Chapter 18  Introduction to engineering practice


Read as one text






The concept of recognised science




The scientific hierarchy


The mechanical engineer


Looking for the science that is pertinent to engineering


How can one learn both theory and practice and link the two?


The science of the “logicians”


Chinese whispers


Mathematical modelling and solving by trial


Engineering drawing


Engineering handbooks


Engineering artefacts









Chapter 18  Introduction to engineering practice


This book has been written and published on the net as I have gone along. It was inevitable that I would learn in the process. When I started writing I did not have any firm ideas about how I would handle engineering practice and now I find that I need to stop and prepare the ground carefully for the rest of this book. I decided to make this preparation into a chapter in its own right.


In section 1 of this book I have been gathering together part of what I can only call recognised science. Recognised science is founded upon Newtonian mechanics as taught in degree courses in physics and the adjective “recognised” means that it is generally agreed, by those who use it with understanding, that it is correct because it has not been found wanting when used for any of the important things that it purports to explain. What I have collected together in the first part of this book is a selection from recognised science that is relevant to mechanical engineering and by relevant I mean that it can and is used with confidence in applications of mechanical engineering. I do not think that I have included anything that is not relevant and, for each topic, I have chosen the science that I find to be the easiest to understand.


Those applications that I have considered are fairly obvious ones like flow in pipes and channels. They are in a sense the old natural philosophy in that they occur in nature in some way like flow of blood in arteries, water in river beds, high velocity flows resulting from natural event, fluid pressure etcetera.


Now I want to turn to applications like pumps, fans and turbines, aeroplanes, wind tunnels, surge tanks and so on and these are all artefacts that have been designed and developed by engineers and often refined using our recognised science. Clearly I am moving into the realm of engineering practice and that is more complex in character than one might imagine.


When I was a student there were textbooks entitled:- “the Theory and Practice of ……..” and it might be of steam engines or of hydraulics. I supposed that the practice would illustrate the way in which the theory had affected the design of, say, steam engines. This supposition on my part was in error because mostly the practice preceded the evolution of the theory that might have affected it. The fact is that often the theory that might be applicable to some mechanical device and the design of the device have developed independently even if they developed at the same time. To this day I am grieved that I spent so much time learning the theory of the steam engine only to find that real steam engines were so constrained by the possible mechanical arrangements that the theory was almost irrelevant.[1] On top of that I now know that there are many other factors that determine the practice of engineering and these need to be explained. I want to explore these factors one by one. I need a place to start and I think that this must be with Thomas Kuhn.


I wrote a version of this chapter and published it to the net with serious reservations; frankly I did not like it. I had spent months on it and I found it difficult to write because there appeared to be no core statement that threw any light on the rest. Within just a few days of uploading the chapter the work of Thomas Kuhn came to my notice and I decided to reconstruct Chapter 18 because Kuhn seems to me to have got to the heart of my problem. His work is published as “ The Structure of Scientific Revolutions” and appeared about 50 years ago.


Kuhn is concerned with science whereas I am concerned with engineering. Kuhn gets to the basic problem for me when he notes that the role of the scientist is to explain how the natural world behaves and then says that "Unlike the engineer ..............the scientist need not choose problems because they urgently need solutions and without regard to the tools available to solve them." If I may turn that round I think that he is saying that scientists can pick and choose what they choose to study but engineers must find solutions to the problems that they are paid to solve.


Kuhn said that science does not proceed, as generally supposed, by incremental stages to reveal more and more clearly the final fundamental truth of whatever science is being used to explore. Instead scientists look at the world and attempt to interpret it using a body of received opinion, “recognised” science, shared assumptions about their science, and expectations about themselves and the world at large that Kuhn calls a paradigm. It seems to me like a mindset and I can recall several shifts of my mindset over the years. This is a crucial observation and accounts for much of the mistaken output of science. It also accounts for the mistaken output of politicians, engineers and everyone else. I want to look at this statement more carefully in the context of engineering.


Kuhn's book is essentially about the process by which an existing paradigm is replaced by a new and better one. Several times he points out that people invest a lot of time in learning a paradigm and how to use it in, say, the design of steam turbines, and need to be convinced of the advantages of some new paradigm before they will even contemplate a change to using it. Kuhn says as often as not, new paradigms are introduced by someone who is competent to think about problems that the existing paradigm could be used to solve but, instead of using that existing paradigm, create a new paradigm to solve them. That person is often young and of course has no investment of time in the old paradigm to forego. This is mirrored in engineering.


The concept of recognised science

I used the idea of recognised science above, now I want to look at it more carefully. Let me start with Newton’s laws of motion. We are all familiar with Newton’s “laws of motion”. It is staggering to think that these laws are so simply stated in a combination of words and symbols and yet they are so accurate that they can be used to guide space probes into orbit round distant planets. Newton’s laws are clearly recognised science in that they work extraordinarily well. Indeed they are the foundation of all the science in this book. And yet, when a space probe has travelled far enough to leave the solar system and enter deep space, it becomes evident that the probe is not quite where Newton’s laws says that we should expect it to be. Then we do not know whether the tiny discrepancy is the result of a tiny error in the statement of Newton’s laws, which we regard as recognised science, or the result of some effect on the probe that we have overlooked or, of course, that we do not know for certain where it is. Kuhn calls this an anomaly and when something as widely used as Newton's laws has such an anomaly we get worried. This follows because it is much too optimistic to suppose that something that we regard as recognised science is infallible. However this tiny error has exposed our mindset that we expect to get the right answer when we use recognised science. When the "man in the street becomes acquainted with some application of science that does not go as expected he rejects the science not the mistaken expectation. But even that word “expect” can be qualified because, in the event, engineers always assess our recognised science by whether it gives us a useful result, that is, one that is adequate for our purpose and then the phrase “adequate for our purpose” can be qualified because while the result may be adequate for our purpose it may not be the best that is possible. As we have seen engineers are in a different position to that of scientists.


Recognised science covers much more than Newton’s laws. It includes the science that explains some aspect of the natural world by the use of Newton’s laws, for example, the flow of gases at both high and low speeds or pressure transients that go with changes in velocity of fluids flowing in circular pipes. It also includes the explanation of the loss of energy to friction in fluids flowing in circular pipes and this does not depend on Newton’s laws. In these cases the recognised science is by no means as accurate as Newton’s laws because some simplifying decisions have been made but, it is the best science that we have found for the problems that must be solved and proves to be adequate for the purposes of engineering.


So we have to look to see what we mean by adequate for our purpose in the context of engineering. Engineering is about making artefacts that achieve some purpose that was specified at the outset. It may be that it was a bridge that, when successful, carries traffic without falling down or swaying or doing anything else that constitutes failure. It may be an engine that is only successful if it produces a required efficiency and runs without overhaul for say 100,000 hours. But all that was required of Stephenson’s steam engine “Rocket” was that it worked better than other competing engines all of which were primitive in construction. All that was required of Parsons’ steam turbine was that it worked better than contemporary steam engines and the fact that its design was so poor was irrelevant. So the criteria for success of an engineering design are not anything absolute but are determined by the expectations of those who design it and by those who use it.


The case of friction in pipes illustrates Kuhn’s ideas beautifully. I described the process by which we gained our recognised science for pipe friction in chapter 7. The scientists proceeded on the expectation that, if they worked for long enough, there would be a revelation of a final and fundamental truth whereas Moody just wanted something that would give useful answers to problems in engineering whether it was the final and fundamental truth or not. Moody stopped the scientists in their tracks but one must accept that Moody could not have given us his chart without all the work of the scientists and, of course, the Moody diagram contains much of that science.


So there is no hard and fast definition of recognised science but I think that the key aspect lies in the words in bold above “It is generally agreed, by those who use it with understanding, that it is correct because it has not been found wanting when used for any of the important things that it purports to explain”. However we must remember that no paradigm can be so good that it cannot be replaced by something better


Now I think that most would agree that understanding Newton’s laws and indeed a lot more of the underlying “theory” of engineering is not easy and that most of us work with a limited understanding. Some people have no idea just how poor their understanding and knowledge base actually is and this leads to the dangerous position that “experts” with a limited comprehension of Newton’s laws give advice that is simply incorrect based on their limited ability. Furthermore, not knowing just how poor their understanding and knowledge really is, they write textbooks and papers and this has given rise to all sorts of spurious explanations of engineering that have, somehow, come to be accepted, by those who know no better, as recognised science. So we have a situation where both genuine science and spurious science co-exist and both are treated as recognised science by different groups. In my experience those who are conversant with genuine recognised science easily spot spurious science for what it is. The reverse is not true, and too often, genuine recognised science is denigrated as theoretical.


