Results 1 to 9 of 9
  1. Username Protected
    Star Contributor

    Posts
    745 Posts
    Thanked 507 times
    Phenom 300 Owner & Pilot
    Join Date
    Joined Oct 2020
       #1  

    ForeFlight Now Uses STAR Altitudes in Planning

    Tyson Weihs, ForeFlight co-founder, just posted on CJP that ForeFlight now uses STAR altitudes in fuel calculations. That has been a longtime request of myself and others.

    We shipped an update today that takes flight planning up a notch on fuel accuracy. When planning a STAR, the flight planning engine now adheres to STAR crossing altitude restrictions and adjusts fuel calculations accordingly. The next time you create a flight, the plan will take these into account and show the restrictions and altitude computed at each STAR waypoint on the navlog
    Here is an example on the JAIKE arrival into TEB. You can see the TOD has been moved back to make JAIKE at FL130, and all the other waypoints match as well.

    Name:  IMG_7359C23CE418-1.jpeg
Views: 1458
Size:  962.3 KB
  2. Username Protected
    Really Frequent Poster

    Posts
    208 Posts
    Thanked 64 times
    Phenom 100 Owner & Pilot
    Join Date
    Joined Oct 2020
    #2  
    I find FF already very accurate, so this will only improve flight planning. Thanks for sharing.
  3. Username Protected
    Member

    Posts
    31 Posts
    Thanked 7 times
    Phenom 300 Owner & Pilot
    Join Date
    Joined Oct 2020
    #3  
    I’m still finding ForeFlight fuel burns off by 150-200 pounds compared to flightplan.com…..anyone have some realistic settings for the 300 I can manually enter into ForeFlight?
  4. Username Protected
    Member

    Posts
    31 Posts
    Thanked 63 times
    Other Services Provider
    Join Date
    Joined Jun 2021
    #4  
    Quote Originally Posted by (Username Protected) View Post
    I’m still finding ForeFlight fuel burns off by 150-200 pounds compared to flightplan.com…..anyone have some realistic settings for the 300 I can manually enter into ForeFlight?
    While I'm not with ForeFlight anymore, having investigated many customer feedback on "fuel burn is off, fltplan is spot on", I can provide the following feedback and troubleshooting guidelines:


    1. By far, the biggest source of fuel burn discrepancies is when crews don't fly the published climb and descent speed schedules, and/or planned cruise speed. If you climb at a different speed schedule, it will be off
    2. In general, cruise models are spot-on with real-world; I have rarely seen more than +/-1% deviation in fuel between TOC and TOD provided planned cruise speed is maintained (error then is often wind and/or temp versus predicted)
    3. Climb and descent can differ a bit because both are "ideal" with no level off and TOC/TOD calculated for that -- this is not always a real-world experience, as you know, especially in Class B airspace
    4. OEMs typically guarantee flight planning performance to be within 3% of actual, so the book data is only as good as that


    It was always interesting to hear how "good" fltplan is in pilots' minds, especially since its model does not consider weight reduction with fuel burn and non-standard temperature condition effects. If you really want to revert back to that method, you can setup your own fltplan formatted climb, cruise and descent model in ForeFlight using the by-altitude performance and input via web. But I'd advise against that as my tests have shown it's not that great in real world.

    Instead I would attempt to identify the source of the fuel burn error and correct it with the provided climb, cruise and descent fuel biases available. This requires flying 3-5 flight and meticulously recording actual fuel and (preferably) delta ISA data:
    • Actual weight and fuel remaining at brake release
    • Fuel remaining at TOC, but ideally at TOC and any navlog waypoint crossed during the climb
    • Same with delta ISA (to check if weather model significantly deviated from actual)
    • Fuel remaining and temp/winds at all cruise waypoints
    • Fuel and temp a TOD, waypoints through descent and after runway exit


