Re: [Flightgear-devel] Pressure errors in FGEnvironment
Christian Mayer writes: As I wrote before, there's a function in the WeatherCM code that calculates the air pressure based on the air pressure at a given altitude and at a given teperature profile. It is based on the well known (but incorrect) baryometric (SP?) formula but doesn't suffer from its limitations. When you feed it, the standard conditions it will return the standard atmosphere. When you adopt that code, you'll automagicly get the correct results. Does the code handle only pressure? There are a few fairly good atmosphere models I can adapt (including the one in JSBSim); I just stuck with the tables for now because they keep the code fast and simple. I want to be able to extrapolate both ways -- if the user supplies a temperature or pressure at altitude, I want to be able to extrapolate the temperature or pressure at sea level, and vice-versa. All the best, David -- David Megginson, [EMAIL PROTECTED], http://www.megginson.com/ ___ Flightgear-devel mailing list [EMAIL PROTECTED] http://mail.flightgear.org/mailman/listinfo/flightgear-devel
re: [Flightgear-devel] Pressure errors in FGEnvironment
Andy Ross writes: Nonetheless, I think I found the problem. In converting the YASim table to the new format, its values were re-encoded as deltas from sea level conditions, with sea level pressure defined as 29.92 inches of mercury. I've fixed that now -- everything is encoded as factors instead. It seems OK at lower altitudes, but please let me know how it affects the 747. All the best, David -- David Megginson, [EMAIL PROTECTED], http://www.megginson.com/ ___ Flightgear-devel mailing list [EMAIL PROTECTED] http://mail.flightgear.org/mailman/listinfo/flightgear-devel
Re: [Flightgear-devel] Pressure errors in FGEnvironment
David Megginson wrote: Christian Mayer writes: As I wrote before, there's a function in the WeatherCM code that calculates the air pressure based on the air pressure at a given altitude and at a given teperature profile. It is based on the well known (but incorrect) baryometric (SP?) formula but doesn't suffer from its limitations. When you feed it, the standard conditions it will return the standard atmosphere. When you adopt that code, you'll automagicly get the correct results. Does the code handle only pressure? The code does only calculate air pressure. There are a few fairly good atmosphere models I can adapt (including the one in JSBSim); The code does comply with the international atmosphere models (IIRC JSBsim uses exactly the same data I used to verify the calculations) I just stuck with the tables for now because they keep the code fast and simple. Tables are faster and simpler. But they aren't really flexible. When you are concerned about performance: How many pressure calculations do we need? Not more than a few per frame. And as the pressure changes are very small during a frame we can even cache the result. And alltogether the number of calculations that are done is very small. Probably the space overhead a table generates would be worse (cf. discusion about inlining code) I want to be able to extrapolate both ways -- if the user supplies a temperature or pressure at altitude, I want to be able to extrapolate the temperature or pressure at sea level, and vice-versa. So the code is better suited for you than the tables. You give the code a temperature profile (which is basicly the table approach) and a pressure value at a give altitude (doesn't need to be sea level). Then you get valid numbers for any altitude that you want. CU, Christian PS: As the air pressure curve is similar to the e-function (e^altitude) it's nowhere linear and thus badly approximated by a table... -- The idea is to die young as late as possible.-- Ashley Montague ___ Flightgear-devel mailing list [EMAIL PROTECTED] http://mail.flightgear.org/mailman/listinfo/flightgear-devel
Re: [Flightgear-devel] Pressure errors in FGEnvironment
--- Christian Mayer [EMAIL PROTECTED] wrote: David Megginson wrote: Christian Mayer writes: As I wrote before, there's a function in the WeatherCM code that calculates the air pressure based on the air pressure at a given altitude and at a given teperature profile. It is based on the well known (but incorrect) baryometric (SP?) formula but doesn't suffer from its limitations. When you feed it, the standard conditions it will return the standard atmosphere. When you adopt that code, you'll automagicly get the correct results. Does the code handle only pressure? The code does only calculate air pressure. There are a few fairly good atmosphere models I can adapt (including the one in JSBSim); The code does comply with the international atmosphere models (IIRC JSBsim uses exactly the same data I used to verify the calculations) I just stuck with the tables for now because they keep the code fast and simple. Tables are faster and simpler. But they aren't really flexible. When you are concerned about performance: How many pressure calculations do we need? Not more than a few per frame. And as the pressure changes are very small during