I am a student in mechanical engineering with a passion for aerospace engineering, especially airplane design.

However, coming from a more mechanical than pure aeronautical background, I feel that even now, doing a masters in what is essentially conceptual airplane design, I am lacking a certain foundation in the methodology of airplane design.

To remedy this I have recently obtained the very well-known books on airplane design by Roskam and Raymer and have started working through these books.

My question is: In the modern age of airplane design, where we are starting to get interested in more novel and radical configurations, are the approaches presented by Roskam and Raymer (and I am sure there are others authors) still valid?

To summarise this design methodology I am talking about:

  1. Determine payload weight
  2. Guess aircraft maximum takeoff weight
  3. Estimate fuel fractions for different mission segments from historical correlations
  4. Calculate tentative empty weight
  5. Determine allowable empty weight by taking advantage of the fact that, historically, there is a logarithmically linear relationship between maximum takeoff weight and empty weight
  6. Compare tentative empty weight and allowable empty weight, if the values are too different, guess a new value for maximum takeoff weight and continue this iterative process until convergence

This design procedure uses historical correlations in the fuel fraction determination and through the historical trend in relationships between MTOW and OEW.

Is this approach still valid for novel configurations?

Most importantly, is it known whether this logarithmically linear relationship between MTOW and OEW independent on configuration?

  • 1
    $\begingroup$ I suppose the methodology would still work, but these days the constraining factors on aircraft designs are often not "aerospace engineering" issues but human or infrastructure ones: The 787 could have been faster, but it needs to meet noise regulations (people don't like loud planes flying overhead). The A380 could have been bigger, but its size already causes trouble operating at some airports. The improved design technologies (e.g. simulation) that Peter mentioned help contribute to working more of that out on paper before building the flying prototype. $\endgroup$
    – voretaq7
    Commented Aug 19, 2015 at 21:38

2 Answers 2


Many studies and presentations you will encounter about radically different configurations with much higher performance are the aviation equivalent of car commercials. Someone tries to get attention in order to get published, or she/he needs to convince investors to fork over their money.

Trust me, the aircraft designers of the last century were no idiots. They had the luck of living in a time when it took only a few years for aircraft designs to become obsolete, so they could go through the whole design process many times and apply what they had learned before.

Someone graduating from university today will be lucky if she/he gets the chance to go through every step of a major design process even once. I myself have brought three aircraft into the air and consider myself very lucky. I had colleagues who retired without ever having seen one of their many projects take off (literally).

The consequence of this is a very mature industry which has found an optimum configuration for most purposes already and is only tinkering with the details. Add to this an ever increasing thicket of regulations which have grown on the experience from existing designs. Sure, the progress in computer control is impressive, and engines get better all the time. Manufacturing precision is ever increasing, and materials still become better and more consistent. But overall, we still build aircraft like we did a generation ago, and I do not expect this to change radically in the future.

What impacts the design process most is certainly the incredible precision of simulation software, which allows to see details which remained hidden in a wind tunnel or a strength test. Things which had to be done consecutively can now be done in parallel, and the precision of knowledge at an early stage of a design is much higher than before. But at the same time aircraft become ever more complex, so the advantage of better simulation is eaten up by the increase in complexity. Add to that the fact that mostly beancounters will have the last word, where before engineers could determine how the work was done, and you will understand that the level of preparatory work is shockingly inadequate in modern aircraft companies. Many of the avoidable early mistakes will require late, expensive fixes (but by that time the stingy beancounter has left the company with a huge bonus for the money he bragged to have saved).

The design process is still the same, and where before the experience of a few people helped to find an overall optimum, large, integrated teams will together determine this optimum. The inexperience of the participants will only show up in delays, when the iterative design cycle needs to be repeated more often than anticipated, but the result will be of high quality - really bad designs are a thing of the past. But cutting corners will still have consequences, so small annoyances will still plague new designs. Consequently, most of the engineering work will involve ironing out these annoyances or retrofitting new systems to existing airframes.

