Wave Tank Question

Dolfiman -Thanks for the link to the LEGI channel. It is a nice 2D tank, but much bigger and heavier than I envision. The Science Museum of Minnesota has a similar type of tank, albeit shorter in length and with a smaller cross section :


I worked there for 10 years as an exhibit designer and project manager, and am quite familiar with the tank. Even this is a 'beast' from my perspective. The thick glass sides and seals on the dry-back piston are heavy, complex, and expensive. I don't have a pic of the piston drive mechanism, but it uses a rolling-ring linear shaft drive mounted above the water level attached to a truss frame behind the piston/paddle face.

This is a dual-tilt stream table produced by my friends at Little River Research and Design (Emriver Stream Tables & Hydraulic Flumes https://emriver.com/) that is more in keeping with the scale and type of experience I'm looking for:


Their geo-models are well-engineered, easy to transport, can be set up by 2 people, and have a variety of instrumentation options. They are used in research and museum environments, with facilitation. You modify conditions, observe, and learn stuff in a ways you never will with screen-based simulation.

 

Interesting paper: "Numerical Studies of Directional Wavemaker Performance" by John F. O’Dea and J. Nicholas Newman. A PDF copy is attached.

The example wavemakers they analyze are 3 formats I would think fall under John Hardiman's term description of a 'snake-type' periodic wavemaker: John - please correct me if you meant differently.
Hinged Flap:

Translating Piston:

Horizontally-articulating Piston:


It looks like all 3 types produce a as a series of rigid face elements, even assuming seals via membranes or pleats between faces. An interesting observation is made on page 8 (italics and bold mine):

"The use of a piston-type wavemaker also makes it possible to use a horizontal articulation (see Figure 2c) in which the actuators force a translational motion at the joint between wave boards. The result is a continuous, piecewise-linear (rather than discontinuous), variation of the wavemaker displacement when oblique waves are generated. This has the potential of reducing the irregularities in the waves caused by the discontinuities seen in Figures 2a and 2b). Figure 17 shows the wave amplitude fluctuation at a distance of 30 m from the wavemaker bank, comparing the discontinuous piston and the horizontally-articulated (tent) piston mode. The 100 m long wavemaker is resolved into a coarse (33 boards), medium (50) and fine (100) spatial discretization. There does not appear to be any obvious advantage to the “tent” mode in this example, and in fact the overall magnitude of the wave in the middle of the basin (x = 30 m) is somewhat reduced when the tent mode is employed. The large-scale fluctuations are believed to be a consequence of the finite width of the total wavemaker, rather than the finite width of the individual boards."

Edinburgh uses this wet-back piston system, often in series, for their shallow-water wave tanks:

If one proposes to use a wet-back wavemaker that has small gaps between discrete wave making surfaces that move in parallel, be they paddles, pistons, or some combination thereof, I'd assume the main issue would be 'leakage' between wavemakers in operation. Even a dry-back system has gaps between individual wavemakers, with flow induced by the pleats, bellows, or membranes. Is a series of wet-back wavemakers elements with 'reasonable' gaps between elements and tank walls a no-go from a modeling standpoint? Avoiding time spent on seal design and maintenance, and wear on plastic glazing would be advantageous.

 

Found a really cool paper:
The magic carpet: an arbitrary spectrum wave maker for internal waves"
by Thomas E. Dobra, Andrew G. W. Lawrie, Stuart B. Dalziel

The magic carpet: an arbitrary spectrum wave maker for internal waves - Experiments in Fluids https://link.springer.com/article/10.1007/s00348-019-2811-5

Small perturbations, but they built a working wave maker with bottom-mounted actuators that is akin to the 3D, conformable ADAPA mold.

 

Dr. O'Dea and I had many interesting discussions about waves, wave testing, wave measuring, structures, data, and "what really matters" during our intersecting careers. I have nothing but the highest respect for him; he was a scientist par excellence and I was a lowly engineer only interested "red light/green light" results. I would suggest you read everything by him and Tim Smith that came out of NSWCCD about waves and wave makers.

 

Thank you for recommendation. Getting to 'what really matters' often isn't easy.

S.H Salter's work on wave generators has intrigued me for quite a while. He approaches design problems with few pre-conceptions, and is dedicated to hands-on experimentation and development of practical lab tools.

