The Power of the Waves

Jospa Blog

May 1, 2009

An account of the April tank test is posted below - the background noise is the wave generator. This test was most encouraging and we believe we are well on the way towards a robust proof of concept. Wikipedia defines this as “Proof of concept is a short and/or incomplete realization (or synopsis) of a certain method or idea(s) to demonstrate its feasibility, or a demonstration in principle, whose purpose is to verify that some concept or theory is probably capable of exploitation in a useful manner. A related (somewhat synonymous) term is “proof of principle”.

Filed under: Uncategorized — admin @ 7:24 pm

IRISH TUBE COMPRESSOR SCOPING TEST AT HMRC 10.35 WEDNESDAY APRIL 8TH 2009

SUMMARY

 

The diameter of readily available tube for test is limited: JOSPA has been using a 50mm dia tube. The question of whether this could work in the tank at HMRC had arisen, and the development had been stuck for some time on the relationships between tube diameter on the one hand and friction and/or viscosity losses and air and water separation/commingling on the other. We hoped this test would at least shed useful light on these relationships and enable us to draw up and proceed with further development; any more would be a bonus.

The results were beyond expectation in describing parameters: the ‘efficient’ parameters are described below. They were certainly encouraging enough to make us feel we are on the right track and that proper development will make the technology work. It also provided strong suggestions on what the immediate next steps and beyond in the development should be. 

 

NOTE: We emphasize that this test cannot be regarded as strictly scientific in the context of proof of the concept, as we introduced the external force of mains water.

A. DESCRIPTION

 

PURPOSE: This test was an attempt to ascertain the band of wave frequencies and heights at which the 50mm dia tube might be suited for tank trials.


METHODOLOGY: Mains water was injected into the tube intermittently via solenoid valves and a crude venturi (the purpose of the venturi being to entrain air). The adjustable parameters were varied and the effects on the flows in the tube were observed and videoed at different settings.

PRESENT: For HMRC: Brian Holmes, Tom Walsh   For JOSPA: Joss Fitzsimons, Stan Kucek

PROCEDURE: Before starting, BH suggested doing a ‘benchmark test’ with no waves without video.  This proved to be a wise move.

B. TESTS

 

No waves

 

1) (Benchmark) Solenoid timer was set at 0.4 seconds and Tom pressed button at (approximately) 1.5 second intervals.  The pattern of air and water inside the tube was not segmented or defined.  Little air seemed to come out the end of the tube.  No regular air, water, air, water pattern was noticeable at the exit.  There was no evidence of a regular pattern at the exit or indeed along the tube.

              Simple Waves

2) With simple waves on. Initially a few trials were tried with periods of 1.5 seconds and 2 seconds.  The solenoid opening times were varied from 0.3-0.6 seconds. A higher opening time could be expected to give a much higher flow, but as these valves were slow acting it was a trial and error matter. Especially with regard to how much air was induced by the water. At no point did the 2 second period look right or as expected.  There were not discrete air and water segments.  The jet of water from the nozzle quickly stagnated on entry to the tube and the little air induced rapidly stagnated, just moving to and fro with little drift.  The appearance was as unimpressive as the original piston trials and the OWC attempts. There was no evidence of the water surfing within the tube or what surfers call “Catching a wave”.

3) The 1.5 second period fared better. I have no written notes but the (short) video shows air and water segments, moving but not racing forward.  An improvement over 2 seconds period, but not the fast moving slugs we were to see later.  All the subsequent trials were at 1.2 or 1.1 seconds.

4) Period 1.2 second, valve opening 0.4 sec, wave height 100mm. From my notes, “Definitely much better”. By better I mean discrete air & water segments racing forward at wave speed. The air segments were clearly running at the wave crests but as they were moving so fast it is not possible to say if the wave led the air slugs or vice verse.  This was dramatic and resembled what is expected to happen at sea.

5) The same (1.2 second & valve opening 0.4sec.) but with wave height 150mm. From my notes: “Better still”.  The extra height of the waves gave further distinction between air and water segments.  Most importantly with these tests at 1.2 second period, the distinction was clear almost to the end of the functional 12m of the tube.

