PHOENICS Developments Work in progress

1
PHOENICS DevelopmentsWork in progress

• Main activities are
• Assisting the new user
• Revealing buried treasures
• Meeting the needs of end-users in
  special-application sectors
• Building on existing strengths
• Providing new features

2
1. Assisting the new userthe Commander, 1

• -- The PHOENICS-Commander top page has a
  'new-user' button
• -- clicking on it leads to welcome, a quick
  start, a slower start, tutorials, and
  ready-to-run cases.
• -- But not enough cases have been selected,
  polished and provided with new-user-suitable
  words.
• -- and tutorials need updates and augmentation.

3
1. Assisting the new userThe Commander, 2

• The Commander has multi-language structure.
• Each language has its own dictionary and its own
  help file.
• BUT only the English and Russian are available.
• Who will provide the other-language equivalents?
• It is quite easy for a native speaker. I shall be
  glad to explain what is necessary.

4
2 Revealing buried treasures1 In-Form

• In-Form is an immensely valuable feature of
  PHOENICS.
• More documentation and exemplification exist for
  it than for any other feature.
• No competitor has anything so powerful. (Or so I
  think. Am I right?)
• BUT it has still not caught on.
• Why not? Advice will be welcome.

5
2 Revealing buried treasures2. PARSOL

• This cut-cell technique is recognised by most
  users as economical of users time and effort.
• BUT some are persuaded (by competitors ?) that
  body-fitting grids are inherently more accurate.
• We need benchmark comparisons.
• Has anyone made some which they could/would
  contribute?

6
2 Revealing buried treasures3. MUSES

• Multiply SharEd Space is a valuable technique,
  providing multi-phase capability.
• It was introduced into PHOENICS (uniquely, I
  believe) several years ago but it became buried.
• It has been used (only) for heat exchangers and
  blast furnaces .
• Because it employed PLANT statements recovery
  involved conversion to In-Form.
• Carrying out this conversion has opened the door
  to NewMuses (discussed below)

7
3 Meeting the needs of end users1
Special-purpose programs

• CHAM has only one well-maintained SPP, namely
  FLAIR,
• However, CVD (for chemical-vapour- deposition
  reactors) and ESTER (for aluminium smelters) have
  had face-lifts.
• The overheads in creating and maintaining an SPP
  are not small. So there will probably be no more.
  
• Nevertheless most potential users of PHOENICS are
  interested in only narrow application sectors.
• We therefore need a new strategy.

8
3 Meeting the needs of end users2
Special-purpose GATEWAYS

• The new strategy the GATEWAY concept, namely
• 1 The PHOENICS package (pre-processors, solver,
  post-processors documentation) is always the
  same.
• 2. For each sector, there is provided a bundle of
  files which, accessed via PRELUDE, provide all
  that the end-user needs but no more.
• 3.The bundle includes
• A store cupboard filled with useful
  objects-with-attributes
• A start-up script
• A few examples
• A new-user tutorial.
• 4. We do not plan to de-activate any parts of
  PHOENICS so a Gateway user still has access to
  all PHOENICS facilities.
• 5. Probably a modest extra charge will be made
  for the files.

9
3 Meeting the needs of end users2 GATEWAYS
(continued)

• GATEWAYS are much easier to construct and
  maintain than SPPs.
• They are best constructed by partnerships between
  CHAM and a specialist company or consultant.
• Examples of application sectors which could be
  served by PHOENICS GATEWAYS include

10
3 Meeting the needs of end users2 GATEWAYS
(continued)

• Shell-and-tube and other heat exchangers
• Steam condensers
• Furnaces and incinerators
• City pollution complete-building studies via
  transfer objects
• Circuit-breakers
• Rocket-exhaust plumes
• Fans, pumps, compressors of various kinds
• Wind farms
• Wave tanks
• Waste-water treatment plant
• and many more.
• Would you like to be a gateway partner? Please
  let me know.
• You will learn more about gateways in the last
  part of this lecture.