a frame we can even cache the result. And alltogether the number of calculations that are done is very small. Probably the space overhead a table generates would be worse (cf. discusion about inlining code) I want to be able to extrapolate both ways -- if the user supplies a temperature or pressure at altitude, I want to be able to extrapolate the temperature or pressure at sea level, and vice-versa. So the code is better suited for you than the tables. You give the code a temperature profile (which is basicly the table approach) and a pressure value at a give altitude (doesn't need to be sea level). Then you get valid numbers for any altitude that you want. CU, Christian PS: As the air pressure curve is similar to the e-function (e^altitude) it's nowhere linear and thus badly approximated by a table... Depends on how many points are in the table. -- The idea is to die young as late as possible. -- Ashley Montague ___ Flightgear-devel mailing list [EMAIL PROTECTED] http://mail.flightgear.org/mailman/listinfo/flightgear-devel __ Do You Yahoo!? Yahoo! - Official partner of 2002 FIFA World Cup http://fifaworldcup.yahoo.com ___ Flightgear-devel mailing list [EMAIL PROTECTED] http://mail.flightgear.org/mailman/listinfo/flightgear-devel
Re: [Flightgear-devel] Pressure errors in FGEnvironment
Tony Peden wrote: PS: As the air pressure curve is similar to the e-function (e^altitude) it's nowhere linear and thus badly approximated by a table... Depends on how many points are in the table. Yes. You can solve all problems with raw iron... I don't know how feelable the sudden performance drops are when you arrive in a new part of the table... CU, Christian -- The idea is to die young as late as possible.-- Ashley Montague ___ Flightgear-devel mailing list [EMAIL PROTECTED] http://mail.flightgear.org/mailman/listinfo/flightgear-devel
Re: [Flightgear-devel] Pressure errors in FGEnvironment
On Wed, 2002-06-05 at 14:15, Christian Mayer wrote: Tony Peden wrote: PS: As the air pressure curve is similar to the e-function (e^altitude) it's nowhere linear and thus badly approximated by a table... Depends on how many points are in the table. Yes. You can solve all problems with raw iron... I don't know how feelable the sudden performance drops are when you arrive in a new part of the table... There's two ways to implement the model with tables. One is the brute force approach that LaRCsim uses. It maintains a tables of T,p, and rho as a function of altitude and interpolates as needed. The second comes up when you implement the model as specified by ICAO, that uses a table of constants that get plugged into one of two equations (depending on whether or not the region is isothermal). JSBSim uses the latter. The first will require a lookup and interpolation every frame its executed while the second may not require a lookup (the table only has eight breakpoints even if it does) and will incur the cost of evaluating the equations (they are e^x, I've no idea how much FPU/CPU time they incur). I don't really know, but I'd guess it's a toss up. CU, Christian -- The idea is to die young as late as possible.-- Ashley Montague ___ Flightgear-devel mailing list [EMAIL PROTECTED] http://mail.flightgear.org/mailman/listinfo/flightgear-devel -- Tony Peden [EMAIL PROTECTED] We all know Linux is great ... it does infinite loops in 5 seconds. -- attributed to Linus Torvalds ___ Flightgear-devel mailing list [EMAIL PROTECTED] http://mail.flightgear.org/mailman/listinfo/flightgear-devel
re: [Flightgear-devel] Pressure errors in FGEnvironment
Andy Ross writes: Nonetheless, I think I found the problem. In converting the YASim table to the new format, its values were re-encoded as deltas from sea level conditions, with sea level pressure defined as 29.92 inches of mercury. But in the FGEnvironment constructor, I see the following initializer: pressure_sea_level_inhg(28), Now, 28 in. Hg is a reasonable sea level pressure. But it's not the same one that the table is assuming. At sea level, it's off by only about 2 inches. But at altitude, it's off by a whopping TWO inches! :) Actually, the value being used is the one set in preferences.xml, 28.5, but you're still almost certainly right. I'll change both to 29.92, but there's more to consider. The idea is that users should be able to set any reasonable sea-level pressure and see reasonable behaviour -- that's why I set the tables up with deltas rather than absolute values. I can see, now, how that would be a problem at higher altitudes, but what should we see? If the altimeter setting at ground level is 28inHG or 31inHG, what would you expect at, say, 20,000 ft? Would a factor rather than an offset most appropriate? All the best, David -- David Megginson, [EMAIL PROTECTED], http://www.megginson.com/ ___ Flightgear-devel mailing list [EMAIL PROTECTED] http://mail.flightgear.org/mailman/listinfo/flightgear-devel
Re: [Flightgear-devel] Pressure errors in FGEnvironment
