Thursday, October 29, 2015

Refinement to Fundamental Limits of Orbits

My recent travels for EVE Vegas 2015 reminded me of the benefits to my concentration of doing math in the airport while waiting for a flight.  In the midst of sounds from airport slot machines and tourists from the mid-west comparing who hates Obama the most, I had a moment of clarity on the subject of asymptotic orbit performance.  This avenue of inquiry was covered in Part II, but I am revisiting it with the loose orbit case now that my new blog post on manual orbit is nearing completion.  So this post is just a minor refinement to the manual orbit section I presented in Part II.   

In EVE's motion model, ships are limited in their maximum velocity.  They are also limited in the maximum acceleration they can achieve.  I made a point of emphasizing this in my notes in Part II.  When it came to bounding manual orbit performance, however, I bounded the maximum angular velocity on orbits that were limited by acceleration only, and I didn't consider that there is a point at which the velocity of the ship cannot increase past VMAX.  

Just for review, recall that a stable orbit is characterized by a balance between centripetal force and the acceleration force that a ship can achieve. 
      $\large F_{centripetal} = F_{ship}$

I know that the maximum acceleration a ship can produce is VMAX/τ, so I can solve for the relationship between velocity and orbit distance,
      $\large F_{centripetal} = \frac{M v^2}{R}$
      $\large F_{ship,MAX} = M\frac{V_{MAX}}{\tau}$ 


Isolating angular velocity by writing the ratio v/R
     $\large \omega_{accel,limit} = \frac{v}{R} = \sqrt{\frac{V_{MAX}}{\tau R}}$                               (10.26-1)


This is what I did in the chapter, however, notice that if I write just the velocity from (10.26-1), beyond a orbit radius of τVMAX, the velocity of the ship would need to be greater than VMAX!  This is obviously not possible, unless there are some external forces acting on the ship.  

So, if we aren't limited by our ship's maximum acceleration when orbiting beyond the characteristic distance, then the obvious limit in this region is the ship's maximum velocity.  That is, 
     $\large \omega_{V_{MAX},limit} \leq \frac{V_{MAX}}{R} $                          (10.26-2)

This limit is obvious, but it is worth correcting since my original writing did not acknowledge this boundary.  The updated bound for angular velocity is now,
      $\large \omega_{maximum} \leq \begin{cases} \sqrt{\frac{V_{MAX}}{\tau R}}, & \text{ if } R \leq \tau V_{MAX} \\ \frac{V_{MAX}}{R}, & \text{ if } R > \tau V_{MAX} \end{cases}$        (10.26-3)

Another minor note I want to point out is the boundary between these two conditions.  An orbit at the characteristic distance, τVMAX, has a limit on angular velocity of 1/τ.  This point is an easy one to remember, so I'm going to use it to distinguish between 'tight' orbits and 'loose' orbits.  An orbit with radius closer than the characteristic distance of your ship is limited by your acceleration towards the center of the orbit, which means that manual orbiting techniques are likely to be of value.  

Below I present a normalized graph showing how the new bound for maximum velocity affects the limit on angular velocity proposed for any ship for parameters VMAX and τ

Figure 10.26-1:  I have graphed the limits to angular velocity normalized to ship motion parameters, and now based on two limitations of ship motion - maximum acceleration and maximum forward velocity.  This graph is normalized to distance units of τVMAX in orbit distance, and (1/τ) in angular velocity, so that at an orbit distance of τVMAX the rate is limited to (1/τ)


Now that I have cleared up the limitations to orbit performance with Equation (10.26-3), I can go back and plot the existing ship data against this limit.  The original figure shows poor agreement between measured orbit angular velocity when the orbit radius is greater than τVMAX.  In the figure below, I have updated the theoretical limit traces to include the velocity limitation.  You can see that in the AB and No Prop. Mod. cases, the angular velocity approaches the theoretical.  This is also true in the case when MWD propulsion mods are used because the angle to the acceleration point is almost the same as accelerating towards the center of the orbit circle.  

Figure 10.26-2: Enyo automatic orbit data for No Prop. Mod., AB, and MWD are compared with theoretical limits. Note that these points correspond to orbit commands for 500m and at intervals up to 7500m.


These tighter bounds on orbit performance will be helpful background when I present one possible approach to manual orbits soon (TM). 


