Sunday, October 21, 2012

Engineered Products and Evidence-of-Suitability

Really beginning to dislike civil engineers supplying services to manufacturers of supposedly engineered structural building products. Notice I didn't use the Americanism pre-engineered, this word itself probably stems from part of the problem. Civil engineers typically provide services for the custom design of one-off products. If live in an industrialised society probably got a million and one complaints about all the infrastructure provided: that is focus is not on the benefit provided but the deficiencies and inconveniences which have to be tolerated. Put simply the designs are inadequate. These are multi-million to billion dollar projects we are stuck with once constructed: the systems are extremely difficult to fix or get rid off. Quality design cannot be achieved by merely crunching numbers and complying with codes of practice. Quality design requires imagination, ingenuity and a certain amount of paranoia.

Now when engineers dump these mega-systems into our environment, their experience is developed from one mega-system to the next. As they learn more deficiencies in design become fewer, but the more that has to be learnt and done under the supervision of the experienced before an individual can start to practice: for society becomes less and less tolerant of the defects. Now consulting engineers are not really operating in a market based economy in the same way as manufacturers do. Manufacturers supply engineered products which either do or do not satisfy the needs of the end-user, supplied at the right time, in the right condition for the right price. Consultants supply a service (a product), and largely irrespective of the quality and suitability of the physical systems supplied. The community criticises the systems, but continues to employ the same group of consultants to produce more of the same: possibly hoping, fingers crossed, that they are gaining experience and the next system will be better. Whilst the systems often suffer faults in workmanship, that can often be directed back at the design, the design failing to be quality robust. If a design is quality robust then at the ideal, cannot help but make to the specification, and need to deliberately go out of the way to not build to specification. The steel reinforced concrete buildings which often fail during earthquakes in relatively poor countries are examples of poor quality robust design: whilst the reinforced concrete was adopted to minimise use of expensive steel, it should be noted that the steel still too expensive and scarce, and such steel was never going to make it into the concrete. Therefore to specify such buildings constitutes poor design. Whilst large scale manufacturing has adopted the principles of quality robust design (QRD), the building and construction industry has been relatively slow to adopt. Most ISO:9000 accredited consultancies, simply have systems concerned with managing flow of contract documents: not concerned with the quality of service or the quality of the designs for the systems specified.

I will make the assertion that given the lack of whole system design, these custom engineered systems in the built environment that consultants deal with: are anything but engineered. We do not refer to a Boeing 747, an Airbus A380, a Holden Commodore (GMH), a printer, a computer or a TV as being pre-engineered: they are all engineered products: or hopefully so.

Peoples typical contact with so called engineers, is when they have contact with a regulatory authority who sends them off to get an engineers report, engineers calculations or an engineers certificate. Engineering is not about meeting regulatory requirements, being registered and filling in silly forms. {Which is what RPEQ's (Queensland) and RBP's (Victoria) seem to think engineering is.}

Engineers do not have ESP or xray vision. They cannot come out to sight, kick the dirt look through walls and certify some construction as compliant with code of practice. Such activity is totally contrary to the basis of engineering, and something being engineered.

If something is engineered, then there is a technical specification describing what it is, in terms of its critical characteristics and expected performance under varying conditions. This extends to all its component parts and raw materials. There is also technical specification concerning the control of all inputs and the process of transforming raw materials into finished end-product. Engineering is always an activity prior to the existence of a product or execution of a process. Scientific observation, measurement and control of a process without prior engineering is applied science not engineering. Engineering specifies an expectation and aims to control the process of achieving such expectation.

Large manufacturing enterprises got divide into the designers and the makers, giving rise to the process of "over-the-wall-design": this is a very inefficient process, for products get specified which cannot be made. Manufacturing aimed to remove this inefficient process. The building and construction industry via the tendering and contracting process is built around "over-the-wall-design". More over the industry is now has a multitude of manufacturers who rely on external consultants: because all that engineering provides is the certificates of compliance to meet regulations.

