I worked on two projects during my winter internship at Princeton Satellite Systems: a two-stage-to-orbit (TSTO) launch vehicle design proposal related to the NASA Space Launch System (SLS) and a satellite conjunction maneuver demo. These both used the Spacecraft Control Toolbox for MATLAB.
One of the main ideas behind the TSTO launch vehicle project is to propose an all-liquid variant of the SLS. Currently, the SLS first stage is mostly powered by two solid rocket boosters (SRB) upgraded from the Space Shuttle SRBs. However, our proposal is to replace the two SRBs with five liquid boosters (LB), each mated with an RS-25 engine. The second stage would remain the same. Using MATLAB, I analyzed the launch and trajectory performance of both variants and found similar performance. Additionally, the total mass of the all-liquid SLS variant would be approximately two-thirds the mass of the SRB-powered spacecraft. An approximate CAD model of the all-liquid SLS version is shown below.
In addition, the LBs can be used independently to power smaller high-performance TSTO launch vehicles that carry around 8,000 kg of payload to low earth orbit. Trajectory plots and a preliminary CAD model are shown below.
My other project this internship was to help out with a satellite conjunction avoidance demo with Ms. Stephanie Thomas. The goal was to create a solution in MATLAB to identify potential satellite-debris conjunctions and develop a method/algorithm to avoid the conjunctions. I mainly worked on testing the code and relevant functions and providing feedback about the solution’s comprehensiveness.
Overall, I greatly enjoyed this internship and the opportunity to work at PSS. I saw firsthand how even a small company can make significant contributions to aerospace and engineering through diverse interests yet specific, impressive skill sets.
IAC 2022 is underway! Annie Price, a former PSS intern, and Mike Paluszek are attending. The Congress has hundreds of technical talks and poster presentations. In addition, there is a huge technology showcase area. Both companies and government organizations have booths. Here are some photos from the show floor.
A 1 N green propellant thruster. It is a few centimeters in length.
I also met engineers from Boeing, DLR, Teledyne, Lockheed Martin, MDA, Sierra, Rolls-Royce, Saudi Arabia, South Korea, Slovakia, Sweden, and many other companies and countries. There were at least three robot arms on display, including one by Kinetik Space.
This one has selectable end-effectors.
I met an engineer who worked on the Apollo program. His area was radiation hardness. He said back then, no one knew much about the problem.
There were excellent talks on Tuesday on formation flying and rendezvous. Annie is presenting our talk on a fusion-powered Titan aircraft on Thursday, in SESSION 10-C3.5, Joint Session on Advanced and Nuclear Power and Propulsion Systems,” In W08 at 13:45.
Hohmann transfers are a well-known maneuver used to change the semi-major axis of an orbit. The Spacecraft Control Toolbox allows you to compute the required velocity changes, and integrate them into a full simulation.
In this demonstration, we create a 6U CubeSat that has 3 orthogonal reaction wheels and a single hydrazine thruster. The thruster is aligned with the body x-axis and must be aligned with the velocity vector to do the maneuver. An ideal Hohmann maneuver is done with impulsive burns at two points in the orbit. In reality, with a thruster, we have to do a finite burn.
rI = [-7000;0;0];
vI = [0;-sqrt(mu/Mag(rI));0];
OrbMnvrHohmann(Mag(rI),rF);
[dV,tOF] = OrbMnvrHohmann(Mag(rI),rF);
The first time OrbMnvrHohmann is called, it generates the plot below of the planned Hohmann transfer. The function computes the delta-V and also the time of flight, which will be used to determine the start time of the second thruster burn.
We create a short script with numerical integration to implement the maneuver using a thruster. The burn durations are computed based on the thrust and the mass of the spacecraft. In this case, they are about three minutes long. The maneuver is quite small, so the mass change is not important. The attitude control system uses the PID3Axis function which is a general-purpose attitude control algorithm. The simulation is a for loop, shown below. The ECI vector for the burn is passed to the attitude control system, which updates every step of the simulation.
% Simulation loop
for k = 1:n
% Update the controller
dC.eci_vector = uBurn(:,kMnvr);
[tRWA, dC] = PID3Axis( x(7:10), dC );
% Start the first burn
inMnvr = false;
if( t(k) > tStart(1) && t(k) < tEnd(1) )
inMnvr = true;
end
% Switch orientation
if( t(k) > tEnd(1) )
kMnvr = 2;
end
% Start the second burn
if( t(k) > tStart(2) && t(k) < tEnd(2) )
inMnvr = true;
end
if( inMnvr )
dRHS.force = thrustE*QTForm(x(7:10),dC.body_vector)*nToKN; % kN
else
dRHS.force = [0;0;0];
end
el = RV2El(x(1:3),x(4:6));
xP(:,k) = [x;tRWA;Mag(dRHS.force)/nToKN;el(1);el(5)];
% Right hand side
dRHS.torqueRWA = -tRWA;
x = RK4(@RHSRWAOrbit,x,dT,0,dRHS);
end
The maneuver logic just waits a quarter orbit then performs the first burn, by applying the thrust along the body vector. It then waits for the time of flight and then starts the next burn. The start and stop times are pre-computed. RK4 is Fourth Order Runge-Kutta, a popular numerical algorithm included with the toolbox.
