adm78 / visualchemeng-js Goto Github PK

This could go below the current css grid structure. Should feature equilibrium relation, describe the Rachford rice solution method etc.

Currently the fact that we have physical feeds (say, the real feed to a distillation column with component concentrations and a temperature) and graphical feeds (i.e. somewhere where we generate particle entities which become part of an ensemble to give a graphical effect). It would be better it we called them something different entirely. Say, change the graphical version of the feed to ParticleSource , GraphicalFeed or GraphicalParticleSource.

This also highlights a potential need to split up the graphical and physical components with the vce library. However, some apps (namely the molecular dynamics app) blur the line between these two things. Needs a bit more though.

Small windows sizes lead to small canvas sizes. Only the first column of particles is initialized and the run button does not start the simulation. Subsequent restarts add more particle onto of the existing ones.

Console error: md.js L57: TypeError: particles[i] is undefined

Options for documentation format:

• sphinx-js https://hacks.mozilla.org/2017/07/introducing-sphinx-js-a-better-way-to-document-large-javascript-projects/
• jsDoc http://usejsdoc.org/about-getting-started.html

Zooming functionality is enabled by default but is not useful in this case. Particularly annoying since the auto-scale/reset axis button have been omitted for viewing purposes.

vcelogo.pdf

Example:

On small displays, the ffect is not too bad, but is an issue on larger ones. We have too options here:
1.) Simply restart the simulation (in most cases we can just call setup).
2.) Regenerate the canvas, translate particles and update display the size any imported images. This has the benefit of maintaining the current state of the simulations (but is obviously a bit more work).

Canvas sizing is not general enough.

Flash app is slow, might be made faster with one of the optimised matter.js forks like https://github.com/bchevalier/matter-js

Sliders and interval flash vars should be reset, flow should be restarted rather than stopped and pause.

p5 canvas does not scale with the window. Event listener could be take directly from kmc.js. If so, it should be put in a vce_utils.js and loaded by both.

A decision has to be made about what to do with particle that no longer lie in the sim box. Simple solution is just to delete them from the particle array and rebuilt the event queue. Alternatively, they could be translated into a new position in the box.

EDMD needs to also use the css grid layout used by pages/kmc_grid.html. This will mean that some features (primarily the sliders) of css./app_layout.css will need to be moved to a seperate file. The app_layout can be more general.

Slider values are not currently connected to the rate constants within modules/kmc/kmc.js.

This occurs in the first 1-2 seconds of streaming. Duplicate data is shifted. This may also be causing the number of plot tags to reduce from all 9 to only 3 (N[A-C] kmc traces) for t>0. This is either an Plotly issue or something in modules/kmc/kmc.js

This should also show on the plotly composition plot axes.

This should serve as a simple introduction to vapour liquid equilibrium.

Conversion and duty are undefined when the initial concentration of at least one of the reactants is 0.

The conversion issue is likely due to the fact that the limiting reactant could not be identified.

This is an issue with the module.reactor.simple.analytical_reactor.AnalyticalReactor.step method.

Application does not critically fail but Ca(t) = Cb(t) = Ca0 = Cb0 for all t.

Causes overspill of container on window size reduction and mobile view

Information mode should explain the physics behind the system being simulated, plus how it is solved from a numerical perspective.

It's not quite clear what's the best way to delivery this information, so we should probably create a testing branch for it.

System selected through drop down menu.

Examples might include methanol water etc. May need some sort of simple database for the equilibrium constants etc.

Some particle initialisation states and user particle inception events lead to particle overlap. The simulation subsequently becomes unstable, manifesting in particles leaving/entering the simulation box.

This is likely the result of the creation of events with event.t < t, causing time to flow in the wrong direction...

This can be quite easily fixed by checking for overlaps at the point of initialisation/particle inceptions. Offending particles could then be randomly moved until a position without overlap is found, before the simulation is resumed (unpaused).

This could potentially be done using a slider with a range and multiple handles. Example: http://jsfiddle.net/LLfWd/60/

Need to expand and contract (fullscreen and generic resize).

xmax, ymax, key image positions will have to be updated. Particle no longer in the boundary will have to be delted.

Optimised fork is here but it a bit behind the current parent repo and looks stale. Depending on the performance gain, it may or may not be worth switching.

For some window sizes, the time text cannot fit in the time box, as a result, the time is not displayed.

Currently not dragging as expected. See valve_test app.