Minutes of the Meeting on 13 November 2000

Present:     Warren Burrower (Greenland Tunnelling), Chris Greenshields (King’s College), Simon Pugh (ESDU), Arris Tijsseling (Eindhoven University), Jim Brown, Della Leslie and Alan Vardy (Dundee University).

Chairman:                   Simon Pugh

Minutes:                   Della Leslie

Meeting commenced at 10.00 hours.

Apologies:     Keith Austin (Flowmaster), Bruno Brunone (Perugia University), Martin Hamilton (EUTech), Anton Heinsbroek (Delft Hydraulics), Arno Kruisbrink (Delft Hydraulics), Ruud Lemmens (Delft Hydraulics), Bjørnar Svingen (Kvaerner), Geoff Taylor (GT Acoustics), Patrick Vaugrante (EDF), Lixiang Zhang (Kunming University of Science and Technology).

In the absence of Keith Austin, who was unable to attend, Simon Pugh chaired the meeting (although Warren was recommended as chairman). With Jim Brown attending the meeting, each person introduced themselves.

The apologies were then given, including: Geoff Taylor retired at the end of June from EUTech. He is now an independent consultant (GT Acoustics) and regrets that he will no longer be able to participate in FSI meeting. Martin Hamilton hoped to attend but due to restructuring in the company was unable. Patrick Vaugrante was unable due to illness; the group wished him a speedy recovery. Also apologies were sent from Mohamed Ghidaoui whom had been invited and hopefully will be able to attend next time.

The meeting began with a general discussion and recent developments. Bruno Brunone has recently been appointed to Chair of Hydraulics. Simon commented on the Minutes of the last meeting, lots of support for our project on guidelines. Arris informed us that Lixiang had visited him in Eindhoven during the summer. Lixiang’s work was discussed (flutter, buckling and FSI). Arris also told us of a recent review paper (submitted) written with Dave Wiggert on the developments of FSI in the last ten years.

Chris Greenshields has a new job beginning in January at L Nabla ltd., although he will spend one day a week at King’s College. He hopes to continue contributing to the group. His position at King’s is to be filled by Jason Reese (presently at Aberdeen), who has been invited to join the group and has expressed interest. Chris discussed some of his past work.

The possible extension of the group was discussed. Simon suggested that we need to define FSI for the group: internal flows only. Arris commented that research in blood flow combines FSI and unsteady friction. Research groups in Aberdeen (Reese, Gorman) and Bath (Johnson, Longmore) were suggested as possible contacts, and to find out in more detail the direction of their research. Alan was anxious to strengthen the numbers of members with relevant practical experience.

Item 1.                   Review of FSI guidelines, Worldwide

Della Leslie gave a presentation reviewing FSI guidelines, worldwide. This gave an overview of a document distributed prior to the meeting. The countries in which standards have been examined are Germany (DVGW: W303), Netherlands (NEN-3650), America (ASME OM-S/G-1990), UK (BS8010-2.8) and Australia (AS2885.1, yet to be obtained). It was noted that Australia had been missed from the review document. It was quite clear that there was very little regarding FSI: only Germany and the Netherlands include any reference to FSI in their standards. In particular it was noted that there are many standards, with each aimed at different types of pipes (e.g. plastic, iron, steel), use, etc. with codes covering every aspect of the systems (design, construction and installation). The large quantity of data available was highlighted with Della illustrating by showing the group a very large pile of papers (several hundred pages), which were only some of the British standards Other published work was outlined and it was noted that of the four items mentioned, three of them had connection with this Advisory Group. The four items were work by Lavooij and Tijsseling, ESDU, Hamilton and Taylor, and a course by the Pipeline Compressor Research Council (attended by Arris).

There is a large amount of data available on pipelines in the US. This includes statistics on accidents, listing causalities, cost, cause etc. It was noted that the cause of a large proportion of accidents is listed as “pipe failure” or “other”. The American pipeline legislation (Title 49 USC and CFR) includes design, construction, inspection, testing, operation and maintenance. It even includes guidelines for the implementation of anti-drug and alcohol misuse programs for operators. Della told the group that the Bill S2438 had in fact failed to be passed by the House, and that this appeared to be politically motivated. It had passed through the Senate with 99-0 votes, but failed to obtain the 2/3 required in House with clearly a Republican/Democrat split. She gave a small extract (several pages) of proceedings from the in House discussion of the Bill. Simon highlighted from this the number of objections to the Bill, particularly from industry.