At the final orbit radius an attitude maneuver is needed to reorient for the final burn.
The spacecraft body rates, in the body frame, during the maneuver are shown below.
The reaction wheel rates are shown below. The simulation does not model any particular wheel. Friction is not included in the simulation, although the right-hand-side function can include friction.
The wheel torques and rocket thrust are shown below. The thruster is a 0.2 lbf hydrazine thruster that is based on the Aerojet-Rocketdyne MR-103. The PID controller does not demand much torque.
The semi-major axis and eccentricity are shown below. The middle portion is during the transfer orbit.
The eccentricity is zero at the start and finish. Note the slope in both eccentricity and semi-major axis due to the finite acceleration. At the end of the simulation, we print the achieved orbital elements:
Final SMA 7099.72 km
SMA error 0.28 km
Final e 1.3e-05
The result is very close to the ideal solution!
This post shows how you can easily integrate attitude and orbit control. Email us for more information! We’d be happy to share the script. We can also offer a 30 day demo to let you explore the software.
Our aerospace theory textbook, Spacecraft Attitude and Orbit Control, has been included with purchases of the Spacecraft Control Toolbox for years and available for purchase as a standalone PDF. We have now compiled our book as an eBook and it is available from Barnes and Noble for Nook:
A third planet, as large as 26% of the mass of Earth, has been discovered orbiting our nearest stellar neighbor, Proxima Centauri .Astronomer João Faria and his collaborators detected Proxima Centauri d using the Echelle Spectrograph for Rocky Exoplanets and Stable Spectroscopic Observations.
It would be exciting to send a spacecraft to enter the Alpha-Centauri system and orbit this planet. At Princeton Satellite System we’ve looked at interstellar flight using the Direct Fusion Drive nuclear fusion propulsion system.
Interstellar Fusion Propulsion
At the 2021 Breakthrough Energy Conference we presented findings for both flyby and orbital missions. Flyby missions are easier, but orbit entry would allow detailed study of the planet. A flyby gets your spacecraft close, but it is moving really fast!
The following charts give an outline of our talk. The first shows the optimal exhaust velocity based on sigma, the ratio of power to mass. Our designs have a sigma from 0.75 to 2 kW/kg. With 2 kW/kg, the optimal exhaust velocity is 4000 km/s. The mission would take about 800 years. Our current designs can’t get exhaust velocities higher than 200 km/s. We’d need another method to produce thrust.
Mission Analysis
The next plot shows a point mission that reaches Alpha Centauri in 500 years. This requires a sigma of about 20. The spacecraft accelerates and decelerates continuously. The mission could be improved by staging, much like on a rocket that launches from the Earth into orbit.
Selected Mission
The next figure shows how the starship would enter the Alpha Centauri system.
Alpha Centauri System Insertion
The final plot shows the orbital maneuvers that lower the orbit and rendezvous with the planet.
Lowering the orbit to rendezvous with the planet.
Even 500 years is a long time! This is over ten times the lifetime of Voyager, but much less than some engineering marvels built on the Earth.
We hope to someday be able to build fusion powered spacecraft that will head into interstellar space!
NASA would like a crew to land on the moon by 2024.
We didn’t have time to write a proposal, but here is our design. We propose a single stage vehicle, that can land from and return to a 15 km circular orbit. It uses 2 Blue Origins BE-3U engines that use cryogenic hydrogen and oxygen. An Orion capsule houses the astronauts. The Orion would take astronauts to and from Gateway and to and from the Earth. Lockheed Martin is building the Orion spacecraft. The European Space Agency is building the service module. A separate transport would bring fuel and payload to the lander. In the future, the lander could be refueled from lunar water.
The dimensions are in meters. The Orion is shown below. We purchased the model from https://hum3d.com.
The landing gear were scaled from the Apollo Lunar module.
It is interesting to compare its size with the Apollo Lunar Module. The Artemis is designed to fit into the 10 m SLS fairing. This a fully reusable lunar vehicle that can be refueled. It is designed for a long-term, sustainable, lunar base.
We use two toroidal hydrogen tanks and two spherical oxygen tanks. The cylinder on the outside is the solar array producing 34 kW of power. Of course, numerous details are omitted. We developed this model using our Spacecraft Control Toolbox. The design script will be available in the Spacecraft Control Toolbox Version 2019.1 due in mid-November.