    It's a decent amount of work, but will help Foreflight investigate things. Ideally all this information is directly noted on a printed navlog. Again, one flight is not enough, 3-5 flights are best. The idea is then to calculate the average delta climb fuel (predicted compared to actual), delta cruise fuel and delta descent fuel over those 5 flights and set teh fuel biases accordingly. Often only climb, and maybe descent will need to be biased. Then repeat the same exercise over another 5 flights to further dial in the correction.
  5. Username Protected
    Really Frequent Poster

    Posts
    208 Posts
    Thanked 64 times
    Phenom 100 Owner & Pilot
    Join Date
    Joined Oct 2020
    #5  
    Thank you (Username Protected), My experience is that Foreflight is spot on. The only time there is discrepancies is when NY/Boston doesn't let you climb or ATC drops you down too soon. No model can do a lot to predict that.
  6. Username Protected
    Frequent Poster

    Posts
    134 Posts
    Thanked 73 times
    Phenom 300 Owner & Pilot
    Join Date
    Joined Nov 2020
    #6  
    Quote Originally Posted by (Username Protected) View Post
    Thank you (Username Protected), My experience is that Foreflight is spot on. The only time there is discrepancies is when NY/Boston doesn't let you climb or ATC drops you down too soon. No model can do a lot to predict that.
    fwiw, I would love to see Foreflight use some historical “real world” climb/descent data (from flightaware, etc) to create something like a 95% probability profile of expected climb/decent for a given route/time of day. I bet controllers are pretty predictable overall (I find they are on routes I fly regularly). Being able to calculate “expected” burn rather than ideal would be awesome.
  7. Username Protected
    Member

    Posts
    7 Posts
    Thanked 8 times
    Phenom 100 Owner & Pilot
    Join Date
    Joined Oct 2020
    #7  
    Sharing some data to add to the discussion. I have been tracking my fuel consumption for a while now. I have taken that fuel burn and compared it to the original ForeFlight estimate. Below is a sample table.

    Name:  Screen Shot 2022-06-16 at 10.27.21 AM.png
Views: 1365
Size:  299.0 KB

    If you eliminate the three outliers that were very short legs (<200nm), specifically the 21-JAN-22 and 20-Feb-22 legs, I burned on average 0.70% more fuel than planned.

    Breaking it down further (all numbers in lbs):
    • Average net fuel = -11
    • Median net fuel = -1 (by this measure, Foreflight is dead-on)
    • Max unplanned fuel burned = 155
    • Max fuel saved = 96
  8. Username Protected
    Member

    Posts
    31 Posts
    Thanked 63 times
    Other Services Provider
    Join Date
    Joined Jun 2021
    #8  
    Quote Originally Posted by (Username Protected) View Post
    fwiw, I would love to see Foreflight use some historical “real world” climb/descent data (from flightaware, etc) to create something like a 95% probability profile of expected climb/decent for a given route/time of day.
    It's in the realm of possibility (and quickly becomes a big data exercise), either using iPad time/distance or Sentry ADS-B device feeding the data. Fuel burn would need to rely on human input, unless FF could tap into the aircraft data bus in the future...an area where I feel Garmin has an advantage, knowing all the aircraft data fo rmany Garmin avionics equipped aircraft. I explored this a bit while I was there, but as it often is, it's a question of limited time bandwidth, other priorities, etc. At the end of the day, sticking with the OEM data for now is the best approach.
  9. Username Protected
    Member

    Posts
    31 Posts
    Thanked 63 times
    Other Services Provider
    Join Date
    Joined Jun 2021
    #9  
    Quote Originally Posted by (Username Protected) View Post
    Sharing some data to add to the discussion. I have been tracking my fuel consumption for a while now.

    I love seeing this! Short hops are by the far the toughest to get right and also introduce the most data scatter. I'd recommend throwing some extra bias fuel into climb or descent as a one-off change just for such short flights (or just load more fuel, knowing you'll need it). Often the OEM low altitude data is a bit lacking, which introduces error.

    For "normal flights", you could try to bias the fuel for further tweaks, but it's dialed in quite nicely as is.

Posting Permissions