Discussion of the standards and related topics continued throughout the presentation. Simon asked who reads those things? The completeness of the review was discussed. Arris suggested looking for standards from Japan and France. Chris suggested a contact in Australia for more information on codes there (CSIRO).

Simon asked if Delft was involved in NEN-3640. Arris said yes, the Committee visited Delft hydraulics to witness FSI tests. It was noted that the German code is based on work by Burmann. It was noted that there seems to be a connection between research in the country and the inclusion of FSI in the standards. Simon asked why there was no FSI in British Standards? Arris suggested looking at the ASME B31 code and handed in ASME Journal relating to this. Arris commented that the ASME – OM standard was for the operation and maintenance of nuclear power plants and suggested looking for other nuclear codes (e.g. Germany and UK).

Simon said that education was a problem in developing standards, with most people working on them not aware of FSI. He said he was cynical about safety issue, stating that engineers are expensive and companies will just pay for work when things go wrong. Chris spoke of work by a water company replacing old pipelines: Logging of loads of data on pipes and using statistics to “predict” failures; the data is not available to third parties. Arris asked if there was an overlap with work by Ruud Lemmens in the Europeans project? Della said that he has expressed interest in the Review document. Simon asked if the standards were all for long pipelines? No, different standards cover all aspects of pipes including piping in connection with boilers. Simon said that there was no need to go much further and that we done a very good job! Both Chris and Simon said that the Australia codes were quite good.

Item 2.                   FSI Research in Dundee: The next 2 years.

Della Leslie gave a presentation about the progress and direction of the current project at Dundee: “Development of practical guidelines for fluid-structure interaction in pipe systems.” First, the work completed so far was discussed, including the review, selection of subsystems and the beginning of analytical/numerical work (this included Mathcad programs for frequency domain work and Fortran programs for time domain work. Programs have been developed that examine a single pipe and a single elbow, with testing on-going to ensure consistency. Arris asked if all the time domain analysis used the method of characteristics? Yes, so far and it was asked if this was the most practical approach? It was explained that the answer is probably “yes” (at least in Dundee’s work) and that it is certainly a sufficiently accurate approach. Some sample results were shown. These showed natural frequencies determined from a single pipe system with a lateral spring at each end (one direction only) and subjected to a lateral impact at one end. Only lateral motion was considered. All types of motion will be considered in further work. The resulting graphs illustrated the variation of the natural frequencies with spring constant: the first two modes were seen to tend towards zero as the spring constant became smaller. A discussion of modes and frequency evolved. Simon suggested a good book for reference by Blevins. He also explained very clearly the modes, with the first two modes being rigid body motion. Arris asked if any damping was included? It was explained that damping is a third order effect in the pipes themselves. Damping might be included in support conditions. A second example was also examined, again a single pipe, considering lateral motion and subject to a lateral impact but with a rotational spring at one end. The system appeared to behave strangely and it was possible that mistakes in the analysis had been made. Della said that she would examine it again to check. Alan explains that the purpose of the examples shown so far is to test the computer software. This is best done with the simplest mathematical conditions. The application to realistic physical conditions comes later.

The choice of subsystems was then presented, together with some of the possibilities to include in the analysis (i.e. what factors should be included: geometry, material properties, support conditions, method of excitation It was also explained that larger systems would also be examined with subsystems as components. Discussion of the use of the subsystems followed. Alan emphasised that that the guidelines are not intended to be the last word in assessing systems. They are intended to help engineers determine whether particular systems are likely to have a problem. Arris said we need a reference to compare FSI results with, e.g. classical waterhammer. Simon asked what the worst case would be?

