September 22, 2002
CAPACITY & DELAY ANALYSIS
Continuing Concerns
Here’s a review of the most recent rendition (19 Sep) of the airfield capacity and delay analysis being conducted in support of the SDF Part 150 update.
The report continues to read as if the entire GA fleet is being marshalled in and out of the FBO ramp (n.b. middle of page 7). This presumably wouldn’t affect the delay calculations themselves, since normal taxi time occurs without regard to traffic in the system. But to the extent it plays a related role in runway preference, it’s worth reiterating that a large portion of the Louisville GA fleet – probably almost all that operating at night – consists of cargo carriers subcontracted to UPS.
These are the "time-critical" operations described later in the report, and with their base at the UPS "Lear Ramp" adjacent to runway 17R/35L, their differential taxi times would be just the opposite of those described in the report. What effect these operations, represented accurately, might have on the rest of the airfield simulation and resulting findings, both for this analysis and the Master Plan, is difficult to guess.
The application of runway preference with increasing traffic loads in the two cases compared here (the "MPU" [Master Plan Update] and the present SIMMOD analysis) remains troublesome. Runway preference should affect two runway-use cases: south-flow arrivals, and north-flow departures. The current and future cases in the present work are no longer identical, but the differences are still very small.
For south-flow arrivals, as traffic increased from 2000 to 2010 the Master Plan relaxed the use of the preferred runway by 14% during the day, and 12% at night. The present work, by contrast, has no change in daytime use, and only a 1% relaxation at night. While the preference itself is mirrored, the decision rule for applying it should be the same in both cases.
The north-flow departure cases are more difficult to compare, because still-unresolved anomalies from the original Master Plan work get in the way. There, the daytime departure preference shifted a whopping 50% from 2000 to 2010 – from a strong preference for 35R to a preference nearly as strong for 35L. At night, as the current narrative reminds us, the 2000 MPU runway use for north departures is almost equal, but in 2010 shifts 15% in favor of 35L. But the scenario for all the MPU cases should be a preference for 35R.
This laydown complicates the present work in other ways as well. At the end of page 4, the report says "The north-flow departure delays are higher than the south-flow departure delays due to the runway usage being more balanced in south-flow operations." But in the very next sentence it cites the balanced MPU 2000 north-flow case just noted. The corresponding south-flow case is balanced too, but the MPU 2000 departure delays are considerably different – in fact, in nighttime IFR the north-flow figure (6.20) is nearly double the south (3.18).
The succeeding narrative on page 5 goes on to speculate that other factors – "change in the flighttracks or departure headings" – may be responsible for different delay results. And these factors may indeed influence differences between north and south flow. But it’s not easy to see how they’d account for differences between matched cases from the MPU and the present SIMMOD work on the proposed alternative. The only difference listed in airspace assumptions for departures between these two cases is the one-mile centerline leg for departures on 35L before diverging as before. This hardly seems the stuff significant delays are made of.
The text goes on to observe that nighttime north-flow departures in IFR conditions, the one case projected to exceed tolerances by 2010, occurs less than one percent of the time – 21 hours per year is the figure used later. This makes it all the more curious that no account is given, here or in the Master Plan, of the capacity of Runway 11/29, whose use should be required more often than that. And since it’s only a single runway, delays incurred in that case should be considerably larger. If the latter isn’t cause for worry, then the former shouldn’t be either.
The rarity of north-flow departures in IFR is underscored in Table 5, whose overall average figures come out well within stated levels of acceptability. Linking these points more closely, and in sequence – perhaps by placing the paragraph that ends on page 5, with the 1% observation, ahead of the paragraph on page 4, that introduces the table – would make the table much more effective.
The cost figures given are presumably direct calculations from the delay statistics, operations count, and rates given. It only bears observing that the narrative at the end of page 7 concludes, "All costs are stated in year 20000 dollars, without any adjustment for inflation of Consumer Price Index." It’s not clear how else to account for rates that go from $1878 per hour in 2000 to $2007 per hour in 2010.
The report’s conclusions return once more to the notion of "limitations on flighttrack headings" to explain the differences in delay between the MPU and the present SIMMOD analysis of the proposed alternative. It says, "Requiring all jet departures to use the same heading lengthens the amount of time between consecutive departures…" This is true of every case but 17R, which has both straight-out and divergent tracks. But it’s no change from current operations modeled in the MPU (though the specific headings cited both there and in this report are still in error).
The one corrective measure proposed in these conclusions, at the bottom of page 7, however, is "to more evenly balance runway usage during peak periods" – dismissed throughout the earlier narrative as not-the-culprit, despite all the anomalies.
With no more credible basis to account for the results given, and with all the incongruous inputs noted both here and in the Master Plan itself, it’s not prudent to lean too hard on the present work for the many cost/benefit tasks it must support. We need to get a much clearer picture of the real factors in play, the accuracy and sensibility of the results, and what we can or should do about them.