Engineering Calculations and Approval

{This is a edited version of an email I sent this week (2019/week 14) attempting to explain nature of calculations and the approval of calculations. The problem I was attempting to deal with is the Queensland legislation. Being edited it has some client specific references removed, and also extends on what I was attempting to explain, with some extra formatting.}

My understanding is that: in Queensland there is a legal requirement that calculations are prepared by an RPEQ or under the supervision of an RPEQ. I also recollect reading somewhere that there is no over-the-shoulder clause which is common for defining the licensed engineer responsible-in-charge.

This requirement does not prevent other people from doing calculations, if you are an owner-builder and so inclined you could do your own calculations. However, your proposal, as shown on your drawings and written specifications would still need to be assessed by an RPEQ. The RPEQ is the approved person (AP) for assessing compliance, it is their opinion the regulatory system gives value. Now whilst they say they do their own calculations, the extent of their own calculations is entirely dependent, on the quality and usefulness of any calculations they receive.

More importantly irrespective of local legislation, we can identify two broad categories of calculations: design-calculations and proof-calculations. Furthermore expect consideration of the following sets of calculations for any given project:

  1. Design-Calculations
  2. Proof-Calculations (Designer)
  3. Proof-Calculations (Independent Reviewer)
  4. Proof-Calculations (Regulator/Approving Authority)

Design-calculations are iterative as design involves trial and error until find a design-solution. Proof-calculations verify that the design-solution is actually a solution. Design-calculations if documented contain all the ways it doesn’t work, Proof-calculations only contain the one way it does work. Proof-calculations by the designer can be considered stage:1 in the chain of evidence-of-suitability.

The designers proof-calculations are the calculations submitted to the regulator (council/RPEQ/CP.Eng), however if the project is critical or new technology, then the designer may seek an independent review before submitting calculations to the regulator. The review is not a check of the calculations, it is a review of the specification, and calculations are done from scratch: either they verify the specification is compliant or refute compliance. The designer and independent reviewer then defend their viewpoints and debate the issues until they reach agreement.

Similar process then occurs when the proof-calculations are passed onto the regulator. The regulator also does their own calculations based on the specification, they do not start by reviewing the submitted calculations. This can be considered as stage:2 in the chain of evidence-of-suitability.

However, if the submitted proof-calculations come from a reliable source, then the regulators proof-calculations are minimal and simple. The regulator can advise the designer to check further issues, and receive additional proof-calculations. Ultimately the regulator can adopt the submitted calculations as being prepared under their supervision and take responsibility for approving the specification on the basis of the submitted calculations. The lower the quality the submitted calculations, the more calculations the regulator has to complete.

So even in South Australia, whilst the councils could accept the calculations from product configurator software, the certifying engineer (CP.Eng) would still be expected to do their own calculations. The extent of such calculations would depend on how reliable they consider the software to be. Software for nail plated roof trusses wasn’t considered very reliable when it came to connection design. Now, if software complies with the Ministers specification for structural software or the ABCB protocol for structural software, then can accept the calculations hidden behind the scenes as acceptable. But the certifying engineer still has to do a minimum level of calculation.

So yes the RPEQ will do their own calculations, but the level of the calculations they do entirely depends on the quality of calculations they receive, and the extent to which they can consider the preparation of the calculations to be under their supervision. If the software is reviewed by an RPEQ against the ABCB protcol for structural software, then the RPEQ only has to do the minimal number of calculations to confirm the software output is relevant to the building proposal. But there is no way that they will be considered to have exercised adequate “duty of care”, with zero calculations of their own.

From my viewpoint therefore your product configurator software has to be good enough that it produces all the calculations I expect to do. Everything it doesn’t do, I have to do calculations for myself. The more calculations I have to do myself, the greater the delay in certification of the individual project. The same would go for an RPEQ: product configurator software has to be good enough that they would consider buying it and using it.

All software, and reference books, engineers use have disclaimers about its reliability, and the responsibility the engineer has to check the information. So I could rely on the AISC/ASI design capacity tables (DCT’s) for hot rolled sections, but not until I have personally verified to my satisfaction that the tables are valid. I don’t have to check all the values, just do random checks. Likewise, frame analysis software like Multiframe, Microstran, Spacegass, I have to check and verify to my satisfaction that it is producing the appropriate results. Some things I cannot check because they are dependent on experiments I cannot repeat. But every building is a real world experiment, sat there waiting to experience the design load, and if I understood all the theory correctly then the structure will resist in the manner I expect.

If you are a manufacturer of manufactured structural products (MSP’s), does your company have or intend on getting professional indemnity insurance to protect you and your buyers from errors in design? Where “design” is carried out by software developed in-house .

I doubt that RPEQ’s do full calculations from scratch for nail plated roof trusses using general purpose structural software. It would take far too long, and cause unreasonable delay in approval of housing. There is thus precedent in accepting output from software.

More over my experience of the Queensland system, over the past 20 plus years, is the self-certified rubbish spewed out of software from Queensland manufacturers. In SA we don’t allow self-certification, and don’t care if its by an RPEQ, self-certification is just unacceptable. The supply agents have hassles getting valid calculations from the Queensland companies to submit to SA councils. This covers such structural products as aluminium balustrades, cold-formed steel sheds and canopies, light gauge cold-formed steel house framing. The problem with the software is that it just prints a specification or cutting list which is declared to be code compliant: there are no proof-calculations and so inadequate documentary evidence-of-suitability. Thus evidence that some RPEQ’s are not doing detailed calculations on a project by project basis.

My expectation is that if product configurator software developed in-house, is independently tested and certified against the ABCB protocol for structural software, and it is clear that the end-users understand the relationship between the calculations and the building proposal, then RPEQ’s can accept the submitted calculations with minimum check calculations on their part. That is they can accept the results are as valid as the reports from Multiframe, Spacegass or the ASI DCT’s, and as acceptable as if they got a junior engineer under their supervision to prepare the calculations for them.

Note that whilst approved software is limited to the building constraints in AS4055 (eg. 16m wide and 6m eaves), it doesn’t stop the software from having a password protected superuser mode. So can design in-house but subject to additional checks for building approval.

In SA a “building surveying technician” can accept the output of approved software, for the larger buildings a “building surveyor” would have to grant approval and the structural calculations would need to be reviewed by a chartered engineer (CP.Eng). A lot of councils now only employ building surveying technicians and out source to private certifiers. So small simple buildings get approved faster than large buildings. For commercial buildings, spans of 15m seem more common than spans of 18m. For larger rural buildings, would need a different structural form, or different materials, and therefore different calculations: at the very minimum need to be able to reinforce the c-sections with flat bar.

Personally I don’t see the reason for the dimensional constraints. The limit should be the natural limit of the structural form and materials being used. No limits need to be set as the calculations conclude the proposal is not feasible.

The point to note is that design-calculations are not subject to legislation, only proof-calculations for granting approval to “go ahead and build” are subject to legislation. The legislation changes from state to state, and country to country. Design-calculations can generally remain the same, but proof-calculations have to change to meet local legislation.

To illustrate I don’t have to design to AS4600, I can use American software to AISI specification, but to get approval I have to demonstrate compliance with the NCC/BCA. To the NCC/BCA, AS4600 is deemed-to-satisfy, it is not a requirement, I therefore have to demonstrate the equivalence of AISI and AS4600 or account for the difference, and demonstrate I have still met the NCC/BCA performance criteria. Its more an exercise of playing technical lawyer than engineer.


Revisions:

  1. [08/04/2014] : Original (email)
  2. [12/04/2014] : Original Published