So we end up with technical information that is a minefield of genuine information, misinformation, that is unwittingly incorrect or misleading, and, unfortunately, disinformation that is deliberately incorrect or misleading.


It is compounded by two other problems. The first is the thorny problem of how to assess the success of an engineering design and the second is the ready acceptance without critical appraisal of many designs that are really nothing like as good as they could be if only they were to be the subject of proper engineering analysis. They both affect the shared assumptions and received opinions that constitute the starting point for many engineering designs.


This all means that Kuhn’s ideas apply equally to engineering where those who practice engineering do so having taken on a mix of information, misinformation and disinformation and use this mix as the basis for design. Add to this an incomplete understanding of fundamental physics and it is hardly surprising that we end up with designs that are accepted as the best when better designs and sometimes much better designs are possible.



I want to introduce examples of information that is incorrect whether unwittingly or deliberately.


Over decades if not centuries or millennia artisans, that is people who make things, will have pondered on how the things that they made actually work. Let me take a real example. Around 1927 Manfred Curry, a medical doctor cum scientist, started to win yacht races with a rig in which the foresheet overlapped the main sheet by a significant amount. The fore-sheet is now called a Genoa and is a fairly standard rig for cruising yachts and for racing yachts. Curry published his explanation of the way in which the Genoa worked. He likened it to the operation of an ordinary tap but it was hopelessly wrong. Gentry, in 1981, offered a much better explanation but Curry’s explanation will never be superseded because it is in print all over the place. One must be charitable and allow that Curry did not set out to deceive but the net result is that we have a spurious explanation looking to the unwary just like conventional science and still being taught to students as part of a “recognised” science of sailing. I need a name for this spurious science and in the context of this book I will call it pseudo-science knowing that many others have used the same words for their interpretation of science.


But what we find when we look for explanations of the modus operandi of engineering devices is often a mix of recognised science, pseudo science and all grades in between. It is very complex and inevitable because almost all of our most important machines were developed long before the appearance of the relevant science and this leaves us with an uneasy “retrofit” of the new science and the explanations, spurious or otherwise, that had formerly been used to develop the machines in the first place. In most cases, when the new science was used to rework the old methods, it was used rather too enthusiastically and many engineers who were happy with the old ways but might have changed, were put off by the apparent complexity of the new way[2] and did not change. This last statement even applies within the science as, for example, the reworking of the science of supersonic flow in a convergent-divergent duct to change from using ordinary thermodynamic notation to using Mach numbers.


This is the situation in engineering and the extreme example must be sailing where Newtonian mechanics and the methods of aeronautics might as well not exist. The recognised theory of sailing is quite impenetrable to me and so is naval architecture for that matter. Heated arguments about whether a piece of fabric attached to a frame of poles and ropes is a wing or a sail seem to me to be quite irrelevant. There would be no argument if the ordinary rules of fluid flow were to be invoked. However sailing goes far back in time and has had plenty of opportunity for its curious version of science to evolve. Someone trained in engineering science soon runs out of patience with this sort of hybrid science because it is so different. The two do not appear to have anything in common. It is hardly surprising that lecturers in engineering do not feel at home with pseudo-science or with practice that does not fit with what they are lecturing.


So let me return to pseudo-science, ie to text looking like recognised science but in fact wrong. When I was lecturing I had no idea that this pseudo-science existed. I ran into it for the first time when I took up model yachting and looked for the science of full-sized yachts only to find that it was wholly pseudo-science. I sought to offer an alternative based on science. It is in my books on this website called “The RC Racing Yacht Explained” and “The Physics of Sailing”. Generally it fell on deaf ears although a few have understood. It took me until I was 82 to realize just how many more full-sized engineering designs were also based on some form of pseudo-science. It is alarming to think that many of these designs are seriously in error and are taken as designs to be copied and emulated and become part of the shared assumptions and received opinions. They are the output of designers with no real understanding of science who are using their pseudo-science instead and these designs are everywhere. Search on the internet the words “ducted fan” to see what misguided devices are taken to be standard. (Perhaps you will not even know why they are wrong.) Anyone learning about engineering must always be on guard when studying engineering artefacts; those artefacts may be very misleading. What is wrong with the design of a vacuum cleaner that it needs 1,000 watts[3] to make it work? How does it happen that the speeds and sweepback angles of modern fighter aeroplanes do not fit with science? How does a new power station using steam turbines manage to have a significantly lower efficiency that the one it replaced? Why are bulbous bows on ships so varied in shape? How did we come to be installing wind powered generators when, if all the factors into account, their performance is so abysmal? Do all houses need footings that are one metre wide and one metre deep and why is it supposed that deeper footings are better than wider footings? Do the earth wires in domestic electricity systems need to be 6 mm in diameter? How can anyone swap the position of the jib of a Bermuda rig so that it is behind the mainsail and claim that it is an improvement? How did it happen that those who effectively voted[4] for the launch of the Challenger Space Shuttle could not see that the second O-ring was not a back up and never could have been? The list is endless.


My list above is of actual designs but the pseudo-science that has been used is underpinned by often-repeated spurious explanations in the media and encyclopaedias of the way that common devices actually work. None can be more wrong that that in books and on television for how a wing generates lift. Just look at the pointless exercise in Wikepedia that purports to explain the aerofoil: the author only draws one section in good proportion. Who could have produced the unlikely diagrams to explain how turbines work in Wikipedia? Wikepedia is in itself the very antithesis of what it was meant to be. It appears to be written by historians in a straightjacket of a format and, where technical matters are concerned, is largely misinformation encapsulating someone’s faulty knowledge of the science and the practice involved. Wikipedia is simply consolidating peudo-science because it lacks any critical element in its text. It is creating Kuhn’s “body of received opinion, “recognised” science, shared assumptions, and expectations about themselves and the world at large” and most of it, to be charitable, is pseudo-science. Anyone who can pick out the valid information from the noise in Wikepedia does not need to read it.


Let me give a specific example of what I mean by misinformation from a supposedly informed source. I wonder whether the engineers of Queen’s University Belfast have read this extract that is thrown up by Google when searching for theory of sailing. It has no basis in normally accepted science.

How a Sail Works (sort of)

 Sails convert the energy in the wind into forward movement of the boat.  They do this in the same way as wings of an aeroplane provide uplift.  The wind causes the sail to take a wing shape.  The wind passes around the sail and because the distance is greater on the leeward side of the sail, the wind must travel faster.  This means the pressure on the leeward side is lower than the pressure on the windward side.  This pressure difference results in a force acting perpendicular to the boom in the downwind direction.  This force can be resolved as a force on the boat pushing it forward and a force on the boat pushing it sideways.  The sideways force results in drift but is (for the most part) avoided by using a centreboard/ daggerboard.

A sail has a particular angle to the wind at which it is most efficient.  This is usually 45 degrees, which means that boats can travel at 45 degrees to the wind.

When moving in a downwind direction the sail no longer acts like a wing, but more like a parachute, catching the wind and moving the boat along with the wind.  It is downwind that boats travel fastest.

Last updated: 1330 Monday 17-08-1998.


This is a perfect example of pseudo-science. It is wrong or half wrong from beginning to end. For example, the last sentence is simply false. The fastest speed is produced on a broad reach. (Recently a wind-powered land yacht made 126 mph but what speed of wind would have been required for this to have been done on a run?) In my view it is a disservice to us all to offer this spurious text under the crest of a reputable University and then put “sort of” in brackets to excuse it. But the author was giving his version of the theory of sailing in good faith possibly because it is what he was taught. How does it happen that he can accept that lift is generated only when the distance from the leading edge to the trailing edge is greater on one side than the other and not see an anomaly for a sail that obviously lifts yet has no difference between these two lengths? That is what makes it so dangerous.


I do not think that I need to go on with this theme. Everywhere one looks the same thing is evident, science and every other identifiable discipline has acquired a set of ill-conceived explanations of how it works; pseudo-science. These explanations are extremely difficult to dislodge and replace with something more cogent and the would-be engineer must be vigilant at all times and think carefully about any text that he may read. The text must be examined to ensure that it really is consistent with accepted science.[5] Just knowing that pseudo-science exists is half the battle to understanding engineering. The would-be engineer must also be ready to find that designs, even from prestigious institutions, do not stand up to scientific analysis. It needs great confidence in one’s grasp of science to view such designs with scepticism. Nevertheless it must be done.


But one must be careful with this pseudo-science. Somehow, in a ponderous way, it may lead ultimately to the best design as anyone who follows the evolution of racing yachts will have observed.[6] Sometimes it leads nowhere. Either way it does so by trial and that trial could be much reduced if more science were to be used. Perversely many reject this science because they rather like doing things by trial.


However all this would not be there if there were no inevitable process to produce it and I need to look for that process. It is buried in the human psyche and in the structure of society and it has many faces.