David Megginson wrote: The idea is that users should be able to set any reasonable sea-level pressure and see reasonable behaviour -- that's why I set the tables up with deltas rather than absolute values. I can see, now, how that would be a problem at higher altitudes, but what should we see? If the altimeter setting at ground level is 28inHG or 31inHG, what would you expect at, say, 20,000 ft? Would a factor rather than an offset most appropriate? Obviously, getting this truly correct is a meteorology problem that requires bunches of scientists and a supercomputer or two to solve. But I'd argue that using a factor would be saner from a flight simulation perspective -- if the sea level pressure at a location is 95% of nominal, then the pilot would naively expect that the air density at all altitudes would be 95% of nominal. Certainly the use of an offset mechanism is going to be surprising, and for the extremes of sea level pressure will lead to super-hurricanes up at altitude. Actually, I'm fairly certain that high altitude phenomena tend to smooth out pressure differences down below, so in fact the relative difference between pressures at the flight levels should actually be less than that at sea level. Maybe you could try a factor that asymptotically approaches 1.0 as altitude increases? I don't have much background in meteorology, though. Andy -- Andrew J. RossNextBus Information Systems Senior Software Engineer Emeryville, CA [EMAIL PROTECTED] http://www.nextbus.com Men go crazy in conflagrations. They only get better one by one. - Sting (misquoted) ___ Flightgear-devel mailing list [EMAIL PROTECTED] http://mail.flightgear.org/mailman/listinfo/flightgear-devel
Re: [Flightgear-devel] Pressure errors in FGEnvironment
Andy Ross wrote: David Megginson wrote: The idea is that users should be able to set any reasonable sea-level pressure and see reasonable behaviour -- that's why I set the tables up with deltas rather than absolute values. I can see, now, how that would be a problem at higher altitudes, but what should we see? If the altimeter setting at ground level is 28inHG or 31inHG, what would you expect at, say, 20,000 ft? Would a factor rather than an offset most appropriate? Obviously, getting this truly correct is a meteorology problem that requires bunches of scientists and a supercomputer or two to solve. But I'd argue that using a factor would be saner from a flight simulation perspective -- if the sea level pressure at a location is 95% of nominal, then the pilot would naively expect that the air density at all altitudes would be 95% of nominal. Certainly the use of an offset mechanism is going to be surprising, and for the extremes of sea level pressure will lead to super-hurricanes up at altitude. Actually, I'm fairly certain that high altitude phenomena tend to smooth out pressure differences down below, so in fact the relative difference between pressures at the flight levels should actually be less than that at sea level. Maybe you could try a factor that asymptotically approaches 1.0 as altitude increases? I don't have much background in meteorology, though. As I wrote before, there's a function in the WeatherCM code that calculates the air pressure based on the air pressure at a given altitude and at a given teperature profile. It is based on the well known (but incorrect) baryometric (SP?) formula but doesn't suffer from its limitations. When you feed it, the standard conditions it will return the standard atmosphere. When you adopt that code, you'll automagicly get the correct results. CU, Christian PS: Based on the formula the phenomena won't smooth out. -- The idea is to die young as late as possible.-- Ashley Montague ___ Flightgear-devel mailing list [EMAIL PROTECTED] http://mail.flightgear.org/mailman/listinfo/flightgear-devel
[Flightgear-devel] Pressure errors in FGEnvironment
I wrote: First, the air pressures returned from the environment system don't agree with the standard atmosphere that YASim uses to do its calibration Heh, funny that. The new environment manager *is* using YASim's numbers. :) Nonetheless, I think I found the problem. In converting the YASim table to the new format, its values were re-encoded as deltas from sea level conditions, with sea level pressure defined as 29.92 inches of mercury. But in the FGEnvironment constructor, I see the following initializer: pressure_sea_level_inhg(28), Now, 28 in. Hg is a reasonable sea level pressure. But it's not the same one that the table is assuming. At sea level, it's off by only about 2 inches. But at altitude, it's off by a whopping TWO inches! :) This value can be set from other places, so I'm not 100% sure this is the problem. I can't test fgfs right now because my fiancee is laying out wedding invitations on the good computer. But it looks pretty suspicious. I'd suggest: -pressure_sea_level_inhg(28), +pressure_sea_level_inhg(29.92), Andy -- Andrew J. RossNextBus Information Systems Senior Software Engineer Emeryville, CA [EMAIL PROTECTED] http://www.nextbus.com Men go crazy in conflagrations. They only get better one by one. - Sting (misquoted) ___ Flightgear-devel mailing list [EMAIL PROTECTED] http://mail.flightgear.org/mailman/listinfo/flightgear-devel
