Periodic control for minimum-fuel aircraft trajectories

Wang

Proceedings of the 33rd Chinese Control Conference

et al. 2014

This paper concentrates on optimization of the steady-state cruise for a hypersonic vehicle.The steady-state cruise is a state with constant altitude and velocity during the flight.The ratio of fuel consumption to the range over a given period of time is defined as the fuel rate,which is the objective of this fuel optimized problem in this work.The optimal steady-state cruise is determined by minimizing the fuel rate with multi-constraints including boundary constraints of states and path constraints.Since this work needs to find the global solution satisfying the nonlinear constraints,we can apply particle swarm optimization(PSO) combined with the sequential quadratic programming(SQP) to find the optimal solution.Thus,this strategy takes the advantage of intelligent optimization algorithms with global search ability and SQP with high precision and handled constraint ability.A demonstration example results show good performance of fuel savings on the nonlinear model of the hypersonic vehicle.

Section: Introductionmentioning

confidence: 72%

Optimization on steady-state cruise for a hypersonic vehicle

Wang

Proceedings of the 33rd Chinese Control Conference

et al. 2014

Journal of Guidance, Control, and Dynamics

“…With a single actuator exciting the structure, the torsional and only one of the lateral modes are significantly excited (see Ref. 4 for results pertaining to this case). The multi-input excitation provides a richer picture of the system's dynamics.…”

Section: Multi-input Systemmentioning

confidence: 97%

“…w/(0) = 0 (4) The thrust, lift, and drag force models are T = 6r max (F,/0, L = C L (p/2)V 2 S, and D = C D (p/2)V 2 S, where C D is given by Eq. (la).…”

Section: (*Cyc)=7«>)mentioning

confidence: 99%

Dynamic decrease of drag by optimal periodic control

Sachs

1991

lo ir°-s. V VT^i •j^v r^ -0.0 10.0 TIME (sec) 20.0 30.0 40.0 ol § DAMPING 0.100 0 o il ).0 TIME (sec ... / ( 1 0 i£_. TO ,_._. 0.0 20.0 30.0 4 Fig. 4 System poles 4-input/l-output sensor X\ (18th-order lattice).algorithm, the two closely spaced modes can be progressively distinguished. Extraneous poles due to noise and/or nonlinearities are recognized by their erratic nature; whereas genuine system poles behave consistently. Multi-Input SystemFor the multi-input time-invariant case, data are recorded while all four actuators are simultaneously excited. The input sequences are digital white noise bandlimited between 0.25 and 2 Hz sampled at 12.5 Hz in order to excite the low-frequency dynamics of the structure. The system poles are successively extracted for 4-input/l-output and 4-input/4-output configurations.Sensor X\ data are run through an 18th-order 4-input/ 1-output lattice filter. The resulting modal frequencies and damping ratios for sensor X\ are shown in Fig. 4: one mode at 0.97 Hz with 0.09 damping and another mode at 0.68 Hz with 0.06 damping. Similar results, not shown, are obtained by running sensor Y\ data through an 18th-order 4-input/l-output lattice filter: one mode at 0.97 Hz with 0.09 damping and one at 0.68 Hz with 0.09 damping. Structural analysis of the truss shows one torsional mode and two lateral modes, one along X and one along Y. Thus, sensor X\ sees the torsional and X-lateral modes, and sensor Y\ the torsional and Ylateral modes. From the results, it can be deduced that the 0.97-Hz mode is the torsional mode since the damping is the same on both X\ and Y\ 9 whereas the two lateral modes are at 0.68 Hz as proved by the difference in damping ratios. Previous single-input identification testing with lattice filters had not clearly revealed the existence of two modes at the same frequency. With a single actuator exciting the structure, the torsional and only one of the lateral modes are significantly excited (see Ref. 4 for results pertaining to this case). The multi-input excitation provides a richer picture of the system's dynamics.The MIMO data are run through an 18th-order 4-input/ 4-output lattice. All sensors X\, X 2 , Y { , and Y 2 are constrained to see the same dynamics. The outcome is a 0.97-Hz mode with 0.09 damping and a 0.68-Hz mode with 0.07 damping. The lattice filter has merged the two lateral modes into a single mode because it is not configured to separate X dynamics from Y-dynamics. To ensure that both lateral modes are distinguished, the lattice should be set up so that X sensors t£if X 2 ) and Y sensors (Y l9 Y 2 ) observe different dynamics, in which case the a, would be 2 x 2 matrices. ConclusionThe vector-channel lattice filter was used to identify the truss experiment's modal frequencies and damping ratios for two important cases: a time-varying system with order change and a multi-input system configuration. The time-varying test was special in that the variation in parameters was significant as well as sudden, but, more importantly, the effective order of the s...

AIAA Guidance, Navigation, and Control Conference

“…See, for example, Refs. [2][3][4][5][6][7][8][9][10][11][12][13]. However, not much prior research accounts for the optimal landing problem.…”

Section: Introductionmentioning

confidence: 97%

Initial Guess Generation for Aircraft Landing Trajectory Optimization

Bakolas¹,

Zhao

Tsiotras

2011

We present a semi-analytic framework for the generation of initial guesses for the numerical solution of the landing trajectory optimization problem of an aircraft. Our approach consists of the following tasks: First, we introduce a geometric framework for the generation of length-suboptimal, curvature-constrained, threedimensional curves, which satisfy the following requirements: 1) the projection of the curves on the horizontal plane correspond to Dubins-like paths, and 2) an aircraft traveling along these curves is descending continuously until it reaches its final destination. Subsequently, we generate a time-parametrization of the geometric path such that the control inputs required to traverse the geometric path do not exceed the control limits of the aircraft. The resulting time-parameterized path along with the control input time histories is subsequently fed as an initial guess into a numerical optimal control solver, such as the one introduced in Ref. [1]. Simulation results demonstrate the advantages of employing the proposed initial guess generation scheme in terms of convergence of the numerical landing trajectory generation.