I found his 1981 paper "Absorbing Wave-makers and Wide Tanks" online and have attached a PDF here. It is wonderfully accessible. You could build stuff from the descriptions and diagrams.

The 1969 thesis by Eric Reginald Chappell titled "Theory and Design of a Wave Generator for a Short Flume" reviews an incredible number of wave-maker mechanisms. He covers about any approach one could imagine. A PDF can be found here:
Theory and design of a wave generator for a short flume https://open.library.ubc.ca/soa/cIRcle/collections/ubctheses/831/items/1.0050588
Several examples vary stroke or path of the paddle face with adjustable to mechanical linkages. While elegant, they would be time-consuming to fabricate and to adjust. Discrete elements with servo control would seem to be more practical 'nowadays.' Almost-instant gratification through software and electronics.

My gut says a 3D wave produced by a flexible sheet in the manner of the ADAPA mold or the aforementioned "Magic Carpet" wavemaker would be optimal for getting the wave you want as close to the wave maker apparatus as possible. My gut also says it'll be hard to generate that wave in any size other than a ripple due to requirements of pushing the surface into a complex, compound curve. If the diaphragm is elastic enough to conform to deep curvature variations, it will also lack sufficient rigidity to produce the desired wave front. Self-dampening would be an issue.

Constructions as used in the rigid or semi-rigid, 3D conformable molds could solve the flex problem, but they are designed to be fixed during use. They often use vacuum-bagging together 2 auxetic surfaces cut from rigid material to lock-in the shape. Or thicker buffer membranes or meshes are bagged onto the initially conformable surface. Solvable for dynamic motion without excessive complexity? Maybe...

Would a grid of adjacent, discrete 'pistons' despite flow discontinuities, be a reasonable approximation to a continuous 3D surface for a horizontal wave generator? Would the difference between final wave forms be minimized with sufficient tank length?

 

Be care that the o-rings of cylinders does not produce a stick-slip effect at each motion inversion, because they can noise the input signal and lead to a wrinkled water surface instead of a smooth one. Especially detectable when producing a regular wave, explanation :
** the stick-slip effect at each motion inversion of the cylinder will give a rectangular signal which will superpose to the input sinusoïdal one.
** the Fourier decomposition of a rectangular signal is equal to : sinus first order + sinus third order + sinus fifth order + ...
** the first order is not a problem, it is embedded in the "gain", i.e.the input amplitude of the mechanism versus the output amplitude of the wave.
** the fifth order and over are no longer a problem because to low energy within.
** the third order is the problem because the energy could not be négligeable, the period being 1/3 times the one of the main wave >>>, so the third order wave length is 1/9 the main wave : you can see 9 small waves on the main wave length, so the impression to see a wrinkled surface.

 

Dolfiman - I'll warn you, my last college math class was multivariable calculus 40 years ago, and I eventually wound ended up with a BA in Art. But this all makes sense and thanks for advance warning on stiction. Whatever type of wave maker I end up designing or adapting, I'll strive for low mass, low friction, and PID control of the actuator(s), with force or pressure instead of position feedback at some point. There are many low-cost and open-source options for control systems using brushed or BLDC motors now. HTD timing pulleys and belt drive systems are easy to put together. A lot of this stuff has become readily available due to Maker culture and the growth of robotics competitions. Prototyping mechanisms is easier than it used to be.

 

The aforementioned 1969 thesis 'Theory and Design of a Wave Maker for a Short Flume' included a bit on causes of transverse wave formation on wave maker plates. This pic illustrates a couple of methods to minimize this problem. The author notes that the use of filter plates in front of the wave maker results in a lower wave height - I'd presume due to pressure drop across the filter.

I can't say that I find vertical ribs or filter arrays in many of the images of online wave makers. Here are some on paddles of an HR Wallingford design:


Edinburgh Designs has video footage of an array of their piston wavemakers keeping a wave front straight in a flume. No ribs. Using force feedback vs position feedback on the paddles, they overcome resonant effects off of the tank sides:
.