When I asked a manufacturer about its product and the materials it uses, I don't expect the material supplier to direct me to some consultant civil engineer in another state. That is absolutely not acceptable. In similar manner I sure hope that manufacturers we work for, don't go around declaring we are their engineers.

If producing an engineered product then it needs a technical specification as previously described. If have an engineered product then it required engineered components and materials, ecah of which requires a technical specification. If supplying components or materials for an engineered product, then your product needs to be an engineered product and have a technical specification. That technical specification needs to identify all the physical properties, all the characteristics necessary to engineer the component or material into the larger engineered system.

If make steel, aluminium, glass, bolts, gears, chains, cables, nets,  etc... then if wish such products to be used in larger systems to be engineered to be compliant with codes of practice and regulations, then all these products require a technical specification. Further more require more than a simple certificate declaring product is compliant with some code: the business needs to be able to provide evidence that it is capable of making product compliant with a code. Such evidence needs to be generated on an on going basis, to verify that the process capability is being maintained. For the benefit of external parties, this capability should be audited on a regular basis by an external party, to witness and verify the claimed capability.

Certificates of compliance and/or conformance, professional licenses, along with permits of regulatory approval all largely represent worthless scrap paper. They all fail to provide a reliable indicator of an engineered product.

Take glaziers they install glass and certify compliant with the glass code AS1288. The certificates are worthless. The requirement of the installation is that it is compliant with the Building Code of Australia (BCA), in particular the structural provisions, which requires compliance with the deemed-to-satisfy provision in the form of the loading code AS1170. The glass code (AS1288) is a design code, the loading requirements to AS1170 first have to be determined before a design-solution can be selected from the available design-charts or otherwise determined from first principles. The certificates issued therefore are inadequate.

The Building Code of Australia (BCA) is largely independent of jurisdiction. That is anyone anywhere in the world can design a building product with the intent of achieving compliance with the BCA. The BCA itself contains some criteria for demonstrating compliance. The requirements are basically centred around providing adequate evidence-of-suitability. Ultimately such evidence would be submitted in a court of inquiry, as a consequence such evidence needs to be reliable and capable of independent verification. A certificate from someone in authority is not adequate evidence. A certificate has to be more than someone declaring the product to be good stuff. A certificate should at the minimum make reference to the technical specification of an engineered product.

Cannot simply declare the glass is to AS1288, or that the steel shed complies with AS1170 and AS4600, such declarations are worthless. The glass has to have a technical specification, the glass has to have specific mechanical properties such as elastic modulus, yield strength and ultimate strength, the chemistry of the glass may also be important. If there is no technical specification then there is no evidence the supplier can produce a glass installation which complies with AS1288. The materials namely the glass is an unknown entity being installed into a still larger assembly of unknown characteristics.

For the steel sheds people running around trying to get lighter frames. This may be an acceptable thing if there was a high level of consistency in the engineering of the product, but there is not. Those consulting civil engineers, I referred to at the beginning, simply issue some simplistic often highly deficient calcs-for-council, which will be used to dump a product into the built environment at the rate of a 1000 or more per manufacturer per year. At least with consultants custom designs there is some potential for improvement from one building to another. With other engineered and manufactured products, there is on going engineering to improve the product and better satisfy the needs of the end-user. Not so with manufacturers in the building industry, no in-house engineering, just rely on external consultants, and have no proper product design or proper management of design and engineering requirements: the engineering is always at the last minute when hit a regulatory requirement for more information. That is meet a problem with an individual project. In the main all these problems are due to poor engineering and inadequate evidence-of-suitability.

Regulatory approval, such as development approval, is only concerned with checking that the proposed development complies with regulatory requirements to the extent assessed: it does not mean the development is actually suitable for purpose or 100% compliant.

