When should I segment my pipeline flow calculations?

It has always been an article of faith that if the density of the downstream piping were within 10% of the density of the upstream piping then the incompressible flow equations would provide adequate results in an inherently compressible fluid. When the downstream density is less than 90% of the upstream density then common practice was to segment the line until that 90% was realized. To get a feel for how valid this assumption is I wrote a MathCad program to bust up a flow problem with downstream density around 50% of upstream density for a single segment and then matched the (pipeline inlet - pipeline outlet) dP while segmenting the line finer and finer.

Assumptions:
[ul]
[li]100% CH4 (MW 16.043, SG 0.5539)[/li]
[li]Std weight 24 inch new steel pipe (ID 23.25 in, efficiency 0.95, absolute roughness 150E-06 ft)[/li]
[li]No water standing in line[/li]
[li]Temp constant at 520R[/li]
[li]Upstream pressure 580 psia [40 bara][/li]
[li]Downstream pressure 290 psia [20 bara][/li]
[li]Line length 62 mi [100 km][/li]
[/ul]

Steps:
[ol 1]
[li]Guess segment downstream pressure[/li]
[li]Calculate average pressure with the guess (using front-end loaded average)[/li]
[li]Calculate compressibility at average and standard pressure[/li]
[li]Calculate density at average pressure and standard pressure[/li]
[li]Calculate viscosity at average pressure[/li]
[li]Calculate Reynolds Number[/li]
[li]Calculate Fanning Friction Factor[/li]
[li]Using Isothermal Gas Flow Equation to calculate downstream pressure[/li]
[li]If calculated downstream pressure more than 100 Pa from guess, iterate 1-9.[/li]
[li]Move segment downstream pressure to upstream pressure and repeat 1-10[/li]
[/ol]

I ran the program for a number of different segments:

[tt]Seg Length...........Max Flow rate..........Density Ratio (worst segment)
100 km.................243.725 MMSCF/day...............48.1%
50 km..................243.759 MMSCF/day...............61.5%
25 km..................243.759 MMSCF/day...............74.3%
10 km..................243.759 MMSCF/day...............87.0%
2 km...................243.765 MMSCF/day...............96.9%
1 km...................243.765 MMSCF/day...............98.4%
500 m..................243.765 MMSCF/day...............99.2%
250 m..................243.765 MMSCF/day...............99.6%
100 m..................243.765 MMSCF/day...............99.8%
10 m...................243.765 MMSCF/day...............99.9+%[/tt]

I found it interesting that above 90% the flow rate was unchanging. I re-ran this several times for flow rates increasing and found that below 90%, the flow rate changed with increasing segmentation, and above 90% a given dP gave the same flow rate. Changes above 30% were pretty small (more than the 0.016% above, but much smaller than the +/-10% accuracy we expect to get from this equation). My conclusion is that I will continue to use the 90% check, but certainly understand why someone would ignore it for pipeline calculations.