From: vince@offshore.ai (Vincent Cate) Newsgroups: sci.space.tech Subject: Re: Basci question about rocket shapes. References: <80d892ce.0305160624.92479a0@posting.google.com> <6eaa11b1.0305241211.4b98df4c@posting.google.com> NNTP-Posting-Host: 209.88.68.230 henry@spsystems.net (Henry Spencer) wrote in message news:... > In article <6eaa11b1.0305241211.4b98df4c@posting.google.com>, > Ian Woollard wrote: > >Clearly the more nearly spherical your structure is the better the > >mass efficiency would be. > > With metals, ignoring intertanks etc., yes. With composites, a cylinder > is just as mass-efficient as a sphere. I had always thought spherical was more efficient but when I wrote some code to look at the mass of pressure tanks I found they were the same mass for the same volume and pressure. It is true that for the same radius a cylinder is twice as thick, but that is not the whole story. I added pressure tank mass calculation to my "rocket equation" applet. Sample inputs for tank mass are at the end. You can run it at: http://www.spacetethers.com/rocketequation.html The tank is assumed to be a half sphere at each end of a cylinder. I seem to get about the same tank mass for the same volume tank and pressure no matter how long the cylinder section is (0 on up). Compare sample runs of "cylinder", "sphere", and "combo". If either a metal or a composite did not work for these formulas, I would have guessed a wound composite since it can have more strength in one direction. Are you sure it is not the other way? The relevant part of my code is below (full code is on my site). Basically I imagine cutting a cylinder in half lengthwise and figure out how much force there is how thick the material needs to be. Same with a sphere. I get that the cylinder needs to be twice as thick as the sphere of the same radius (which I remember from school). This does make approximations that assume the tank walls are thin compared to other dimensions. It is also assuming that the pressure is about the same everywhere inside the tank. For a tall rocket under several Gs acceleration, this is not true. Because of this a sphere will have an advantage. Is this the only advantage a sphere has? For real pressures and heights, how big a deal is this? Do they make rocket tanks thicker at the bottom? If anyone finds any errors in the code below, please let me know. -- Vince private void outputTank(double TankPSI, double TankMegaPascals, double TankGramsPerCC, double TankRadiusMeters, double TankCylinderMeters) throws Exception { double PI = java.lang.Math.PI; double PSItoMetric = 6894.757; double Pressure = TankPSI * PSItoMetric; double TankDiameter = 2 * TankRadiusMeters; double GramsPerCCtoKgPerMeter2 = 1000000/1000; double Density = TankGramsPerCC * GramsPerCCtoKgPerMeter2; double Rsquared = TankRadiusMeters * TankRadiusMeters; double Rcubed = Rsquared * TankRadiusMeters; double Pascals = TankMegaPascals * 1000000; double SplitCylinderArea = TankCylinderMeters * TankDiameter; double SplitCylinderForce = Pressure * SplitCylinderArea; double CylinderSplitLine = 2 * TankCylinderMeters; double CylinderThickness = SplitCylinderForce / (CylinderSplitLine * Pascals); double CylinderArea = PI * TankDiameter * TankCylinderMeters; double CylinderMass = CylinderThickness * CylinderArea * Density; double CylinderVolume = PI * Rsquared * TankCylinderMeters; if (TankCylinderMeters <= 0) { CylinderThickness=0; CylinderMass=0; CylinderVolume=0; } double SplitSphereArea = PI * Rsquared; double SplitSphereForce = Pressure * SplitSphereArea; double SphereSplitLine = PI * TankDiameter; double SphereThickness = SplitSphereForce / (SphereSplitLine * Pascals); double SphereArea = 4 * PI * Rsquared; // both halves double SphereMass = SphereThickness * SphereArea * Density; double TotalMass = CylinderMass + SphereMass; double SphereVolume = 4/3 * PI * Rcubed; double TankVolume = SphereVolume + CylinderVolume; double TankGallons = TankVolume * 264.172048; MyPrint("CylinderThickness in cm ", CylinderThickness*100); MyPrint("SphereThickness in cm ", SphereThickness*100); MyPrint("CylinderMass in Kg ", CylinderMass); MyPrint("SphereMass in Kg ", SphereMass); MyPrint("Total mass in Kg ", TotalMass); MyPrint("Mass in pounds ", TotalMass * 2.2046); MyPrint("TankVolume in meters cubed ", TankVolume); MyPrint("TankVolume in US liquid gallons ", TankGallons); }