The scientific hierarchy

Science has a hierarchy in the sense that some science is seen to be of greater status than some other science. On the face of it this is ridiculous but it is real and, of course, counter-productive. There are those who think that Mathematics is the Queen of Sciences complete with capital letters. It is debatable whether mathematics is a science at all but the implication is that sciences are inferior to mathematics. This attitude permeates all science to our great disadvantage. Whole departments in universities are viewed as belonging to a pecking order and some disciplines that are not so dependent on mathematics as others are seen to be very low in the pecking order.[7] I cannot see why a mathematician, or anyone else, should think that what a mathematician does is any more worthy than what I do when I design something. Experience tells me that very few mathematicians can do some engineering but many engineers can do some mathematics. It would be much better to regard the whole spectrum of science and mathematics as just a set of complex tools available to anyone who cares to learn those sections that are useful in pursuing the job in hand and not ascribe some superiority to mathematics.[8] There is no need for a hierarchy. However this position will never change and, as a direct result, textbooks often contain unnecessary mathematics just to preserve a perceived position in the hierarchy. This just adds to the confusion and in some ways it is worse than pseudo-science.


However, instead of this hierarchy disappearing, it is reinforced in science by another social influence on science. It is social status. There are those who say that all men are born equal and at the same time think that everyone should enjoy an above-average status in life! They appear to deny the easily observable characteristics of people at large. People naturally separate into classes, to use a politically incorrect word, and this can be seen to be true by watching people buying newspapers. Newspapers in England are called by names such as quality dailies, tabloids, popular press and so on. These names, in themselves, indicate classes and one needs only to match purchasers to their choice of newspaper to become certain that tastes differ and it is but a small step to categorise them as upper class, middle class and lower class and take the other steps to professional and working class and white collar and blue collar classes and so on. Classifying people seems to run right through history starting with group leaders from prehistory days.


This seems to be intrinsic to every society and not something to be stamped out by equality laws drawn up by people who are so important that their equality law does not actually apply to them.


My interest lies not in social classes but the link between social class and the science that is pertinent to engineering. In my introduction to Section 1 this book I used the phrase “slaves do arithmetic” and I want to revisit it. I go back to its source. I found it in John MacLeish’s book “Number”, ISBN 0 00 654484 3 that I thought was to be admired. It takes us back to Ancient Greece and I quote in red and blue :-


The Greeks made a unique, but relatively minor, contribution to mathematical logic. They tidied up the body of knowledge about geometry, making it into an abstract system knit together by deductive logic. (This was an achievement of genius by Euclid.) They created a climate of opinion (my bold) in which arithmetic was also considered as an abstract system, with no practical application. But from the standpoint of modern, practical science, these contributions had the same kind of baleful, stultifying effect as the Greek views on society - for example, on education or the status of women - which still bedevil our cultural and social values.


Arithmetic and logistics

In modern times, arithmetic means something quite different from what it meant to the Greeks. Scientists (of today) describe it as 'the language of science', a major tool in the study of nature and society. The Greeks, however, thought of it as a form of abstract wisdom, with no connection whatever with practical activity.


Calculation is clearly a mental art. It is part of a battery of skills. Children (and slaves) can be taught to use logistics in the form of drills (what we call mechanical arithmetic), with a minimum of mental content. But it can also be taught at the highest level, as in mathematical physics, and at all stages in between. Skills are used by craftsmen and craftswomen to produce artefacts of value and practical use. They relate to the real world. Physical skills always involve some mental component, for example educated judgement or the calculation of size and scale; even the most mental of skills (for example, composing music) involve some physical activity.


In this day and age, these remarks are platitudes. But they would have made ancient Greek “thinkers” laugh out loud. Calculation (logistics), for them, was a skill like fishing or barbering; arithmetic was a different and higher form of art, whose practice was for free citizens only, forbidden by law to slaves. This attitude applied even to numbers and numerals. To the modern thinker, they seem identical. But to ancient Greeks they were utterly different, their nature depending on their context and especially  who was involved in the operation. In arithmetic, numbers were regarded as abstract, spiritual entities; numerals in logistics (written exactly the same) were regarded as base, concrete quantities, with no 'existence' independent of the objects that they described. (Most 'moderns' would agree with this last statement.)


Greek scholars constantly made distinctions of this wrong-headed kind: education versus training, sport versus work, free person versus slave. These distinctions were definitions to guide the inquirer; they were linked to matters of life-style. They were also blinkers, hiding the real connections between things, and distorting every subject to which they were applied.------- It was a cardinal principle of this logic that experience was ruled out (since it came to us not from reasoning but through our senses). Experience was not allowed to correct, or even to throw light on, the process of reasoning. Indeed, it was the main function of reason to correct the false conclusions drawn from the errors imposed by our partial view of reality.


McLeish is clear and needs no further explanation from me. In a nutshell the Greeks, and no doubt other races, had some very odd ideas and those ideas are not only difficult to replace but probably still survive in some modified form today. However there are further implications. McLeish has said that who did calculation and who did arithmetic was settled not by ability but by social class. The study of the abstract was associated with the upper class, that is, free citizens, and the use of arithmetic to calculate was associated with a much lower class, the slaves. In addition the free citizens were affluent and the slaves were not affluent. It requires no great stroke of enlightenment to see a parallel with engineering science where the abstract study of engineering, wholly divorced from practice, is now seen in educational circles to be much superior to actually making something and it has acquired social status. Furthermore it can also be taught by lecturers who know nothing about engineering practice. The incentive to join those pursuing the abstract study of engineering has never been greater. Everyone wants a degree about the abstract and not a diploma about the practical even if that diploma does contain applied engineering science.


If I may put this position into a single paragraph, it seems to me that there are people who are qualified by examination to call themselves engineers who will almost certainly know lots of abstract things derived from science that might just be useful to someone who actually makes things and possibly nothing about the way in which things are made nor of the materials from which they are made. Further to that they will think that this makes them superior to those who make things. They expect elevated salaries and to have seats on committees that direct the activities of those who make things. They give themselves grandiose titles and buy big computers to explore ever more abstract ideas that are totally divorced from the needs of design or manufacture. They are regarded as the elite!!


One might observe that all this flies in the face of what is required for an engineering business to be successful.


If this is the case it is fair to ask who actually does engineering now and it seems to me very often to be those with little, if any, formal training, no knowledge of how to apply Newton’s laws and generally a total dearth of acquaintance with the science that underpins engineering proper. They design using their own pseudo-science that has grown up without any reference to Newtonian mechanics and is now the subject of examinations. Even professors of engineering contribute to pseudo-science.[9]


If one looks to see how engineering should proceed one must start with a definition of an engineer.


The mechanical engineer

This textbook is clearly aimed at those who aspire to be engineers, that is, in the upper echelons of those who practice engineering and who seek to achieve this by study of both the fundamental principles of the science that is relevant to engineering and by learning engineering practice. Such people are often called professional engineers. The question that one must now answer is “What does a professional engineer do?”


There is one definition of an engineer that says that he is a man who can do for a dollar what any fool can do for 5 dollars. It is saying that mechanical engineers do something special that most other people cannot do. It seems to me that a better definition of the professional engineer, that this text is aimed at, is someone who can design and either make something, or cause something to be made, that requires an input from science, either implicitly or explicitly, and that actually works as intended.[10] I do not think that there is any question that the sine qua non for someone to be an engineer is a good working knowledge of the science that is pertinent to engineering, knowledge of good practice in engineering and a healthy scepticism. These three taken together let engineers pick out how artefacts actually work both in fact and in principle and not to make many mistakes because they can check with confidence against science at all times. When it comes to design there is another ingredient that involves instinctively seeing how things work and therefore the main thrust of a design and having a good idea what sizes and speeds and shapes are most practical. This last is almost impossible to teach and it comes from an innate interest in engineering. For me design is the most important activity of engineering.


Looking for the science that is pertinent to engineering

Science exists in the minds of men and women and as we have seen Kuhn says that all of us look at the world using a body of received opinion, “recognised” science, shared assumptions, and expectations about ourselves and the world at large. In my view minds are very slippery places where all things that are inconvenient can be ignored.[11] If an attempt is made to write down a definitive science these inconvenient things cannot be ignored and they must be faced and resolved[12] but, in the end, what is written down is the output of single minds with their inevitable limitations.


Kuhn says that a text book on a particular subject is written to bring together all the paradigms that are currently recognised as the best. Perhaps that applies to science but I am not convinced that it applies to engineering. It would be wonderful if we could say that all texts are written for altruistic reasons or even by competent authors but this is not usually the case. I once heard a lecturer say that those who should write textbooks never do and those who do write textbooks never should.[13] In part I am writing this chapter of my book because it interests me to try to disentangle the various influences that affect the style and content of written science.