I'm reminded of the observation of John O'Dea ['Numerical AnaNumerical Studies of Directional Wavemaker Performance'] that it appears the discontinuities across piston faces on a translating piston array vs. a horizontally-articulated array ('piston snake') didn't matter much on the desired wave a distance down the tank. All of the above makes me wonder yet again if a 3D wave maker composed of individual round pistons transitioning to square cells is non-starter. Here goes:



Thoughts for prototyping materials include: Delrin pistons With O-rings; PVC pipe cylinder with milled flats; CNC-routed PVC plate manifold; 3D printed or cast transition ducts; Plascore PC or PP honeycomb (1/8" dia. x 1" t cell). Incidentally, the 90% open area of the Plascore square matches the piston area - this was not planned. In a lab wind tunnel, flow straightening typically happens at the upstream of the test section over an area 9-10 times larger than the test section. Less pressure drop with low velocity flow. I know how it would work out here, other than I'd expect a proportionally larger loss through the honeycomb, particularly at high stroke rates. In some ways, this isn't so different than the laminar flow straighteners used in pressurized water lines. in industry, or to make arty fountains.

 

I played around a bit more with the geometry of the wet-back wave maker from my initial post. Dimensions are in inches:


This is sized for a nominal 24" deep flume. Water level would obviously be lower if you were generating bigger waves. It ought to be able to emulate anything from a top-hinged to a bottom-hinged wave maker, and anything in between. The geometry shown seemed to be a reasonable compromise between reasonable unit size and decent stroke length. It's still a long unit. Buoyant forces will be large. The cams could be rod actuated, but this drawing is based on a timing belt or curved rack and pinion drive scheme. The 'extra' displacement at the bottom of the area displacement portions (light blue) are a given with the geometry. This area is quite small in piston mode, and larger in the bottom hinged flap mode. I'm not sure how big of a deal that is.

 

Looked long and hard for this one:
"...a translation of a series of four French articles by F. Suquet.l' Fo Bi~sel,9 ,and a group of engineers at the Laboratoire Dauphinois dlRydraulique (Neyrpic.)"

The original articles were written in 1951-52. The Saint Anthony Falls Laboratory at the University of Minnesota translated the articles into English in 1954, as part of a research contract with the David Taylor Model Basin. The official title is 'Project Report No. 39:Laboratory Wave -generating Apparatus.' A reduced size PDF is attached.

The SAFL is in Minneapolis, only a few miles from me. I've never visited. Time to fix that!

 

Here are the original articles in French, in case of :
LES APPAREILS GÉNÉRATEURS DE HOULE EN LABORATOIRE (shf-lhb.org)
and
LES APPAREILS GÉNÉRATEURS DE HOULE EN LABORATOIRE (shf-lhb.org)

 

Thanks, Dolfiman. A higher res PDF of the English translation is available at http://conservancy.umn.edu/bitstream/11299/108277/1/pr039.pdf. Chrome doesn't like it - 'not a secure site.' Use at your own risk, but that's where I downloaded it from with no issues.

The document does review a flexible flexible membrane 2D wave maker, and has a photo of it. Looks like it's from early-mid 1900's. There is also an illo of a 3D membrane wave maker from a 1916 magazine article. Can't figure that one out from the illo. Anyway, mention is made of how precisely a wave can be developed close to the paddle face. Advantage is a clean wave in a much shorter tank. Disadvantage was cost and complexity of the mechanism and seals. Remember though that it was built when related motion this complex had to be driven by mechanical linkages. Not the case now with software and servo systems. The article also mentioned that making a longer tank to accommodate a simpler wave maker had costs too. Not just more $$$ into the tank, but requirements for a larger facility.

 

The above does lead to a specific question: What would be the disadvantage of separating a series of individual wave makers, be they piston or hinged flap, by a thin vertical wall between each element? It seems to me this is analogous to using vertical fences on the face of a paddle or the walls separating chambers of a 3D pneumatic wave generator.

 

Just received this book:
PHYSICAL MODELS AND LABORATORY TECHNIQUES IN COASTAL ENGINEERING by Steven A. Hughes.
jehardiman, dolfiman, and others: A nice compendium on physical models for wavemakers and related coastal processes. What I was hoping for as a bridge between theory and 'practice' in physical modeling. Starts at Dimensional Analysis and Principles of Similitude and works its way through physical modeling at scale for wave generation, sediment transport, etc.

What I'm liking so far is that although described as a 'graduate-level text', it is comprehensible to one who's math ended with calculus and intro diff eq, - 35 years ago. It's a bridge between analysis and physical modeling without getting into to the specificity of computational models or analyses that one comes across in many theses, dissertations, or reports. Nuts and bolts. Its about correlating phenomena that can be observed at 'model scale' with 'full scale.'

 

This just out on the NSWCCD MASK Tank.