 

6) We tried to change the solenoid opening time on the run but this seemed to distract from the button pressing for the solenoids, and both notes and video are poor due to stoppages. Over the next few trials it became clear that we were not improving things by changing significantly from a solenoid opening time of 0.4 seconds.  The useful solenoid opening time range for the solenoids lies in the range of 0.3-0.5 seconds, on these 1.2 second waves.  Shorter opening times delivered too little water and induced too little air.  Above 0.5 seconds the flow from the nozzle is almost continuous when using wave periods of 1.2 seconds, but would suit longer wave periods.
        
7) A few attempts were made at a period of 1.1 seconds. Some success is evident from the videos but there were problems due to unusual wave behaviour which seem to have resulted from reflections.  An attempt with a wave height of 200mm does not show the expected improvement with this higher wave. In fact at 200mm the behaviour worsened, with a lot of break up of the air/water segments. At 50mm wave height the behaviour was better but best at 100mm.  We had already established during setting up that a period of 1.1 seconds was at the lower limit of what our modified solenoid valves could do.  That was one problem.  We seem to have also had a problem with reflections, but I am not sure that this is the case with low amplitude 1.1 second waves.  For whatever combination of these reasons, the results using 1.1 second waves were not as good as with 1.2 second.

Panchromatic waves


8)  TEST WITH PANCHROMATIC WAVES.  This was tried without the expectation of an immediate positive result. But this is what happened.  Tom judged when to activate the solenoids while watching the oncoming waves. The air water definition and flow looked surprisingly good, just as well as with the simple waves at 1.2 seconds. The main difference was that with ‘lulls’ in the wave train the speed of flow dropped but never stalled, then as the oncoming waves increased the flow speed increased in response.  This was an unexpectedly good outcome.  A second panchromatic trial was then tried but this time with the solenoids operated on a regular time basis with no reference to the oncoming waves.  The result was very poor by comparison resulting in an almost immediate stall at the entry to the tube.  Clearly there is a case for monitoring oncoming waves and controlling an ‘intelligent’ feed system.


C. SUMMARY ON SIMPLE WAVES:

 

We got consistent ‘good results’ using a period of 1.2 seconds, wave heights of 100mm and 150mm, and solenoid opening times of about 0.4 seconds.  By good results I mean discrete segmented water & air moving rapidly along to the exit, with the air segments running near the wave crests. There was some variation, but this is explained by the rough & ready in-feed system and the fact that the timing & phase were manually controlled.  In all cases a higher ratio of air would have been better, but we were already aware of flow limitations through the solenoids.  The fact that the water and air segments were not inclined to mix, but remain discrete, is evidence that while friction within the tube is indeed serious it is less than feared.  This is a valuable result of this trial. 

Longer and shorter periods both showed substantially poorer results than the 1.2 second period.

 


D. DISCUSSION ON THE ABOVE:

 

These conditions define a small operating window to be sure, but one where the surfing behaved impressively and at speed.  The only available explanation for the big difference between a 2 second period and a 1.2 second period is the difference in tube friction resulting from the difference in wave speed. A visible difference resulting from the change in wave period was expected in advance but fortunately the period at which the change happened was longer than feared. The attempts at shorter periods of 1.1 second were not so successful and are probably attributable to the slowness of the solenoids, manual timing and the crudeness of the venturi design. These trials provide evidence that we can indeed use a 50mm bore tube in tank tests, but within only a small window of these variables.  The tube friction at 1.1 or 1.2 second may still be too high to clearly demonstrate an increase in head due to the waves alone.  However, with this new knowledge we are now in a good position to open this operating window so that tube friction becomes less of a worry.  Small changes aimed at reducing the period could make a big difference.  Quicker solenoid speed, automatic solenoid switching and better venturi design could permit us to use a substantially shorter period – perhaps less than 1 second. This would result in lower speeds within the tube and substantially reduced friction (which is roughly proportional to the square of the velocity). This is critically important in whether we can ever achieve a demonstrable pressure increase due to waves alone within this tank using this 50mm bore tube.  With a better air/water feed system we can be reasonably confident that tube friction will not wipe out any gain in head produced by the waves.E. FEED SYSTEM REQUIREMENTS:

There are many ways of feeding air & water into the tube, whether using pistons, an OWC, solenoids or a revolving or other arrangement.  Regardless of what method we use we will have to ensure that the total flow air & water is even and not stop/start (even if the two are independently stop/start).  So what would be the ideal system?  The combined flow of air and water should be variable around a small range to ‘tune’ to a particular linear wave speed.  Control of the ratio of air to water would also be desirable.  A pressure sensor (or perhaps a damped manometer) should be located at the entry to the tube.  This would be the reference, above which any pressure gain within the tube would be a positive.  This pressure reference will become very important as it is difficult to know what pressure we are putting into the tube by any other means.