An author may really create nothing more than an agglomeration of other peoples texts complete with their misinformation. I do not think that this will do. Inevitably, in order to write this book, I had to refer to other texts but I decided that I would treat everything critically and remove anything that appeared to be ill-conceived or plainly unnecessary or wrong or derived from pseudo-science or unnecessarily mathematical. I hope that what is left is coherent and easy to follow. Unfortunately every writer must take a decision to write at some instant in time, knowing that, in what may be just a short while after the text is so say finalised, some new thought will occur that could have altered the text to advantage. It has happened to me. Engineering science cannot be set in stone.


What is written is in several forms. There are textbooks that the authors, at least, regard as a basis for formal learning. There are many papers on specialist topics and perhaps lots of collections of allied papers. Textbooks and published papers tend towards the academic treatment of engineering science. They rely basically on Newtonian mechanics. Then there is the vast field of published text, handbooks in all the fields of engineering and standards that describes engineering practice and includes much data that has been collected by observation and experiment. No doubt they also include material that might be at odds with a reader’s received opinions[14]. Unhappily when someone who is conversant with engineering science comes upon such material the credibility of all the contents of the book become suspect.[15] This data that is gathered by experiment is often called empirical. I do not like using the word “empirical” because it has come to be seen as describing parts of engineering science that do not yield to Newtonian mechanics for various reasons and therefore do not have academic status. When academics have to use data gathered empirically they use descriptive words like “semi-empirical” to excuse themselves leaving the reader to presume that the other half is mathematical and therefore respectable. They expect their science to give answers and are disappointed when it does not and do nothing about it. The whole field of science and practice should be seen as coherent with some applications depending heavily on Newtonian mechanics and other applications, that are just as important, depending on data gathered by experiment. Instead of coherence, there is a vast gulf between what appears in textbooks and what appears in engineering handbooks. It is as if they do not belong together or perhaps that people do not expect them to belong together. One must face the fact that when using empirical data care must be exercised to identify dubious material and if this is the only data available to think carefully whether there is some way that it can be used satisfactorily.


We have lost our way when it comes to science and engineering and it is now difficult to find science that is linked to engineering; all one can find is science that exists in the absence of practice. We need to know how this has come about and I think that it is intrinsic in a sentence from paragraph four of this introduction and I quote “The fact is that the theory that might be applicable to some mechanical device and the design of the device have often arisen independently of each other even if they co-existed.”


It was not until I started to write this chapter that I gave any thought to the history of textbooks in mechanical engineering. I recall in 1945 when I was studying for a Higher National Certificate that there was a list of recommended textbooks. Anyone could see that they were not the last words on the several subjects but they appeared to be the best available. I now see that those books were part of an ongoing evolution of textbooks and it was too soon for them to be the last words. Indeed some major sections of engineering science like automatic control have evolved since 1945.


I write this in 2010, that is 65 years on from 1945 and, if you go back another 65 years to 1880 the major works of Reynolds and Rayleigh were still in the future. So my experience goes half way back to the time when textbooks, as we know them, on fluids could not have been written. It seems to me that the foundations of mechanical engineering as a coherent subject based on a combination of engineering practice side by side with what I will call loosely “theory” was in place by 1900 when a great surge in the evolution of engineering took place and in some cases the theory altered and improved the practice. Gibson’s excellent book called “Hydraulics and its Applications” first appeared in 1908 and I have a copy of the third edition reprinted in 1938. The preface to that edition of 1930 tells us what was happening.


In order to bring the present edition up to date, it has been necessary to rewrite much of the book. The rapid development, during the past ten years, in the design of hydraulic prime movers has necessitated recasting the sections dealing with turbines, and a section has been added dealing with the wave transmission of energy in pipelines.


A chapter has also been added dealing with the principle of dynamical similarity and its applications to fluid motions and resistances, and the section dealing with the measurement of fluid flow has been largely extended.


Such errors as have been noticed in the second edition have been corrected, and the Author would thank those readers whose courtesy in intimating these has rendered their elimination possible.



October, 1924.


Here we have an example of a new paradigm being promulgated through Gibson's book. What was Gibson telling us when he talks of the rapid development of water turbines? At the end of the 19th century the prevailing idea about how water turbines worked was that the water hit the blades just as wind was thought to “hit” the sails of windmills. Look at the two pictures of a water turbine of that time. In figure 18-1 the circular top plate with its sector openings was fitted into the base of a water filled chamber at a depth of two or three metres. When the sector gates were opened water was fed to the wheel. The water simply fell on to the wheel. This was hopelessly inefficient. The water fell on to blades that were crudely made from plate as in figure 18-2. It was hopelessly misconceived. A moment’s thought would have shown that the water must flow across the blade once the blade moves. Once this was recognised the “theory” of water turbines followed very quickly and led to the turbines that we now use with efficiencies in the 90%’s. Theory got past a block in the development of the water turbine that trial could not pass.


Text Box:  
Fig 18-3
Gibson’s book is a joy to read. It is lavishly embellished with high quality engineering drawings of all sorts of hydraulic machinery that depend on the application of the theory given in the book and it tells us just how good the standard of engineering was at the time. I give just one example in figure 18-3 which is of a Pelton Wheel turbine. It is timeless and the only changes to the design that have taken place in the last 80 years are modifications that have been made to make it more cheaply. (I will return to this theme. It is important.)


However the things that do not appear in Gibson’s book must not be overlooked. Gibson has given us the theory of hydraulics and that depends crucially on Newtonian mechanics and the work of others like Bernoulli. Their work had been in existence for decades. He gives us details of many applications that depend on this theory, but there are lots of applications of hydraulics that involve so many imponderables that the theory cannot be applied and they are dealt with by experiment and so do not appear in the book. Such applications might be pumping slurries like concrete or even gravel in suspension in flowing water, the whole field of oil hydraulics, separators, intakes to engines, natural flows of all sorts, combustion, and the conveying of things like grain in bulk or gases for domestic heating. Usually our difficulties stem from the fact that we find ourselves unable to quantify properties like density and parameters like friction or to define system boundaries. This makes it difficult to gather data in any systematic way. Then we are dependent on data that has been gathered by measurement and systematised where possible for future use. Some call this empirical data and regard its use as a failure of science and therefore to be avoided. Engineers cannot hold this view because a design is required and must be found by any means that is available. (Kuhn's distinction between scientists and engineers.) The gathering and storage of empirical data, by its very nature, cannot have an underlying structure like that for theory-based data as evinced by Rayleigh and Reynolds but that does not mean that there is nothing to be gained from some conformity. But there is a great deal of data that, in its current form, is so often used satisfactorily that it will not be replaced.


There are other early books by Adison (published 1934 - 48) and Lewitt (1923 - 43) that have changes in content to reflect the evolution of practice. However the thing that stands out in these books is the way that the science and the applications form a seamless whole and the authors did not even consider the possibility that anyone would teach engineering without a context of practice. Yet text books would emerge that concentrated on science that was decked out with pages of mathematical formulae to the virtual exclusion of anything practical. There is a price to be paid for that development.


Text Box:  
Fig 18-4
Figure 18-4 comes from Robertson and Crowe. It is their idea of a Pelton wheel. It is what two people have produced from a basis of theory and seemingly no knowledge of practice. Had they ever seen a Pelton wheel running they would know that the wheel would have lots of buckets (typically 24) because, if the jet is not cut into short lengths, some of the water leaves the buckets before it can give up its energy and they would also know that the water leaves the buckets horizontally. They would also know that their buckets are too small for the pitch circle. If they cannot go from their theorising to a practical wheel how do they suppose that their readers can? It does nor stop there, the mechanical detail is ludicrous and is just as misleading in a different way as the paper on sailing from Queen’s University, Belfast.


Herein lies the problem with the books that emanated from America during the time of the Second World War and have spawned textbooks devoid of practice because theory and practice did not seem to fit together[16]. If you separate theory from practice you lose the practice and there is no way to decide what theory is relevant to engineers and theory burgeons without limits. This process is aided by the use of computing and if one reads the content of the internet it is clear that, too often, the computer is also being used in the vain hope that it will show how some mechanical system works when the programming must start from some statement of the modus operandi that the program cannot change.


In effect textbooks have been emasculated. To an engineer they are largely irrelevant except to obtain a degree. But the old textbooks are still there even if they are out of print and the practice of engineering is timeless. We are unlikely to see anything that is fundamentally new so the old books are not really out of date. They are much better than the nothing that we actually have.


How can one learn both theory and practice and link the two?

I can say how I learnt. At 16 years old I started a five-year apprenticeship in the railway works of the Great Western Railway. During that apprenticeship I studied initially part-time and then full-time for a British vocational qualification called the Higher National Certificate that would fit me for a role in the design office of the Railway Works. As I have always enjoyed engineering I learnt a great deal about how things were made in the railway works. Just by keeping my eyes open I learnt like how wheels were attached to shafts, piston rings were made to spring open, how riveting was done, how boilers were lagged and clad and these were all things that were not seen to be in my field at all. I also learned from the hundreds of hours spent practicing engineering drawing. The exercises were of real engineering objects like valves, gears, engine parts and mechanical joints like, say, between a shaft and a flywheel or a Hooke’s joint. The ability to draw has been an absolute boon. Drawing by hand is much more versatile than computer aided draughting.


Then I took a degree over three years and, when I graduated, I had, by current standards, a good grounding in both aspects of engineering. I was 24, had spent 21 years in education and training and had yet to earn enough to support a family. By sheer chance I became a lecturer and then, when I tried to use the material that I had supposedly learnt in my degree course to design laboratory equipment, none of all this learning worked in the way that I thought that it should. There was obviously a long way to go and it took me a couple of decades and the big thing that I learnt was not to put too much trust in books and to look out for where thay might be wrong.


It is fair to ask how I came to be in this position. The basic problem was that no one seemed to know how to use the theory in design except at the trivial level of, say, working out a friction loss in a pipe. At that time, and probably now, design proceeded by evolution although it was often just guesswork. Steam locomotives were designed by development not by research and development. The goal appeared to be to build the most powerful engine that could go through loading gauges with a limit on its length imposed by the distance that a man could throw coal using a shovel[17]. Significantly the railway saw no need for a laboratory devoted to research into flow patterns through boilers or new methods of construction[18] and I find very little difference in basic construction between Stephenson’s Rocket and the last steam loco to be built.


In my degree course we were asked to design a single-cylinder diesel engine of some stated horsepower. The method was to find an existing design and, from that, deduce dimensions for a new engine. It involved using an indicator diagram to find maximum forces, the variation of, say, turning moment and using what had been learnt in the courses on the theory of machines and strength of materials in conjunction with factors of safety to find these dimensions. It was a process that could never have led to the design for the lightest diesel engine of this power. To be fair, the idea of mathematical modelling per se were still in the future as were the computers that would make mathematical modelling a practical and very important tool for design.


This was compounded by an attitude of mind that supposed that there was some inexorable process by which a design could be evolved from some basic premises if only one could find it. (See Kuhn.) This same attitude of mind included the false position that the trial solution is an admission of failure and therefore unthinkable I now know that there is no inexorable process, only an exciting cerebral activity that, with luck, finds a route through a fog of ideas to find the main thrust of the design and the trial solution, when made by computer, is far and away the engineer’s most powerful design tool. I think that this attitude that a trial solution is a failure is the mindset of the mathematician who by definition creates an edifice of logic from some rules of his own choosing, like the statement that not only are there numbers that everyone uses every day but there are numbers that are these everyday numbers multiplied by . I did not find out about the power of the trial solution except by chance in an examination question, set externally, requiring a trial solution. I was slow to see its great power but the arrival of computers changed everything and made such solutions very quick.[19]


In my day there were engineering handbooks of which Kempe’s Engineer’s handbook that ran to 2,600 pages was the best. It contained engineering practice in every field of engineering. It was totally ignored, if not actively denigrated, in the Department of Engineering in which I worked. The lecturers with their university degrees and no practical experience were not at home in these pages yet this is what their students would have to use in their careers. To my eternal shame I went along with their view at first. Then Design as a separate topic began to be talked about and I attended conferences about problem solving and other “pie-in-the-sky” ideas. It seemed to me that we needed some proper engineering design that showed how to use engineering science in combination with handbooks not yet another set of abstract ideas. When my turn came to take design I resolved to marry the abstract material that we taught with the content of Kempe’s and the result was sheer exhilaration[20]. I found that the two could be joined together to make something that was greater than either taken separately. I thoroughly enjoyed my design classes and so did the students but when I was instructed to change to sterile CAD I gave up those classes and went back to lecturing. No matter, I had changed into an engineer because I knew that I could both design and make things that worked.


The fact is that learning to become an engineer, even when you are interested, takes a very long time. It is simply not possible in three years full-time study at a university from age 18 especially in the absence of practical training.


If learning about engineering to my level involves a very large commitment of time to study it is fair to ask whether such a commitment is possible these days. I do not think that society is geared to such a commitment. Children start school at five years old when I started at three years and one month sitting at little desks in little chairs. They go through school learning things that are coloured by left wing interference using fashionable methods that are always changing and end up unprepared for the rigour of studying, between age 18 and 21, for a degree in engineering even one that has been watered down to save money. With luck a graduate might follow a postgraduate apprenticeship but it is most likely that he will take up a post in engineering and try to learn on the job using pseudo-science or its output. It is a process that is terribly wasteful of talent and works to the disadvantage of the world at large.


I have given this personal experience to illustrate just how much time has been required for me to become reasonably conversant with engineering. It highlights the problem of finding enough time in formal structured education to become conversant with both the science and the practice of engineering and it has been recognised for many years and many attempts made to resolve the problem. This problem is compounded by the fact that a large proportion of the engineering students are not really interested in engineering but suppose that in some magical way they could be “processed” by education to become proficient engineers. This is not unique to engineering, it seems to apply to every other field in which one might obtain a degree.


The science of the “logicians”

If there is pseudo-science that comes from the “slaves” what has come from the logicians? The classic example of this is in chapter 7 of this book about the friction loss in a pipe. The physicists, quite properly, investigated this in their own way and they found it to be a problem that could be never ending. They had on open-ended brief and sought to refine their work indefinitely. Lewis Moody wanted to use their work in engineering and produced his celebrated diagram that ended the investigation for all time. Moody showed that the research work had progressed far enough for him to be able to predict friction losses in practical pipes with more than sufficient accuracy for most applications. Reynolds, Stanton, Pannel, Nikuradze et al have been followed by men pursuing new goals in a similar way using all sorts of computers and the like but lacking their Moody to bring it to a halt should it reach the point of being useful.


I now look at the work of the NACA on aerofoils and it seems to me to be very academic and that it never had a Moody to make it all practical.


It looks to me very much as if there is a great deal of activity on the part of those who work in the absence of the constraints of practice that is of little value to engineers. Not to mince words, I cannot understand most of it, and I think that part of the problem is, as Kuhn points out, that I am not supposed to. It is deliberately intended to be arcane so that its circulation is limited. In effect it is in code.


I spent a long time, 20 years, trying to understand the bulbous bow.[21] I looked all over for an explanation of how a bulb fitted to the forefoot of a hull could reduce the drag. There were several unlikely suggestions that were on their way to becoming pseudo- science. There were lots of computer-generated diagrams giving pressure and velocity distributions around a given bow and under the hull but no explanations of the way the bulb worked. There was also a great deal of very opaque text. I may be quite wrong but it seems to me that if you do not start with a rational explanation of how the bulb works you cannot expect the computer to explain it for you because this computer activity depends on you to tell it what to do. At some stage you have to explore the response of the computer and if you do not have an explanation of the way the bulb works you cannot tell which of the responses is the best to help you home in on a useful result.


Like the logicians of ancient Greece the abstract process is often conducted in the absence of an input from reality. It seems to me that this is a comfortable position to be in because seemingly erudite papers can be presented at great expense to like-minded people who politely listen and soon forget and who will soon be presenting similar papers and wanting an audience. I may be cynical but it would be nice to see a few useful things coming from this activity on the internet in words that anyone can understand. The fact is that even science-based engineering design proceeds at a relatively low mathematical level and engineers know that there are many factors that are not mathematical or scientific that can obviate science at a stroke. I once asked a civil engineer how one could calculate the required thickness for the walls of an underground tunnel. He said that I should understand that the designer’s fee was a small fixed percentage of the overall cost of the project and the cost of the wall of the tunnel was the major part of the cost. Enough said. There are hundreds of influences like this that can bring fancy ideas to nothing. Some mighty computer may be used to design a new wing for an airliner but, if the performance of the wing depends on it being scrupulously clean you might need to think again about this new wing when it is in its normally-used state. If you know nothing of these overwhelming constraints on engineering design the theorising is likely to be pointless.


Chinese whispers

This is a major problem. My dictionary defines Chinese whispers in this way :- “a game in which a message is passed, in a whisper, from one to another in a group of people with the result that the final version is different from the original, often amazingly.”


Let me give an example of this in engineering. As I have said in Chapter 11 you can read the collected papers of William Froude and a memoir of his life and his work. It is hard to see how these definitive documents could have given rise the remarkable entry in Wikepedia (Jan 1 2010) except by Chinese whispers. It appears to be factually incorrect about Froude and his career and work and to attribute things to Froude that he did not do.


It is difficult to avoid Chinese whispers. In Chapter 11 of this book on ship resistance I found it to be necessary to paraphrase the original book on Froude in order to put its content into modern parlance so that anyone reading my book will have a version that includes some of my interpretation. Anyone reading my book as a reference for some new text might then change it again. It seems that information is almost impossible to transmit without it losing its original meaning and the more times that it is transmitted the more it changes and becomes embellished. By 2010 this process has had plenty of opportunity to change to the entry in Wikepedia. This is why, 22 years ago, I went back to the original papers to see just what were the “original messages” sent out by Rayleigh, Reynolds, Stanton and Pannell and Froude. Perhaps I should have included Mach because I am not certain what he actually did.


The most common way to learn about engineering is to be taught, that is lectured. Lecturers are always working against the clock and the notes that a student takes away from the lecture room are necessarily condensed and only a skeleton that the student can flesh out from his memory. Worked examples are there to reinforce the memory and to show how the texts can be used in practice. A lecturer will have started by listening to other lecturers who also started by listening to lecturers and there has been plenty of time for the Chinese whispers to do their work and for omissions and distortions to occur. It is easy for the content of lectures to simply dwindle with time and only the need to set examination papers of eight or nine questions stops it from dwindling below some irreducible minimum.


Everything is going the wrong way in engineering education.[22]


Mathematical modelling and solving by trial

The evolution of mathematical modelling and mathematics packages like Mathcad on computers has made life much easier for the mechanical engineer. At a stroke it removes much of the labour of differentiation and integration and of using every other mathematical function and it plots graphs instantly. This makes the exploration of mathematical models of fluid systems very easy to explore by trial. I have sought to illustrate this process throughout this textbook. When creating mathematical models of some real fluid system it pays dividends to keep in mind the accuracy needed for that system to be acceptable. There is no need to size a large pipe to the millimetre if the diameters of available pipes go by increments of 100 mm.


Go to the book on “The physics of sailing” on this website and read Section 1 entitled “The single soft sail” and then the paragraph “The performance of the single soft sail” to see what can be deduced from very much simplified mathematical models. Such an explanation could never come out of pseudo-science.


Engineering handbooks

If the content of an engineering degree proves to be inadequate for use by a practicing mechanical engineer, there is no other recourse than to look at what is on record about practical engineering. As I have said, most of the lecturers where I worked decried handbooks and that is an unsustainable position to take but one must remember that they had little experience on which to draw.[23] I am not familiar with the current content of handbooks but mostly they stretch back a long way in engineering and much of the content may be timeless and there is little incentive to change so my 1976 copy of Kempe’s will not be out of date in many topics. How can involute gears change?


When I tried to use Kempe’s Engineering handbook in conjunction with the science that we were teaching I found that the key to understanding was to try to understand the way in which the originator of the data sought to collect the data and then to store it for use. I think of that as being an underlying strategy but I found that other lecturers thought that I was trying to create a new jargon like that used by local government officers to describe their activity so that no one understands what they are doing. But understanding the strategy used and not dismissing it is central to success[24]. I have drawn attention to the commonly-used strategies in the rest of this text. The one most often used in fluids is the creation of a rational expression that is dimensionally consistent and then to use it in conjunction with one or more dimensionless coefficients that can be determined by experiment. But there are several others, for example, the way in which the stability of a container ship is determined by continuously finding its centre of gravity as containers are loaded based on notional centres of gravity for the containers. That really is a strategy. In thinking about these strategies one must recognise that a significant proportion of the data that was gathered prior to say 1900[25] was collected piecemeal and was not designed to fit in with any grand plan of data gathering. Measurements were made in what we now think of as unlikely units and the data in these units was related for plotting graphs and deducing suitable mathematical relationship for calculation. That data is still correct and usable even if we do not like the units. But we have to be careful because the “constants” will almost certainly have units. For example, it may be that, over a useful range, a quantity  in gallons per hour is related to some dimension  in feet by an equation of the form  where ,  and  are constants. Then  must have units of gallons/hour,  is a number and  has some very unlikely dimensions. The best advice when faced with this is to use the equation as it stands, convert your input data to gallons per hour and feet and convert the outcome to the units that you normally use. There are too many pitfalls to trying to convert the equation.


Engineering handbooks contain vast quantities of reliable information that has been built up over time. It is contained in text, drawings and diagrams, photos, tables, charts and graphs. It is the graphs that contain the most condensed information. This is the result of working to some experimental plan that includes the variables to be measured and the way in which the data is to be presented in the graphs. Now all we need is a cadre of lecturers who can link engineering science and the content of handbooks in courses in design!


Engineering drawing

There is no question in my mind that the decision to stop teaching engineering drawing was the greatest single error made by educationalists. One could see how it came about. It was widely felt that teachers of engineering drawing did not need to have degrees. It was assumed that what these lecturers needed to know was how to get would-be engineers to draw using pencils, drawing instruments and boards and tee squares etcetera and that the subject matter was not important. But this was in reality an opportunity to learn to use drawing to depict engineering artefacts to true scale and to learn how to use drawing to solve problems in mechanics by graphical methods. In my view this is the foundation of mental modelling. These are invaluable to the practicing engineer both as information about engineering practice and as tools of the designer.


Drawing would have had as much time as, say, thermodynamics in the first year. As overall class contact time was reduced some of the course content had to be reduced and the first to be reduced was engineering drawing because the very fact that it could be taught by a non-graduate meant that it was of low status in the eyes of those who had graduated without practicing engineering drawing. They did not understand that teaching drawing properly requires very considerable ability. Soon there was too little time to become sufficiently proficient to meet the requirements of awarding bodies. Engineering drawing was abandoned and with it went the last link with practice.


For a short time drawing using computers made a comeback but it was too time consuming and those charged with teaching it had little affinity with the topic and it quickly disappeared again.


To someone like me, who can draw, computer aided drawing is not very satisfying. No doubt it is invaluable to those companies who manufacture using computer controlled machinery. There is now a trend to replace engineering drawing with the sort of imagery used extensively in television. The outcome seems to be more akin to a cartoon than anything real and the viewer is left to put back the reality.


Engineering artefacts

Engineering is nothing without artefacts, that is, things made by people to achieve some purpose. The need for artefacts goes back a long time in history. No one can say when the first boat was constructed but its construction would have been prompted by a need to do something for which it was necessary for people to be able to go afloat on water. Its construction would have been facilitated by the use of tools that may have been no more than hand-held knapped flint. Whoever made it did not have the benefit of any insight that might be described as “natural philosophy” which is the form in which science first appeared. Instead some process of evolution by ingenuity, and by trial and failure, was started that reached its peak in the wooden-walled sailing ships of the nineteenth century. On the way what can only be described as good practice in the building and operating of ships came to be understood and codified. This good practice was not influenced by science as we know it now but sailing ships had a very good chance of returning from a voyage so the good practice could be described as capable of being successful. Froude sheds some light on the period of the nineteenth century when the methods used to design ships began to change and it is clear that it was extremely difficult to start again and the original practice is still there getting in the way. Even the attitude that sees no need to change the existing paradigm to fit in with recognised science is unchanged. This situation is mirrored all over engineering wherever the design stretches back into history.


It is clear that our society depends on its success on certain machines. For a steady supply of electricity we depend primarily on steam turbines because boilers can burn all sorts of grades of fuel. Steam turbines are augmented by gas turbines driving alternators that can have very high efficiencies with a low first cost but run on gas and other clean fuels and cannot burn heavy oils. But for this limitation we would generate most of our power using gas turbines. Some countries are lucky enough to have hydroelectricity. For land-based transport we depend on internal combustion engines using petrol or diesel fuel but, for railways we use electricity that is generated in the normal way or diesel fuelled engines. For transport by sea we use large diesel engines and very crude fuel. Aeroplanes use gas turbine engines running on what is essentially light diesel fuel. The heating of buildings is largely dependent on gas and electricity. Presumably it is the business of mechanical engineers to be involved with all of these machines and especially those that depend on Newtonian mechanics for their design.


Society also depends on other industries not the least of which is the food industry where a large proportion of our food is processed by machine. Packaging for food is made by machine and is big business, indeed it makes possible the transport of food over long distances from market to consumer. Motorcars are made in fully automated factories as is electronic equipment. It is a moot point whether these are the fields of the mechanical engineer or the production engineer. It may be that a mechanical engineer could succeed in production engineering but the production engineer will always be short of science if he tackles mechanical engineering.


Here I want to concentrate on the machines that depend on science for their partial or complete success.


It seems to me that the turning point in the creation of science came during the latter half of the nineteenth century. When James Watt (1736-1819) began his search in 1763 for ways to improve on the performance of the Newcomen engine that had dominated mine drainage for 40 years from 1720 to 1760 no one knew what volume of steam would be generated from one pound of water at atmospheric pressure let alone at any other pressure. Yet by 1900 all this was “known" as were the properties of gases and a great mass of physical data. In addition Newtonian mechanics, that had been very slow to find its way into use by any but a select few[26], had gradually been adopted more widely[27]. The subsequent science that emerged then depended on whether the machine existed before the end of the nineteenth century. Typically the aeroplane did not exist in 1900 and its science grew up in the presence of engineering science and the work of Reynolds, Rayleigh and Froude. It is a modern science. The boat had existed for many centuries and many ideas that purported to explain various aspects of boats and ships, but were actually highly questionable, had taken root and found widespread acceptance before the rise of science as we now know it. Unhappily many of those ideas persist and some are taught as being correct and we end up with a hybrid sort of science that is difficult to dislodge. However we must not suppose that aerodynamics is free from ideas that do not help the engineer. To me it is hard to reconcile the Kutta-Joukovski theorem or Lanchester’s ideas with anything that a working engineer might use. They seem like a leftover from mathematicians feeling their way to an understanding of the action of aerofoils only to be overtaken by practical testing in a frame work of modern science. Their work lives on to pad out degree courses. The engineer is much more likely to need to know which aerofoils have been found to be successful and to be able to design from a knowledge of the practical problems of operating an aeroplane including maintaining it to the standards needed to have the design performance.


Parsons’ first steam turbine in 1884 shows no evidence for any understanding of how any turbine actually works. It was Parsons’ good fortune that the most primitive turbine would have a higher efficiency than contemporary steam engines. The steam engine cannot escape the serious disadvantage that live steam and exhaust steam alternately occupy the cylinders and the inlet passages and this leads to condensation of the incoming steam and to the subsequent evaporation of this condensate into the exhaust steam. This results in a very serious loss of heat seemingly directly from inlet to exhaust. The steam turbine has no such problem because the steam flows steadily through the machine and is always in thermal equilibrium with the casing and the blades. The fact is that, like the sails of a ship and the sails of the primitive, wind-powered corn mill, it will work, albeit inefficiently, even if it is made to a poor standard. Parsons appears to have accepted the wisdom of the times that turbines worked by the impact of high-speed steam on the blades when a moment’s thought shows that the steam flows, at least in part, across the blades. (It is hard to know whether de Laval, the inventor of the impulse turbine shared this concept.) We still have an echo of this idea in the names of the two types of steam turbine, impulse and reaction. Steam turbines have grown in size until the blades in the low-pressure stages are 1.3 metres long. It was not long before the mode of operation of turbine blading came to be understood but with this new understanding came the need to make blades that were twisted and of a highly cambered aerofoil shape. Machines to make such blading did not exist and turbine designers used great ingenuity to create blades that were good approximations to what was really needed by generating shapes from cylindrical surfaces and flats. I thought that they were a credit to the designers but, as turbine outputs increased, the failure to achieve the potential efficiency through these approximations involved a very significant loss of income. A steam turbine in a power station will work with a boiler pressure of about 150 bar and exhaust at about 0.07 bar. If the steam just flowed from end to end expanding as it goes water droplets would form in it and severely erode the blading in the lower pressure stages. The science of thermodynamics has been used to overcome this problem and, at the same time, to improve the efficiency. Thermodynamics appears to have evolved to suit steam, and not internal combustion engines, and its application to the steam turbine plant showed that, if the steam is bled away at intervals during the expansion and reheated in the boiler before returning to the next stage of the turbine, the problem of wet steam can be solved and the overall efficiency actually improves.


It is unlikely that steam-turbine engineers could ever have designed the gas-turbine engine. For the gas turbine to work at all the power required to drive the compressor had to be less than the output of the turbine. Blades made by the approximations of the steam turbine designers simply was not good enough, they had to be made to the shapes that were thrown up by the use of engineering science. We now find light-weight gas turbine engines working at pressures that exceed those of diesel engines and with efficiencies to match. The performance in terms of output and longevity of gas turbine engines is simply astonishing. It does not seem to me that thermodynamics has much to do with this improvement beyond the implications of the Carnot cycle. It is all to do with the mechanics of turbine blading and with production engineering.


It is hardy surprising that some steam turbines that may have already run for 30 years are being re-bladed with blades made using production methods developed for gas turbines.


The steam locomotive and the railway altered the way of life of the whole of Europe, Asia, and North America and the rest of the world to a lesser extent. It has now been superseded. It was inevitable. The steam locomotive had the same problems of heat loss caused by the alternating use of the cylinder and the steam passages but it had a severe disadvantage of not being able to work to a condenser and so not utilising the steam effectively. The locomotive was required to produce high torque for accelerating a heavy train and then produce low torque when the train was running at normal speed and to do it without gears. This led to the provision of very large cylinders that were really too large for most of the time. However this did permit the use of valve gear that was essentially symmetrical to let the locomotive haul in reverse. The boiler of a locomotive raised steam very well indeed but the need to make the boiler as large as possible within the loading-gauge restrictions led to a compromise between the size of the boiler and the thickness of the lagging. You could feel the heat being radiated from the boiler as the loco pulled into a station. It was a labour intensive engine to maintain and the work was decidedly unpleasant. So far as I can tell science played little part in the history of the steam locomotive. It was the product of practical engineers and the only significant change that they made was to introduce superheating in about 1912. In the presence of better systems of traction it was a dead end. I have the impression that its legacy in the minds of railway engineers lives on.


The wind-powered corn mill also came to dead end but one should not dismiss it out of hand. The significant elements that entered the mainstream of engineering were the wooden gearing as a lead-in to metal gears and the sweeps or sails that turned to extract energy from the wind that carried all sorts of automatic control systems. The earliest form of sails was the canvas sail which no doubt could be furled round its pole to match the sail area to the wind just as a modern jib sail on a yacht is furled round its stay. Corn milling must have required eternal vigilance to manage the sails in the wind. In England at least, the canvas was mounted on sweeps made in the form of a wooden lattice on a stout spar. There were obvious difficulties with manhandling the canvas to match the wind speed and turning the mill or the sweeps to face into the wind. This brought out the ingenuity of the millwrights to develop the fantail that automatically turned either the whole mill or the bonnet that carried the sweeps and the sweep fitted with wooden louvres that could “spill” the wind. In the end those louvres were pivoted and ganged together to open with increasing wind speed against a leaf spring. The sweeps became interesting technically and examples are preserved for engineers to study. There can be no doubt that the millwright worked on the premise that the wind hit the sail just as the wind hits the spinnaker of a yacht. Yet the fact that the sweeps move at a speed that is of the same order of magnitude as the wind-speed must mean that the wind, at least in part, flows across the sail. The millwrights did find that the sweep must be twisted and that a flat sweep was almost useless so they must have had a contradiction between the need for this twist and their idea of the air hitting the sweep[28]. Whatever they thought, the arrival of the engine and then the electric motor brought development to an end and when next a wind-powered device appeared it was the wind-powered electricity generator and that uses aerofoils shaped blades.


A similar fate overtook small-scale water-power. Mostly water-power was derived from the use of water wheels that were built over and beside rivers, streams etc. Water was more reliable day-to-day than wind because the flow primarily varied seasonally and not with the weather. However the number and position of sites was restricted by topography and often the wheel was not in the most convenient position relative to where the power was required. This did not matter much for milling corn but, where wheels drove factory machinery, all sorts of schemes were devised to carry water from its original bed to the vicinity of the factory.[29] Steam power, the internal combustion engine and electricity proved to be much more reliable.


I am not wholly certain but it seems to me that there were only two sorts of water wheel, those in which the water was loaded on to the top of the wheel and called overshot wheels and the other sort where the water entered the blades from one side and flowed under the wheel called undershot wheels. They are very different in principle because in the overshot wheel the free surface of the water was at atmospheric pressure and in the undershot wheel the pressure varied as the water flowed under the wheel. These wheels were constructed to evolving designs but, if you look carefully at enough of those that remain, it is clear that achieving the highest efficiency was not the uppermost thought. Nevertheless there are some wheels that are close to the optimum in design. I think that a great stumbling block was not hydraulic at all but the failure to recognise the need to use bearings with caps. It was normal to use a half bearing and this was almost impossible to lubricate. The shaft wore rapidly and ultimately broke and, in the case of undershot wheels, the wear led to the wheel rubbing on the masonry. A great deal of know-how went into water wheels and windmills and it is not surprising that millwrights became the jacks-of all trades that maintained the fabric and machinery of factories.


Milling of cereals goes back to the time when people started to grow grasses that produced seeds that could not be digested unless the outer shells could be removed prior to cooking. A mechanical process was needed and this led to the pounding that is still used in Africa and to milling between stones that started with the quern, progressed through windmill; and ended with modern milling equipment. It has been a wholly practical process throughout and it is no surprise to see so many water-powered mills.


The internal combustion engine clearly owes its name to the pre-existence of the steam engine where fuel is very obviously burnt outside of the cylinders. Where thermodynamics seems almost to have developed for the steam engine with its closed cycle for water as the working fluid, when it came to using air as a working fluid, with the air supporting the combustion of the fuel, the cycle was necessarily open and Carnot’s concept of a reversible cycle of two heat transfers and two work transfers was not directly applicable. Thermodynamics turned out to be not so useful for the internal combustion engine. Whilst thermodynamics still showed that the goal must be high temperatures the possible maximum temperature that could be achieved depended on the fuel to be burnt and how it was be introduced to the cylinder. High efficiencies came about as a result of improved manufacturing methods, fuel injection and engine management systems and not through thermodynamics. The modern internal combustion engine is reliable, has a high power-to-weight ratio and is approaching the maximum efficiency that it is ever likely to have.



You will see that I regard the task of the engineer as very complex and must necessarily make use a very tangled mass of information that may or may not be correct. I have endeavoured to show how this situation comes about and why it is unlikely to change. The result is that we must all be very wary of everything that we read and it is a great help in assessing the value of any information on offer if it can be set against its correct background of engineering science.


I have attempted to describe some of the non-engineering influences that have muddied the waters and the main message is that we need to look carefully at our applications even if they appear to come from an impeccable source. They will almost certainly contain the consequence of the existence of pseudo-science and other incomplete explanations of how they work that is part of their history.


I think that I have given enough examples to show that whilst the design of engineering artefacts may be improved by the use of engineering science the ultimate design depends on lots of other factors that the engineer must learn about and take into account. I will attempt to explain these things in the rest of this book as well as giving an explanation of the application of the engineering science.


Looking back I wonder whether I have not been much too parochial and viewed all this from the standpoint of one familiar with the science that is used in engineering. I have the feeling that engineering may look quite different when viewed from the other end. Even so I am better off with my engineering science than any group of engineers without it.


[1] This is equally true for the internal combustion engine.

[2] Why is the first part of Abbott and von Doenhoff quite so off-putting?

[3] One and one third horse power to pick up dust!!!

[4] How can engineering matters be resolved by a vote?

[5] No matter how persuasive the blurb may be, solar heat collectors do not work when the sky is overcast and the collector is under 150 mm of snow. It is extremely hard to square the claims made for condensing boilers with the laws of physics. Nor do wind powered turbines produce power when there is no wind.

[6] Sailing gives a good example of pseudo-science at work. Yachts can sail at about 40° to the wind at a quite respectable speed. In order to do so the mainsail is set with its boom along the centre line. If the sail is reasonably flat it is obvious that it cannot generate a force to drive the yacht. The jib is set at about 22° to the centre line. The pseudo-scientist wrongly regards the jib as too small to drive the yacht at its observed speed and concludes that the jib must be altering the flow over the mainsail so that it can generate a forward force. The sail is observed to have an area close to the mast that has a large angle to the centre line of the hull and, if the presence of the jib somehow increases the pressure on this area, it could drive the yacht. This unspecified area now becomes the “power point” and the jib is now adjusted to make this “power point” drive the yacht as fast as possible. In fact, of course, the jib does drive the yacht working in the flow of air deflected to leeward by the main. So the pseudo-science partially achieves the desired result for the wrong reasons but of course will not lead to ways of improving the set-up.

[7] This pecking order extends to whole institutions. Why the abstract thinkers imagine their work is more important that that of the men who maintain their supply of electricity or fix their gas boiler I cannot think.

[8] There are some disciplines that are not dependent on science that can only acquire any status by converting to a science, for example Social Science.

[9] I recall reading a paper on the stresses in a square key used in a tapered joint between a shaft and a flywheel. In such a tapered joint the connection is by friction brought about when a nut draws the shaft into the flywheel. If there is any stress in the key the joint has failed.

[10] Once I had retired I had a chance to find out whether I can do this and, to my relief, what I designed and made from science did work as intended.. I have recorded some of my projects on this website and I think that my best effort was the miniature air-cushion vehicle although the wing-sailed yacht is a close second. The air-cushion vehicle relies solely on my knowledge of science and on my skill as a craftsman to design and build a device that uses 110 watts to do better what others do for 500 watts. That is what engineers do.

[11] This introduction is made up of lots of inconvenient bits of knowledge that I have been trying to get into some sort of order for years yet have always ignored them because it needs such an effort to sort it out. Now I am determined not to ignore them.

[12] This is what I am doing in writing this book and experimenting. In section 1 I have had to face many things that I have consistently ignored and often it is a very protracted process. Each time that I nail one I experience a great thrill.

[13] There are notable exceptions e.g. Prandtl.

[14] Civil engineers use trial solutions much more than mechanical engineers and some mechanical engineers pooh-pooh this and reject the whole book. It is very short sighted.

[15] It is interesting that old handbooks have a confidence to them that we have largely lost.

[16] I recall learning about the Porter governor for steam engines when I was 18. By the time I met it again at age 21 it had changed to the conical pendulum and when I asked how I could restore the mechanical detail to produce a governor I was told that this was the new way of looking at mechanical engineering. I did not believe it then and it is now evident that it was not viable.

[17] This may seem to be a strange statement but the output of the boiler depended on the area of the grate on which the coal was burnt. That area depended on the gauge of the track, which is fixed and the distance that a man can throw coal.

[18] Can you believe that the frames of a loco were set up over a pit and the cylinder set clamped between them before they were drilled by portable air-operated drilling machines. When that was complete the motion was set up using gauges and chalk-lines.

[19] The importance of  the trial solution cannot be overstated. I recall a Ph D student Fred Maillardet who had been asked to find ways of designing artificial legs for children. It was no use making an artificial leg the same size as the existing real leg, it needed to be for the leg that the child would have in a year’s time so that the leg might last for two years. Mr (now Professor) Maillardet had an analogue computer with a VDU for its output and this computer was to be set up by comparison with a film of a child walking. All the parameters were known with the exception of the magnitude of the friction of a knee joint. Months were wasted in tests on joints from cadavers and searching for a figure from knowledgeable people. One day he told me this story and I told him to put the friction to zero and see what happens. An hour later he came back to say that the film and the VDU were in total agreement. He was back on track and produced a thesis of immense value. Looking back I do not think that, at that time, I would have suggested that he try a small value if the agreement had not been so good. But I think that Fred would have done it for himself.



[20] When I did this I always set a design that was new to me and one that would require reference to a handbook or to other data. Then everyone could make design choices and see the consequences. I used to have a first session when we discussed the problem and possible approaches and then we all attempted to carry the design forward during the week. I wrote mine up and took my work to the class. In the second session we discussed our progress and this gave those who unable to make any headway a chance to find out what others had done. In this way we more or less kept together but the frontrunners could forge ahead. I tried to complete a design in six weeks. They were interesting exercises



[21] Se my article on “The bow wave” on this website.

[22] When I was 16 and in my first year of my apprenticeship I worked from 8 am until 5.30 pm five days a week and on Saturday morning. I attended evening classes four nights a week from 7 pm until 9 pm and did my homework on Sunday. Success in the examinations was rewarded with one day off each week work to attend college and even more homework. We thought that it was worthwhile.


When I was studying for a degree we worked nine sessions of three and four hours each week and worked every evening on homework, lab reports and design and drawing.


[23] When I asked the stress analyst how much interference I needed to shrink a flywheel on to a shaft I received several pages on designing turbine wheels for uniform stress. I presume that he did not know about heating the flywheel to make the hole in it expand to slide on to the shaft.

[24] Some of our lecturers sought to denigrate these books by referring to factors of safety as factors of ignorance. I used to wonder just where the ignorance actually laid.

[25] It was the work of Rayleigh that gave us a grand plan.

[26] It was a new paradigm and was greeted with hostility on several counts not the least of which was Newton's decision to just postulate the existence of gravity without explanation. We are still in the same position

[27] Newton is said to have been challenged for authorship of some explanations of physical phenomena and wrote in Latin and in a convoluted way so that it would be very difficult for other to claim the principia.

[28] In Kuhn's words their existing paradigm was throwing up anomalies and was ripe for change. Despite the century or so that has elapsed people still write letters to the national press in which it is evident that the old paradigm still exist and they think that the wind hits the blades.

[29] In Cornwall in UK I came across a water wheel that was used for pumping. The pump and the wheel were linked by a wire rope about 100 metres long that passed through two windows of a building