Simon asked about EDF? Della continued by outlining the planned experimental work at both Dundee and EDF. It has been proposed that EDF will examine subsystem 5 experimentally in BANCO and a second static rig. This would allow for comparison between open and closed systems. Also it provides the possibility for periodic excitation Della said she would be spending some time at EDF next year, to help where possible with experimental work and also to use in-house software CIRCUS. In Dundee two 3-D systems are to examined experimental, one of which corresponds to subsystem 4 and the second a close variation of subsystem 6 but with the T-joint rotated through 90 degrees. Alan said the sole purpose of the experiments would be to assess/validate theoretical models. A discussion on subsystems and the experiments evolved, which continued at the end of the meeting. Alan asked if the subsystems were O.K.? Simon asked what sort of things cause problems? He suggested simple things, e.g. low-pressure at top of a hill, high pressure at bottom of a hill. Supports were discussed (e.g. frictional). Alan asked whether we should be considering single pipes, branched, or network. It was noted that in the comments sent by Martin, one of the systems mentioned is highly branched. Chris suggested examining fewer subsystems but more variation in boundary conditions. This was discussed, resulting in the discussion to examine systems 1,2,4,5 theoretically, i.e. remove branched systems (system 6) and remove system 3 (although this can be considered a special case of system 4). Experimentally at Dundee it was agreed that both experiments should be 3-D (system 3 and a variation of system 6 with T-piece rotated through 90º).

Item 3.                   The Trans-Arabian Pipeline

Arris Tijsseling gave a very interesting presentation on the Trans-Arabian Pipeline (Tapline). The presentation was based on a report by John Makkinje (1951). The pipeline is approximately 1213.6km long, and runs between Saudi Arabia and Lebanon (via Jordan, Syria). It was the largest pipeline of the time, and was an American Joint Venture (Esso, Chevron, Mobil, Texaco). Arris said the politics of the pipeline included the CIA (1949) and sabotage (1967).

Designed by Bechtel: 1200km long (600km are above ground), 900mm diameter with supports every 20m (A-frames up to 10m high) and contained no expansion loops. There were also heavy anchors at bends. The building of the pipeline took from 1947-1950, using 16000 men. Arris then described some of the aspects of operation, testing and inspection. The pump station increased the line pressure to a maximum of 64bar and hydrostatic testing was carried out. Inspection showed bullet holes in the pipeline! This could be explained by two factors: gazelle and rabbit hide around pipeline and hunters caused the damage, plus also during testing the pipeline was filled with water, and is in a desert. Longitudinal movements of the anchor/supports were also found, but why this occurred was not determined. The study of pipes conveying fluid was invited by Ashley and Haviland (1950) in an attempt to explain vibrations observed in the Tapline. There have been a number of accidents involving the pipeline, e.g. Trucks hitting the line, sabotage and fire. Arris gave an example in which a vehicle hit the line causing a rupture: the pipe was at 60bar pressure; with 10km length till next check valve station. The vehicle was abandoned and the heat from the engine caused ignition of the oil. The result was a very large fire!

Arris then described the possibilities of waterhammer in the pipeline: the capacity was increased through auxiliary pumping units, which were unattended and could trip off without warning. Pressure waves on top of the increased steady-state pressure were anticipated to be disastrous.

The pipeline was “mothballed” in 1976, although deliveries to Jordan were maintained.

Arris explained why he did this presentation (what lead him to it). It started with Della asking for more information about NEN3650 (a Dutch standard). On obtaining a copy, one reference of interest was by Ludwig and Johnson (1950), whose work was in fact special cases of recent work by Leslie and Tijsseling (2000) on the attenuation of wavefronts. Ludwig’s work is based on the Tapline, hence presentation. Arris told the group of new work on jump attenuation in accelerating/decelerating flows.

Item 4.                   The dangers of waterhammer and beer

Chris Greenshields gave a presentation based on some work completed during the summer. This was for a company, A-can technology, who are developing a product – a new design of container. The problem is that during events/concerts it takes too much time to pour a pint of beer and bottles are unsuitable (often a no glass policy). A possible solution is a cross between bottle and pint container (i.e. pre-packed pints). One aspect of safety to consider is the throwing of bottles. Question: Can safety be improved?