THERE ARE THREE POSSIBILITIES TO CONSIDER

1. REVOLVING FEED: A revolving feed similar to early revolving demo. unit was planned last Christmas, but fear of tube friction led us down this long road of testing with solenoids.  A revolving feed would not have the delayed action of the solenoids and so shorter periods would be possible.  The quantity of air taken in by a revolving feed would be much better defined than with the solenoids, and variations due to manual timing of the solenoids would be eliminated.

A revolving feed would however bring some new difficulties to be overcome.  As a revolving tube dips in and out of the water surface the amount of air and water it takes in will vary greatly if the timing is not exactly in phase with the waves.  If the rotating tube dia is the same as the working tube dia, we must have a circumference length of approximately one wavelength in the tank.  Thus any simple revolving feed design will be tied to one, invariable, wavelength, quite a limitation.  To overcome this we may need to have a wave-free chamber located within the tank so that the speed of the revolving feeder can be varied without causing a mismatch with the tank waves.  This wave-free chamber could extend from near the bottom of the tank, where the wave movement is minimal, to above the wave crests. The chamber could be designed to present a narrow obstacle to the oncoming waves, if the axis of rotation is perpendicular to the wave direction and has a 90 degree elbow following the rotating seal and leading to the working tube. It would allow a steady water supply from below the wave movement.

We will also need to consider centrifugal force.  Centripetal force will tend to suck the air/water slugs back out of the tube and into the swirling inlet.  To reduce this effect we may use a larger dia tube for the rotating in-feed and a smaller radius of rotation.  For example if the cross section of the rotating tube is double the area of the working tube, we can halve the radius of rotation and still match the flows.  This would halve the centripetal force.  It may even be possible to have two separate rotating in-feeds to the one working tube; this would halve the rotating speed and quarter the centripetal force.

Another idea is to mount an in feed tank with an independent water supply above the top of the wave surface. This would pressurize the working tube by about 0.8m but the differential at the exit should remain the same.  A difficulty we need to be aware of in mounting the wave free chamber above or to the side of the wave tank is the tendency of the air to separate from its associated water.  Siphoning from a wave free tank over the side of the main tank would present this problem.  Whatever method we use, it would be desirable to measure the pressure at the start of the working tube using a fixed manometer of small bore to minimize fluctuations in the reading.

2. LONG REVOLVING DEMO UNIT: Also discussed with Brian Holmes is an improved revolving demo unit.  The present unit is neat and simple but it does not scale well to a realistic situation.  The angle at which waves rise from the horizontal are invariably less than 30 degrees, but our demo unit has steepness up to 90 degrees on some loops.  By stretching out the axis of the roller and gradually spiraling the tube along the length a much more realistic comparison could be made.  Furthermore by choosing all three - the roller diameter, the tube diameter and the roller length in proper scale proportions - a highly representative model could be made.  It would be far easier to feed than a tank set-up.  While far longer and more expensive than the present revolving unit it would be much more representative and would undoubtedly provide very good research information upon which to base a successful sea trial.  Indeed it poses the question as to whether conducting a sea trial without first building such a unit would be wise.

3. OTHER TANK METHODS

: Every approach has its own drawbacks. For example we could feed both air and water at an even flow through two flowmeters.  Then the waves themselves would sort out the segments as the air rises to the crests and the water to the troughs.  However this even input, while attractive, would use up some of the short tank length as the water & air sort themselves out as well as presenting the difficult problem of measuring the pressure in a tube which is flexing up and down. 

A possibility suggested by Brian Holmes involves the use of a long (270m) UK - based towing tank.  With such a length we could get much closer to actual scale, with all its attendant advantages.  It would involve much greater expense than heretofore.

Powered by WordPress