Interlude

I will be including a musical interlude with each blog post as an instructional aid.  I hope that you will find that it brings much clarity to your thinking.  Use caution when listening at elevated volumes or in populated areas.



Sunday, October 4, 2015

List of Future Topics

[update Oct 2015. ] 
Because there are many possible topics that I can work on, and little time to complete posts about them, I am breaking with my previous tradition of only presenting results if I have also completed taking data on the phenomena.  I will instead be splitting the theory and data into two separate posts.  Expect delays between theory posts and completed data/correlation posts.  

     Santorine




  The project continues...

I started working on math and physics topics in EVE in 2009.  This effort culminated in the first online presentation of the work in 2010 and wide readership in the EVE community and beyond.  Since that time I've moved from google knol, to google docs, and now to a blog format.  

The technical work is also progressing.  While I do have time to work on theory, it takes considerably longer to write up the notes and longer still to add experimental results.  There are many more topics to explore...


"This is just the beginning!" 

 I will be maintaining a list of topics that are in need of further work.  This list has existed for a long time, and progress is slow because I am busy with life.  If you are interested in adding a topic for analysis, or you wish to contribute insights from your own work, please let me know.  
  • Develop analytical approach to manual orbit techniques and compare performance with built-in command methods. 
  • Write introductory/forward material.  This section was removed in 2012, so an update on this is long overdue. I may also use this opportunity to comment on the significance of the motion models in context of gaming mechanics in general.  
  • Code:  Provide @tamber with motion code for his visualization tool.  [Added August 2016]
  • ExperimentalWrite up my existing notes on warp alignment time, and then compare with data.  This is straightforward, but not as interesting as some of the other topics because it isn't about ship interaction. If I were to include bump interaction in the time-to-warp analysis, it would be more interesting, but I don't have an elegant experiment for this... yet.  [Added Sept 2016]
  • Code: Update all tables with data from an API/dB scan.  I'm not a "coder" so I don't have rapid facility with writing database tools.  Most of the data in this writing, however, dates back to the age of Dominion, so when I get time I will learn some python and extract all the data again for my plot script. 
  • Advanced topics in bumping larger ships.  This requires a bit more work on the physics side before it is ready for a write-up.  [Theory posted Nov 20/2015. Data posted Jan 1/2016.]
  • Effect of modules on maximum angular velocity summary table is complete.  Effect for bump distance, however, still needs to be completed.  This would be in Section II/III blog posts.
  • Writing: Improve the clarity of the keep-at-range  and approach counters to orbiting targets.  
  • Complete the addition of keep-at-range and approach counters for ships approaching along a constant vector.  This is similar to the existing method in Part III.  
  • Experimental work: In Part II, I cover automatic orbits both in closed form and experimentally.  There is clearly a 'fudge' factor that is added to the orbit distance results.  I showed that this was constant for a given hull, without regard for the fitted modules.  Is this dependent on hull even if motion parameters are identical?  Find the common theme in this R0 factor.
  • Experimental work: Test server evaluation of "Drebuchet" collision energy and section write-up.  Simple experiment should yield a previously unknown energy constant.[Added July 2016 - RnK posted some data in their tactical video.]
  • I've completed part of the analysis of so-called 'sling-shot' tackling maneuvers.  The write-up, however, requires a lot of background so this is a long term addition to the blog.
  • Experimental work: The Time-to-warp calculation is trivial based on the writing so far assuming there are no external forces influencing ship motion.  Design an experiment that allows me to measure the effect of bumping on change in align time.  
  • Target prioritization theory!  Formal methods are in no way limited to the physics of ship motion.  This subject has been touched on by many theorists before but I have some added formalism to contribute to the discussion. [1st installment - Mar. 2018]
  • Based on the analysis that I did on 'Keep At Range' and 'Approach' as methods to control angular velocity, I will compare with other popular methods for tracking targets.  The theory is mostly written up already, I just need to make a spreadsheet of interactions. [Topic added Nov. 2015]
  • Correct an error in the 'Mass Matters' section in Part I, thanks to a comment from a friendly reader. [Topic added Dec. 2015, Completed Dec. 21/2015]
  •  Complete Part II section on closed-form orbits and show how orbit distance is being affected.  Summarize with object size summary. [Added Jan. 2016].  
 Please consider commenting on this article if there are other topics that you think are interesting and relevant. 


Best Regards,

S. Santorine