The aim of the project is to get an idea of impact force of a thrown container vs. distance thrown. This was divided into three phases. Phase 1: Throw characteristics, i.e. straight-line speed and rotational speed (affects emptying of liquid). Phase 2: Drop impact tests and emptying tests. These gave plots showing peak impact vs. volume of liquid and volume of liquid vs. time. Phase 3: Bottle characterisation.

Five containers were tested: Bottles with opening sizes 55mm; 30mm and 19mm, UPbottle (rigid, Evian type bottle) and an aluminium can (volume of each bottle were about 440ml). The emptying tests showed great differences between the bottles. The 55mm bottle took less than 1 second, whilst the 19mm bottle took greater than 12 seconds. Chris said that the water loss rate appeared to be proportional to the area (of opening) squared.

Impact tests were performed in which the containers were dropped, repeating for various water volumes. The test speeds were 2.15,4.3, 6.45 m/s, and water volume included 0, 75, 150ml and full containers. The reproducibility of the experiments was very good. Chris showed plots of typical waveforms (impact force vs. time) for each container for full containers dropped at 4.3m/s. These showed some clear differences between bottle types. Plots of peak force vs. impact speed showed a linear relation, whilst force vs. liquid volume gave plots that appeared to increase, then ‘level off’ with increasing volume.

The impact of a long tube using classical waterhammer theory was discussed (Joukowsky formula). Chris noted that impact force does not vary with volume, impact time is proportional to volume (and height), and impact force is proportional to impact speed. How does this compare to the containers?

The wave speed from classical waterhammer theory was calculated for each container by determining representative data (i.e. diameter, thickness, Young’s modulus, etc.). Plots were shown of waterhammer theory against experiment. Each plot showed that the theory and experiment had the same gradient, but for the data to match a time offset was required. This varied between containers (from 0.3ms-1.2ms). The question was asked: why is there this offset?

Chris tries to explain the offset. He said there was possible a connection with work by Reid (2000, Int. J. Impact Eng. 24, 133-153) which related valve closure time to the period of oscillation of radial deformation, t. Suggesting that the offset equalled 0.5*tau? Alan asked about deformation of the containers. Chris asked whether just deformation of the base or the whole container. He showed the group examples from each container. Further discussion of this followed. Alan referred to the Dundee experiment rise time, suggesting some sort of connection. Arris asked the experiments related to Joukowsky – rough calculations showed the behaviour was Joukowsky. Alan suggested a method of reducing the force was to reduce the wave speed. Chris said that he though the rate of emptying was the most important factor. Final Chris outlined some future work: he has an undergraduate student working on bottle emptying; a PhD student working on fluid transients in flexible tubes and on 1 Jan 2001 will begin work for Nabla ltd.

Afterthought: It seems possible that some people with malicious intent will get hold of the screw tops and hence throw missiles with very different characteristics. Really evil throwers might develop skill in ensuring that the lid falls off on impact.

Afterthought-2: Perhaps we should have a competition at the next meeting to see who can throw the greatest quantity of liquid over a specified distance!

Item 5.                          Place and date of next meeting

The date of the next meeting is yet to be decided. Arris proposed to have a combined FSI/unsteady-friction meeting, one day before the IAHR Work Group Meeting next June in Trondheim, Norway, (Similar to what we have done in Delft this year). That is to hold the meeting just before the conference and those who wished could also attend it: there is no registration fee and one of the named topics is FSI. It would also give the opportunity to visit Bjørnar laboratory, and could bring some other participants.

Alan expressed concern of whether members of this group would be able to attend (due to time/cost). The next meeting will be very important for the development of the current project here at Dundee. The decision is left open awaiting feedback from all those concerned. Sample costs will be determined.

Example: For a meeting on Sunday and Lab visit Monday a.m.

                   Saturday                   London (07:40) – Trondheim (13:50)

                   Monday                   Trondheim (18:20) – London (21:05)

Cost from London (Heathrow): Approx. £250 airfare +2 nights hotel + subsistence

Item 6.                          Chairman’s Closure

Closure of meeting at 17:00 hours (approximately).