Performance Calculations...

[Piltdown Man]

Calculation of performance is one of those things we have to do every day, either using short cuts designed by very clever performance engineers or long hand using a combination of absolute numbers, tables and graphs. The latter always appear to be real bugger's muddles and of questionable parentage. My question is this, how are these long hand tables and charts made? Are they made from absolute numbers obtained from test flights and subsequently turned into formulae or do we get blocks of interpolated numbers derived from these flights which generate numbers which have been "breathed upon"? Or is another method used to generate the charts, graphs and tables we use? Or is this a "how long is a piece of string?" question?

PM

[decurion]

The performance data are based on flight test results or on models that have been validated against flight test results. It depends where you look at. In the AFM all data are (flight test) certified. If you look at the FCOM (or AOM) things can be different. Although is it is based on the AFM there can be advisory performance data in the FCOM which are not certified and often not based on flight tests but on engineering models.

[john_tullamarine]

Bit more detail ..

(A) OEM

Engineering calculations generate a starting aerodynamic model using fairly standard equations of motion, engine manufacturer data and best guess aerodynamic characteristic constants for the design. The smaller OEM may then proceed to a flight test article while the larger OEM will spend a lot of time, effort and money in wind tunnel modelling to refine the characteristics for the aerodynamic model. It is interesting to note that the Wright brothers spent a lot of effort with simple wind tunnel modelling, as I recall, from the back of a truck at highway speeds.

One or more flight test articles then develop the aerodynamic model, refining equation constants etc. At the end of this process, the aerodynamicists will have a set of equations which can predict the real world performance with a high level of accuracy. Generally, with a suitable amount of windtunnel preparatory work, the aerodynamic models don't suffer wholesale reworking, rather more a case of tweaking the equations here and there to make the output reflect the real world flight test observations and measurements. ie mathematical elegance is a nice starting point but, at the end of the day, we want accurate numbers and whatever empiricism is required to make up for model imperfections is all part of the game.

These models are then used to develop the approved flight manual performance data as part of the certification process. Whether the data is presented as tabular or graphical data is a matter for OEM preference.

As decurion observes, additional (often regulator non-approved) OEM data may not have the same extent of flight test verification behind it but, in the general process, such extrapolations will be reasonable for operational use.

(B) Operator

The operator generally uses a combination of basic engineering work and the AFM to develop flightcrew performance documentation.

For the end product numerical data, a fairly rigorous adherence to AFM data is adopted as flightcrew use data may well have to withstand legal scrutiny in the event of an enquiry ... one would feel a tad left out in the cold if the operational data were to be shown to be unconservative with respect to the AFM data at such an enquiry.

Generation of numerical data usually follows one of the following paths -

(a) use of an OEM provided computer program based on the AFM models. This will be the most accurate way to go. Mutt is our resident guru on the OEM programs as he gets to play with a respectable sample range of OEM products.

(b) the ops engineer develops a homegrown computer model of the AFM performance model. General process is -

(i) read a BIG bunch of data points from the AFM charts or use the AFM tables

Two fairly standard ways of proceeding then - either

(ii) set up a suitable data lookup and interpolation program, or

(iii) run a laborious series of regressions to determine lots of equations of line segments from which a more accurate interpolation process can be developed. Note that multivariate regressions generally are of not much use as the aim is to match the AFM data closely.

Either approach then uses input data relating to the runway to calculate a MTOW for each limit being considered. This is just an automation of doing the AFM calculations manually. A lot of effort goes into setting it all up but then it fairly hums along ... OS and I have done more than a few of these exercises over the years.

(c) for smaller operations (ie only a handful of runways) the effort involved in computerising the AFM performance may not be warranted and the generation of RTOW tables may be done wholely manually (although a series of intermediate crossplots usually will be generated to save time and improve repeatability/accuracy).


Whichever of (a), (b), or (c) one adopts the end result is a bunch of tabular data which will be presented in whatever format is flavour of the month for a particular operator.

The other variation on this is to produce generalised (and simplified) performance data for flightcrew use to determine a conservative RTOW figure. However the processes are largely common.


There are a few areas where draughting artfulness comes into play with respect to the generation of carpet graphical presentations (3D graphs, in effect) but the rest of it is pretty ho hum, tedious and mundane stuff.

Ops engineers are not necessarily bright folk ... although some certainly are ... More important criteria are methodical processing, sound attention to detail, and good housekeeping


If these posts don't answer your question in the way you were after ... please do add a few more questions and we'll give it another go.