is necessary to use AS1170.2 for other structures.

Damage to sheds during tropical cyclone Larry, and tropical cyclone Yasi, is not all together indicative of low quality non-compliant buildings. Many of the photos show the sheds collapsed, but it also shows the sheds still anchored to the original site: basic objective achieved. The BCA does not specify serviceability requirements, that is left as a subjective judgment for the designers and end-users, relative to the specific application.

One major problem with the shed industry is it runs around declaring their product complies with the BCA. Who cares? The product is required to comply with the BCA, so just provide the BCA evidence-of-suitability which demonstrates it complies. If it merely complies with the BCA then it is the lowest quality product permitted in the market. So forget about marketing BCA compliance, market how the product exceeds the BCA and provides higher levels of serviveability. We may not have tropical cyclones in South Australia, but we do experience tornadoes in the remote outback. Tornadoes are typically outside the scope of the BCA, but still need to be designed for. Design is not about code compliance it is about making the product fit-for-function, whilst giving due consideration to uncertainty and variability in its use and manufacture.

Now part of the problem is a failure to understand, that the fundamental law governing all supply is that for fair trading which requires goods are suitable for purpose. Once a product is released to the market or into the environment it will be used for all manner of purposes beyond the intentions of the designers. Technical specifications and product literature therefore need to make explicit the suitability of the product and the evidence-of-suitability. Shed manufacturers compete on price because one piece of junk is the same as any other, there is no added value for the higher price.

The building industry is not serious about the quality and performance of its products, it is largely why it is regulated. Buildings are failing because the component parts are not up to specification, and that is largely because the specifications of the major product itself is lacking.  Australia's relative isolation, and often monopolistic enterprises has seemingly resulted in many specifications being based on assumption. For example steel always from BHP, therefore don't really need to specify in detail. Coldformed steel sections always from Lysaght, so similarly don't need to specify in detail. This however is no longer the case, and those cheap c-sections are probably half the price because they are made from steel with half the strength. It is not the internal pressure coefficients that the ASI shed group should be concerning itself with. The products supplied need proper technical specicifications. Part of which requires putting  wind class N2 windows, along with wind class N2 doors, in a wind class N2 shed, on a wind class N2 site.

All the manufacturers need to get up to speed with the wind classification system, and the ASI shed group shouldn't be advising, near mandating that the wind classification system should not be used. Doing so makes te ASI shed group part of the problem. Door and window manufacturers need to provide their products with proper technical specifications. Due to the requirements of the glazing code, windows likely to be compliant, but door manufacturers do not yet appear to be paying any attention to the BCA. The doors to your house as well as the doors to your shed are highly likely non-compliant with the BCA. Consequently doors can be blown of their supports at less than design wind loads and lead to the development of high internal pressures not otherwise accounted for. Basically the economical design of the building is not taking into consideration the failure of the industry's ability to supply doors to a technical specification.

For a simple analogy. The walls to sheds have frames at 3m centres, with girts spanning 3m metres typically spaced less than 1.2m, the ribs in the wall cladding span the 1.2 m between girts. The roller doors are typically 6m wide, no frame behind, the ribs of the door cladding span 6m. The door clearly has no where near the resistance as the adjacent wall: the door is the weakest point on the wall. The wind will push the door in, causing it to balloon, until it stretches so much it slips from the door guides and is then torn from the building, at which point high internal pressures occur in the building and will end up loosing more than just the door. Make sure the doors are compatible with the specification of the shed.

Complacency is another problem in the building industry. People don't want any hassle, they go to shed manufacturers because there is an expectation that all design problems have properly resolved in the past. So delays because this door not compliant are typically over looked with, just get it finished: a doors a door. This is a suppliers problem. The supplier is supposed to have designed the product and be ahead of the local council and regulating authorities, not behind. There will always be delays experienced during development approval, if rely on the council to advise what regulations have to be complied with. The designers task is to assert which regulations are relevant and that they have been complied with. Sales people are not designers, and custom manufacture does not equate to custom design. If the shed supplier does not employ design personnel on staff, then buyers should seek the services of an independent consultant. Suppliers should seek to inform and educate the public.

Anycase I will essay shed design in more detail at a later date, along with design-for-failure. Other writings on wind loading to AS1170.2 with comparisons to ASCE7-05 can be found in the SEAint archives.

SEAint Archives: