It is becoming increasingly harder to avoid the proliferation of people using AI. This is a subject that I have noted before on its use in naval architecture, but it is becoming so dominant that its instant acceptability, rather than the slop it creates, is becoming the narrative.
In the days before computers, facts mattered. Also, one could trust what one saw with their own eyes: the axiom of seeing is believing, the notion of if it can be seen (as in real) it must be true.
It is hard to imagine a photograph from the last century being doctored so much that the resulting image not only defied the norms of engineering but also believability. Such an image would be instantly derided as nonsense and rubbish and clearly an attempt to subvert fiction as fact. The author of such an image or concept of their “creation” would be laughed at and they would walk away with their tails between their legs and usually never heard of again, with any and all their credibility lost.
There is precedence for this. Some may recall when two eminent scientists, such as Pons and Fleischmann, announced they had achieved cold fusion. On March 23, 1989, they grabbed the world’s attention by announcing at a press conference in Salt Lake City in Utah that they had observed controlled nuclear fusion in a glass jar. There then followed a worldwide debate about its verification and validity; it seems no one could replicate the results they cited.
Eventually, Massachusetts Institute of Technology researchers found serious flaws, and a US Department of Energy panel concluded in October that same year that there really was nothing to cold fusion. A lot of time and money were subsequently wasted in going from their announcement to their humiliation some six months later.
If we fast forward to today, whilst perhaps not as earth-shattering in its suggestion of changing the course of the world’s energy, AI – or more correctly, the use of AI by engineers, as well as wannabe-engineers – is no different in its presentation and promise as those offered by Pons and Fleischmann.
If we look at seakeeping, this in itself is a terribly complex subject and there is no single attribute of a ship’s characteristics that that can encompass this complex subject – just a series of links and connections and applications of parameters that slowly influence a ship’s motion to a larger or lesser degree, all governed by often very complex mathematics.
To begin, we need to define seakeeping. Simply defined, seakeeping is the ability to remain at sea in all conditions and carry out the duties of the vessel. However, the subject of seakeeping tends to be a complex and a subjective performance measure in the absence of real hard data. What this means is that we need to understand the relationship between the sea and waves with that of the response of a vessel to the waves. Therefore, we need to look at the basic equation of motion of a floating body in a sea state.
The natural periods of a vessel are associated with the six degrees of freedom: three in translation – heave, surge, and sway – and three rotational – roll, pitch, and yaw. Only those with a natural restoring ability – heave, pitch, and roll – have natural periods associated with the vessel, since this is the basic principle of simple oscillatory motion.
A very simplified period of motion of a floating object is given by the following expression:
These coefficients can be summarised as:
a = Inertia moment → virtual mass moment of inertia, inc. added mass
b = Damping moment → a force in opposite direction, like a bilge keel
c = Restoring moment → relates to shape of hull
The forcing moment is the wave. When this term is set to zero, it provides the ability to find the natural period of motion of the vessel. The natural period of motion of a vessel is the key to understanding basic seakeeping. This is the start of these links and connections (mathematically) that help explain and define a vessel’s response to waves.
So, to explain a little more and delve a little deeper again, we need to understand how the sea and a vessel interact with each other and how this links with the natural period of motion of a vessel. We can relate these natural periods to that of wave heights, their periods, and, importantly, the speed at which the vessel is approaching the waves. This can be represented mathematically by the following equation:
From this we can calculate the period, or frequency of encounter (we) knowing basic characteristics of a wave – its period, length, speed – as well as the heading, relative to a vessel at speed (Vs) and an angle (cosw) to the wave train, head seas being 0° and stern/following seas 180°.
Why is all this important? Because when a vessel’s natural period of motion becomes close to or is coincident with the encounter period of the waves as it moves through a series of waves, that is when we get resonance – uncomfortable or even violent motions and accelerations. Try driving over speed humps on the road where the distance apart of each hump is coincident with the wheelbase of the car. Not pleasant!
The above is a very simplified breakdown (if such a thing can be done effectively) of the key topics to research and understand to ascertain which parameters of a vessel influence any or all of the above.
Manipulation of the above, an understanding of the waterplane area, the shape and length of the hull, as well as the direction in which the vessel is travelling relative to the waves begins to form a pattern of what is being influenced and how. And then, after many years of delving into such a subject, one may begin to see a clearer pattern emerge of what “kind of works” and what “does not work” and what physical attributes of a ship’s geometry and design influence such outcomes.
This then brings us to those “creators” wishing to be seen as real engineers, or even in some cases, real engineers, beginning to use AI to solve a complex problem that they are unable to solve themselves by study and research, and presenting such output as “real” or “surely why can’t this can be done”. This brings us back to 1989 and bold claims.
We move from the more “simple” errors to the more “bizarre” creations of large monohull passenger vessels with a twin SWATH hull attached under it, for better seakeeping, because foils can lift it free of the water. The creator thought about designing a passenger ferry, so out “popped” a large-volume monohull for many passengers, with some styling. Nothing wrong with that, but then he had an afterthought of seakeeping, so to add to the appeal, they added a SWATH hull form under it.
This ignores the fact that from a resistance and propulsion point of view, having large hanging struts and pods attached to a wide body monohull is not hydrodynamically efficient by any stretch of the imagination and such foils would need to be very large (i.e., more drag to the lift vessel). The notion of ensuring seakeeping by adding a SWATH hull underneath the monohull beggars belief. A little knowledge is dangerous in this context.
This is apart from the fact that AI thinks such vessels are possible is ludicrous. Where is the algorithm getting this “real engineering” from? It should be a red flag in itself. But more worryingly, it seems there is little pushback against such AI slop by engineers in the same manner that Pons and Fleischmann experienced. The only reason why such AI slop is posted/advertised on a regular basis is that it is not being debunked and laughed at loudly enough to humiliate the creator into taking it down. Seeing is not believing in this context; rather, seeing is laughable.
It does seem that the danger is not so much that AI is being subcontracted by those with next to zero knowledge and expertise in their quest to create something, but by the deafening silence of engineers for not calling out the nonsense that it is when presented.
To paraphrase: "The only thing necessary for the triumph of AI is for good engineers to do nothing."