Other elements of the lander were designed for different purposes. The GN&C system is based on our Army Precision Attitude Control System.
Our control system is based on a robotic lander we designed some time ago. We have full C++ code for the control and guidance system.
The architecture for Earth/Moon transportation system is shown below. Eventually, a Direct Fusion Drive freighter would be the main way of moving cargo between Earth orbit, lunar orbit and Gateway. The lander would remain in lunar orbit. Humans would go to the moon using fast orbital transfer, much like during Apollo.
Our next blog post will show how we get from Gateway to and from our 15 km starting orbit. A subsequent post will demonstrate our lunar landing guidance that uses a neural network for navigation based on images of the surface. Using it for landing would require higher resolution images than we have today, but short of building a lunar GPS system, it might be more cost-effective to have a satellite assembling images from low lunar orbit.
We will also update this blog post from time to time. Stay tuned!
The spectacular success of the NASA New Horizons mission has led to many new discoveries about Pluto. The next step would be to send an orbiter. That isn’t easy to do with chemical propulsion but could be done with Direct Fusion Drive.
We’ve done a preliminary mission analysis for a Pluto orbital mission. We are baselining a Delta IV Heavy that can put up to 9,306 kg into interplanetary orbits. These plots show various parameters versus mission duration. The maximum duration is the same as the New Horizons mission, 10 years.
Let’s use the 4 year mission as a baseline. It would use a 2 MW DFD engine to reach Pluto in about 4 years and go into orbit. The engine would thrust for 270 days out of the 4 year mission producing 110 km/s delta-V. The trajectory is shown below
Once there, almost 2 MW of power would be available for the science mission, over 10000 times as much power as is available to New Horizons! The New Horizons bit rate is no more than 3000 bits per second. The high power would allow for a bit rate of over 135 Mbps for data transmission back to Earth using the JPL Deep Space Optical Communications System and a 30 kW laser transmitter. The time in transit is much shorter than New Horizons and would produce significant savings on operations costs. Launch times would be more flexible since gravity assists would not be needed.
DFD would use deuterium and helium-3 as fuels. Only 1700 L of helium-3 would be needed for this project. Current U.S. production of helium-3 is about 8,000 L per year.
Since we would be going all the way to Pluto it would make sense to include a lander. One way to power the lander is using laser power beamed from the orbiter. Here are results for a possible system, beaming over 30 Wh per pass from a 200 km orbital altitude.
Currently, experiments are taking place in the Princeton Field Reversed Configuration laboratory. Here is the machine in operation at the Princeton Plasma Physics Laboratory:
The next step is to build a slightly larger machine to demonstrate fusion. Fusion power generation has been demonstrated in the Princeton Plasma Physics Laboratory Tokamak Fusion Test Reactor and the Joint European Torus but never in a machine using helium-3. A flight engine would follow. Its small size would keep the development and production costs down.
DFD would enable many challenging missions include human exploration of Mars, Europa landers and interstellar probes.
If you are sending a spacecraft to the moon, you will be interested in lunar topography. A new function in the Spacecraft Control Toolbox lets you superimpose a height map onto any sphere.
The function RSHMoon.m gives you the Clementine spacecraft topographic data using a spherical harmonic expansion of the rangefinder data.
A new function, PlanetWithTerrain.m, lets you superimpose this data onto a sphere.
The rocket equation gives the ratio of the initial mass to the final mass given a velocity change and an exhaust velocity.
$$\frac{m_i}{m_f} = e^\frac{\Delta V}{V_e}$$
This seems to say that given enough fuel we could get an infinite velocity change! To see what the maximum possible velocity change could be we need to account for the structural fraction. The structural fraction multiplied by the mass of fuel gives the mass of the structure needed to support and contain the fuel. The rocket equation now is as follows
where is the mass of propellant, is the structural fraction, and is the mass of all other hardware. If we let the mass of propellant go to infinity, and solve for the velocity change, we get:
$$ \frac{\Delta V}{V_e} = \log{\frac{1+f}{f}}$$
The following plot shows the ratio of velocity change to exhaust velocity for a range of structural fractions.
There is a lot of interest in lunar landing missions for both scientific exploration and commercial purposes. Commercial applications might include mining helium-3 for future nuclear fusion power plants on earth and mining water for rocket fuel.
The Spacecraft Control Toolbox makes it easy to do preliminary planning for lunar missions. In this blog we present a single MATLAB script that takes a spacecraft from a low Earth parking orbit to the lunar surface! Here is the final segment, the descent to the moon.
We ended up with a 30 kg dry mass for a spacecraft that can use an ECAPS 220 N HPGP thruster for delta-v.