The BCA has building-solutions and alternate-solutions. However the BCA is also highly lacking in defining the critical quality characteristics of a building and its component parts. In fact for a building code, it has very little content relating to buildings. Approval of new building materials for example is highly dependent on the individuals involved rather than the actual code. So straw bale, rammed earth, SIP's, ICF's, plaster wall panels, aerated concrete panels, polystyrene wall panels all problematic in terms of getting local approval. Yet its not as if the BCA deemed-to-satisfy materials like clay brick, concrete, steel, timber and glass actually have desirable characteristics in their own right as building materials. For example brick not particularly good, its soaks up water, that is why it is typically part of a cavity wall system.

Now when it comes to structural provisions of the BCA, it should be noted that AS1170, AS4600, AS1288 and all other Australian standards related to structures are all deemed-to-satisfy provisions. It is not necessary to use AS1170, it is necessary to meet the BCA performance criteria for structures. Deviating from the deem-to-satisfy provisions moves into BCA alternative solutions. Alternate solution or deemed-to-satisfy building solution, either way the fundamental requirement is to provide adequate evidence-of-suitability.

For a product to be engineered however it has do do more than simply meet BCA criteria, most especially at the end of the day, the performance criteria take precedence over the deemed-to-satisfy provisions, and when failure ultimately occurs the communities expectations and perceptions of what is considered fit-for-function over ride mere code compliance.

In terms of cold-formed sheds there is much that can be done to reduce frame size.

1) Adjust risk and life expectancy: {this changes: importance level, mean return period and wind speed}
2) Reduce internal pressure coefficients
3) Distribute wall wind pressure with height
4) Adjust serviceability requirements
a) Change serviceability risk and life expectancy
b) Change internal pressure coefficients
c) Change deflection limits
5) Allow stressed skin, or diaphragm design
6) Allow plastic or collapse method of design
7) Have monitored system with warning devices
8) Design mode of failure to be minimum hazard. (eg. failsafe)

Whilst such is possible there is requirement to justify the options taken, identify what constitutes fitness-for-function and suitability of purpose, then provide the evidence which demonstrates how the expected level of performance is to be achieved. It does not have anything to do with the BCA or any other national standard: providing such evidence is fundamental to science based design or design-science. The BCA and other national standards simply provide one set off criteria, one subjective judgement as to what one group of individuals considers to be suitable for purpose.

Something which merely complies with AS1170 and AS4600 is not fit-for-function. The codes only provide requirements for loads and the capacities of primary structural members, the structural capability of the components which enable those members to be connected together are largely ignored. The capability of the member is entirely dependent on the capability of the connections. In general the assessment of connections and the local effects on attached members is relatively inadequate. The engineering can be considered less than it should be, and the proposed structure does not comply with the structural performance requirements of the BCA, though it may appear to comply with the deemed-to-satisfy provisions AS1170 and AS4600.

The independent technical assessment required by the SA Development Act is entirely dependent on the knowledge and experience of the people making the assessment. Whilst some of these people are less competent than desired, the system is far better than having self-certifying RPEQ's and RBP's who never get questioned about not knowing what they are doing. When their self-certified products arrive in SA, then the certificates tend to be rejected, the agents for the product then seem to experience difficulty getting appropriate information to submit for approval and approach local consultants. If the products were properly engineered this would not be a problem.

It should be noted this is a different issue the other way round. When products from SA enter the other states, the issue is different: the suppliers have engineering calculations and drawings: what they don't have is an appropriate form filled in by the appropriate person. A good captive market those RPEQ's and RBP's have in Queensland and Victoria, but some contention of designing for the site and filling in silly forms does not constitute engineering. They seem to be getting away with dishing out high priced low quality rubbish. Registration, and licensing not very useful: not quality robust.

Certificates are not good enough, nor are calculations and drawings. Everything has to be part of an engineered whole, for it to be engineered and quality robust. Engineering is dependent on prior knowledge.

Now for some greater confusion. Has stated in prior posts, objective delete the word: engineer, and engineering from vocabulary. Since I have an idealistic perspective that engineering pushes forward the frontiers of science and technology. Engineering operates in a chaotic wilderness of uncertainty.