11
4. Building on existing strengths1. Multi-phase
flow

• CHAM pioneered the CFD simulation of two-phase
  flow, funded by the nuclear industry in the USA
  and UK.
• The input file library contains many examples.
• Competing CFD codes are weaker than PHOENICS in
  this application area but they will catch up.
• The (buried) MUSES technique was our first step
  towards multi-phase-flow simulation.
• The current In-Formization project has revealed
• 1 (bad) that its PLANT embodiment was not quite
  correctly implemented and
• 2 (good) when it is correctly implemented it
  allows many more new applications, and
  improvements to pre-existing features
  (e.g.MOFOR), than had been previously recognized.
• Exploitation of this new recognition is one of
  the tasks being worked on in Moscow.

12
4. Building on existing strengths2. Simultaneous
flow solid stress

• Another unique buried-treasure strength of
  PHOENICS is its SFT capability.
• Its early implementation had deficiencies, now
  removed.
• However exemplification, and problem set-up
  assistance via VR-Editor or (more probably
  PRELUDE) must be supplied if its value is to be
  recognized by users.
• Do any of you have active projects needing
  simultaneous solid-stress computation? If so,
  CHAM will be glad to assist.
• Joint consultancy projects would be the best way.
• This is another of the tasks being worked on in
  Moscow.

13
4. Building on existing strengths3. PARSOL

• PARSOL has a weakness, which its name betrays
• PARt SOLid reveals that its authors thought of
  cells as having one part fluid and the other
  solid.
• Then, thinking solids had no need for the
  pressure or velocity variables, they constructed
  unsymmetrical coding.
• Since then, we have recognised the need to handle
  cut cells in which both parts are fluid
• Or more generally these which may be divided into
  three, typically (but not necessarily) two fluid
  parts and one solid.

14
4. Building on existing strengths3. PARSOL
(continued)

• The current NewParsol project is therefore
  directed to enabling PHOENICS to handle
  doubly-cut (and therefore 3-part) cells.
• During its course, the possibility of using Local
  Body-Fitted-Coordinate grids was explored, and
  found to be satisfactory for some cases.
• However, human-resource limitations have caused
  it to be put on hold.
• Nevertheless, the diversion allowed time and
  provided stimulus for the invention of a neater
  NewParsol strategy than was first conceived.
• This is now being investigated. Its
  implementation will involve both Moscow and
  Wimbledon.
• A difficulty to be overcome is the
  Henry-King-ism of some the legacy coding.
  However a cure for this disease is being
  administered.
• The new strategy allows for the stresses to be
  computed in the solids, as well as heat fluxes
  both along and normal to the.interfaces.

15
5 Providing new featuresUSP 1 the motive

• The only radically new feature being worked on
  (mainly in Moscow) is UnStructured Phoenics,
  known as USP.
• The motive is NOT (as it may be for competitors)
  to handle curved-surface bodies for PARSOL
  handles these satisfactorily.
• The motive is to cut out the waste of time and
  storage entailed by the un-needed fine-grid
  regions which PARSOL (in a structured grid)
  generates far from the bodies.
• I will also cut out waste when only a small part
  of the domain is physically interesting as in the
  next picture..

16
USP ignores most cells
17
5 Providing new featuresUSP 2 the method

• USP is a part of the standard PHOENICS package,
  which remains able to work in structured or in
  USP mode.
• USP employs a standard-PHOENICS cartesian or
  polar grid as its starting point.
• It proceeds by replacing pairs, quartets or
  octets of cells by single cells, until the
  required economical grid is arrived at.
• It retains PARSOLs sub-cells near curved-surface
  objects.
• USP employs a collocated scheme for the pressure
  and velocities.

18
5 Providing new featuresUSP 2 the method
(continued)

• Most of the new coding is being written in
  Fortran 90.
• New features are tested as they are introduced,
  where possible by comparison of results with
  those of structured PHOENICS.
• Activation of USP requires few actions by the
  user except settings of grid coarsening factors.
• The main work is being performed in Moscow but
  testing has begun in Wimbledon.

19
5 Providing new featuresUSP 3. Current status

• All the main elements of USP (unstructured
  storage, addressing scheme, coefficient and
  residual calculation, conjugate-gradient solver)
  exist and work satisfactorily)
• Boundary-condition and source specifications via
  standard Q1s are accepted.
• Tests are proceeding systematically and
  successfully.

20
5 Providing new featuresUSP 4. Main outstanding
matters

• Acceptance of In-Form input.
• Decisions about grid-file format.
• Decisions about automatic grid generation.
• Decisions about visual display of result.
• The date for beta release (hopefully June 2007).

21
Phoenics developments

• Next topic PRELUDE