Trajectory Calculation: Lab 2 Lecture Notes

Trajectory Calculation Lab 2 Lecture Notes Nomenclature t time h altitude V velocity, positive upwards F total force, positive upwards D aerodynamic drag (˙) time derivative ( = d( )/dt ) ρ g m CD A i air density gravitational acceleration mass drag coefficient drag reference area time index Trajectory equations The vertical trajectory of a rocket is described by the altitude and velocity, h(t), V (t), which are functions of time. These are called state variables of the rocket. Figure 1 shows plots of these functions for a typical ballistic trajectory. In this case, the initial values for the two state variables h0 and V0 are prescribed. h h0 t V V V0 h t Figure 1: Time traces of altitude and velocity for a ballistic rocket trajectory. The trajectories are governed by Ordinary Differential Equations (ODEs) which give the time rate of change of each state variable. These are obtained from the definition of velocity, and from Newton’s 2nd Law. ḣ = V V̇ = F/m (1) (2) Here, F is the total force on the rocket. For the ballistic case with no thrust, F consists only of the gravity force and the aerodynamic drag force. F = � −mg − D , if V > 0 −mg + D , if V < 0 (3) The two cases in (3) are required because F is defined positive up, so the drag D can subtract or add to F depending in the sign of V . In contrast, the gravity force contribution −mg is always negative. 1 A convenient way to express the drag is 1 2 ρV CD A 2 D = (4) The reference area A used to define the drag coefficient CD is arbitrary, but a good choice is the rocket’s frontal area. Although CD in general depends on the Reynolds number, it can be often assumed to be constant throughout the ballistic flight. Typical values of CD vary from 0.1 for a well streamlined body, to 1.0 or more for an unstreamlined or bluff body. With the above total force and drag expressions, the governing ODEs are written as follows. ḣ = V V̇ (5) = −g − 1 CD A ρV |V | 2 m (6) By replacing V 2 with V |V |, the drag contribution now has the correct sign for both the V > 0 and V < 0 cases. Numerical Integration In the presence of drag, or CD > 0, the equation system (5),(6) cannot be integrated analytically. We must therefore resort to numerical integration. Discretization Before numerically integrating equations (5) and (6), we must first discretize them. We replace the continuous time variable t with a time index indicated by the subscript i, so that the state variables h,V , are defined only at discrete times t0 , t1 , t2 . . . ti . . . t → ti h(t) → hi V (t) → Vi h hi t t 0 t 1 t 2 ... t i V Vi t 0 t 1 t 2 ... t i t Figure 2: Continuous time traces approximated by discrete time traces. 2 The governing ODEs (5) and (6) can then be used to determine the discrete rates at each time level. ḣi = Vi V̇i = −g − (7) 1 CD A ρVi |Vi | 2 m (8) As shown in Figure 3, the rates can also be approximately related to the changes between two successive times. dh Δh hi+1 − hi ≃ = dt Δt ti+1 − ti dV ΔV Vi+1 − Vi = ≃ = dt Δt ti+1 − ti ḣi = V̇i (9) (10) Equating (7) with (9), and (8) with (10), gives the following difference equations governing the discrete state variables. hi+1 − hi = Vi ti+1 − ti 1 CD A Vi+1 − Vi = −g − ρVi |Vi| ti+1 − ti 2 m hi +1 hi (11) (12) Δh . hi Δt ti t i +1 Figure 3: Time rate ḣ approximated with finite difference Δh/Δt. Time stepping (time integration) Time stepping is the successive application of the difference equations (11),(12) to generate the sequence of state variables hi ,Vi . To start the process, it is necessary to first specify initial conditions, just like in the continuous case. These initial conditions are simply the state variable values h0 ,V0 corresponding to the first time index i = 0. Then, given the values at any i, we can compute values at i+1 by rearranging equations (11) and (12). hi+1 = hi + Vi (ti+1 − ti ) � � CD A 1 (ti+1 − ti ) Vi+1 = Vi + −g − ρVi |Vi| 2 m Equations (13) and (14) are an example of Forward Euler Integration. 3 (13) (14) i t h V 0 0 h0 V0 1 input data 2 ... formula i i +1 (15) (13) (14) ... Figure 4: Spreadsheet for time stepping. Arrows show functional dependencies. Numerical implementation A spreadsheet provides a fairly simple means to implement the time stepping equations (13) and (14). Such a spreadsheet program is illustrated in Figure 4. The time ti+1 in equations (13) and (14) is most conveniently defined from ti and a specified time step, denoted by Δt. ti+1 = ti + Δt (15) It is most convenient to make this Δt to have the same value for all time indices i, so that equation (15) can be coded into the spreadsheet to compute each time value ti+1 , as indicated in Figure 4. This is much easier than typing in each ti value by hand. More spreadsheet rows can be added to advance the calculation in time for as long as needed. Typically there will be some termination criteria, which will depend on the case at hand. For the rocket, suitable termination criteria might be any of the following. hi+1 < h0 rocket fell back to earth hi+1 < hi rocket has started to descend Vi+1 < 0 rocket has started to descend Accuracy The discrete sequences hi , Vi are only approximations to the true analytic solutions h(t), V (t) of the governing ODEs. We can define discretization errors as Ehi = hi − h(ti ) EVi = Vi − V (ti ) (16) (17) although h(t) and V (t) may or may not be available. A discretization method which is consistent with the continuous ODEs has the property that |E| → 0 as Δt → 0 The method described above is in fact consistent, so that we can make the errors arbitrarily small just by making Δt sufficiently small. 4