Going back in time, have an ancient building a stone wall, it gets so high then it collapses. The wall has to be built, resolving the problem was the work of an engineer. Designing stone walls is no longer the work of an engineer because the problem has been solved. To resolve the problem, the process is placed under observation and measurement, controlled and compared against an hypothesis regarding expected behaviour. The hypothesis is validated and the project is completed. The acquired knowledge is added to the body of prior knowledge. Attempting to solve problems as bump into them, is very expensive, wasteful of time and material resources, hazardous and otherwise highly inefficient.

It is preferable that prior scientific knowledge is available at design time, and all potential problems resolved and prepared for during the planning and design stage. For the vast majority of things this is possible, problems typically encountered are due to a lack of science, planning and design.

Applied science is important it makes observations, measures and collects data about systems, natural and artificial. Applied science may develop theories from the data to predict future behaviour of a system. Engineering science, adds the applied science knowledge to the engineering body of knowledge and expands and develops with respect to specific technological applications. For example idealistically, if we must have a division: then engineers designed and built the steam engine, applied science investigated and developed the science of thermodynamics, engineering science adopted thermodynamics and used it to further develop steam powered machinery. This further is added to the body of knowledge of technical science which enables the design of a steam powered device for a specific purpose.

Without a system developing the body of knowledge of technical science, just bumbling around in the dark hoping a product will achieve a desired level of performance.

So we can assess the performance of connections in hotrolled steel because the ASI and other organisations have sponsored research, either physical or literature, to compile and publish guidelines for connection design. No such research has been done for cold-formed steel design, not by industry organisations, nor public funded organisations for the common good, nor by businesses constructing products involving such connections. Everything is assumed ok on the basis of an authoritarian paper shuffling system.

Low quality imported steel, low quality imported bolts and other such issues are not a problem for a fully engineered product, produced by an engineered system. Nor does a lack of coordination, become a problem, just because of increasing availability of manufactured products to the building industry. So there is an increasing retail sector to the industry: but whether choose to make or buy still require a technical specification. Either a technical specification is sent to the supplier for them to comply with, or the supplier provides a technical specification against which the required specification can be compared to see if the product will be suitable. This should not be an exercise of paper shuffling  signature collecting, and identifying who to blame in future.

Manufacturers of engineered products should have technical personnel on staff. At the very minimum they should have adequate knowledge of the technical science for the product to be able to write technical specifications for the product. They should not rely on external consultants for design of their product, as generic as it may appear to be, unless they can properly manage the requirements for ensuring they have adequate evidence-of-suitability. The evidence-of-suitability relates to the product in its own right, and to the product as it relates to an end-users specific needs. An M16-PC(4.6) bolt may be suitable in its own right,  but from the technical specification for the bolt someone needs to determine if the bolt is suitable for a specific connection and thus provide a technical specification for the connection.

The technical specification and associated evidence-of-suitability all have to be adequate to the extent, that someone can independently verify the suitability with no other information beyond readily available publications. In terms of regulatory approval they are not going to conduct physical testing themselves to verify test results, the test results therefore have to be reliable on their own terms.

It should be noted that the onus is always on the proponent to defend the suitability of their proposal for development approval or any other regulatory approval. If going to test to demonstrate suitability, then should get that test witnessed by one or more independent parties: or otherwise send samples to at least two independent test laboratories. Also try to avoid having it tested and certified to the code: such is irritating when simply just trying to determine the products capability and all the lab can do is produce stock standard reports declaring compliant or non-compliant. Need to determine the products capability, its suitability for any specific purpose and compliance with any specific code is another set of issues: and not entirely with in the capabilities of the personnel of the test laboratories to judge.

We do not have a robust techno-scientific culture, and we need to develop such culture if we are to avoid some serious disasters in the future. The issue is not whether the product looks alright today, but whether the non-visual characteristics of the product will function as expected at some distant point in the future.

Sun 2012-Oct-21  02:12AM