To answer your specific question: Yes, the relation between OEW and MTOW will still hold in the future, and radically new configurations will show an advantage only on paper, before they have been thoroughly designed and test-flown. Better materials and methodologies will help to improve the ratio between OEW and MTOW, but this advantage risks to be eaten up by the desire to add bells and whistles everywhere.

  • 1
    $\begingroup$ @Jonny: It is good practice around here to accept an answer only after two or three days, so you have a choice of good answers and don't discourage others to add their view. $\endgroup$ Commented Aug 19, 2015 at 20:29
  • 1
    $\begingroup$ Noted. Also after asking my question I realised that I might evoke some opinions rather than clear answers, which is also a bit against the grain of SE. But I think your answer was specific and technical enough (-: $\endgroup$
    – Jonny
    Commented Aug 19, 2015 at 20:39
  • $\begingroup$ What does it mean OEW ? $\endgroup$ Commented May 17, 2017 at 8:37
  • $\begingroup$ @d.pensopositivo: Operating empty weight. This is structure + systems + pilots + equipment, but no fuel, lubrication and payload. Also called OWE (operating weight empty). $\endgroup$ Commented May 17, 2017 at 13:35

To add to @PeterKämpf's answer: the basic design methodology will be unchanged so long as the laws of physics remain unchanged. But the method of application of the methodology is changing very fast, and will continue to do so.

As a mech eng student, the OP is unlikely to have any real-world experience of the engineering environment in which Roskam's 8-volume set of books were first published, in 1985-1990 (according to Wikipedia). That is 25 to 30 years ago. As just one example, the fastest computer in the world back then (the Cray-2) had a clock speed slower than any modern PC except for cheap laptops or tablets, and less RAM than an top-of-the range modern cellphone. And with a price tag of around $20m, that computing resource would be shared between hundreds or perhaps thousands of users, not sitting on an engineer's desk for his or her exclusive personal use.

Today you don't need to "estimate fuel fractions for different mission segments from historical correlations". You can look at the accurate real-time data being collected automatically by every modern aircraft that your company has built, and look at it for negligible marginal cost. You don't need to approximate from a "historical log-lin correlation" - again, you can look at the actual data, in as much detail as you need.

This accelerating rate of change presents educators with some serious challenges. The university engineering departments don't have access to all that real-time data, and even if they did they are (correctly, IMO) trying to teach some general principles, not the specifics of what Company X does this week. But the danger is that the students work through the case studies in an excellent collection of 25-year-old books, and without any other input, they get the wrong idea that those case studies still represent the current state of the art.

  • $\begingroup$ Thank you, very valuable insight. I suppose that as someone wanting to work in the industry it wil nevertheless be valuable to get grips with the methodologies, whether their implementation is the same as in the days of Roskam or not. I know that both Roskam and Raymer currently market design codes based on the methodologies and methods presented in their books. $\endgroup$
    – Jonny
    Commented Aug 20, 2015 at 5:30
  • $\begingroup$ So they must still be valid (the methods as well as the implementation). I think one should realize that not all aircraft manufacturers are the size of Boeing, Airbus, Embraer or Bomabrdier etc and that many of the smaller companies still us these methods and correlations as is. Of course this is just my thought, not even an opinion. $\endgroup$
    – Jonny
    Commented Aug 20, 2015 at 5:31
  • 1
    $\begingroup$ The term “historical data” does not mean the data is old, it simply means data collected from previous actual use as opposed to data calculated or measured in a specially constructed experiment. Of course you'll use the most recent data available. Also detail is not always what you need; it just gives you more accurate estimate of the “historical log-lin correlation”. $\endgroup$
    – Jan Hudec
    Commented Aug 20, 2015 at 6:00
  • 1
    $\begingroup$ Add to the improvement in computing power the improvement in algorithms (which, by the way, has been bigger than that in computing power over the last 30 years in a number of applications), and you can see what the change has been. But it still needs an educated mind to understand the results, no matter how pretty the colorful output will be. $\endgroup$ Commented Aug 20, 2015 at 6:11
  • $\begingroup$ What does it mean IMO and OP ? $\endgroup$ Commented May 17, 2017 at 8:37

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .