Lightweight, Damage-Tolerant Radiator for In-Space Power and Propulsion

The desire to explore deep space destinations with high-power and high-speed spacecraft inspired this work. Nuclear Electric Propulsion (NEP), shown to provide orders of magnitude higher specific impulse and propulsion efficiency over traditional chemical rockets, has been identified as an enabling technology for this goal. One of large obstacle to launching an NEP vehicle is total mass. Increasing the specific power (kW/kg) of the heat radiator component is necessary to meet NASA’s mass targets. This work evaluated a novel lightweight, high-temperature carbon fiber radiator designed to meet the mass requirements of future NEP missions. The research is grouped into three major sections: 1) a micro-scale radiation study, 2) two studies on bench-scale experimental and analytical investigations, and 3) large-scale radiator system modeling. The four studies presented in this work collectively demonstrate that a carbon fiber heat radiator offers many benefits over traditional metallic and composite radiators. This research elevated the technology readiness level of the carbon fiber fin from 2 to 4. Studying various aspects of the fin concept has provided a thorough understanding of the important design considerations. A micro-scale model of radiative heat transfer was used to explore the thermal radiation from an array of fibers. This model and its results were presented in Chapter 2. Calculations show that: (1) The dependence of the effective emissivity of a carbon fiber fin, that is the total hemispherical emissivity of the projected fin surface area, on fiber volume fraction was investigated using a Monte Carlo Ray Tracing model. It was determined that the maximum effective emissivity occurs at the smallest fiber volume fraction that is still optically thick, which occurs between 15-20% for a 0.5 mm thick array of uniformly packed fibers. As the fiber volume fraction decreases, scattering among the fibers increases the amount of radiation that escapes to space as compared to a flat fin surface, thereby increasing the effective emissivity of the fin. (2) This increase is quite significant since, for example, fibers with a surface emissivity of 0.7 will have an effective emissivity of 0.83 when packed in an array with about 13% fiber volume fraction. (3) The fiber volume fraction of the 2,000 count fiber tows supplied by Mitsubishi is about 12%, which is close to the ideal packing density for maximizing effective emissivity as predicted by the MCRT model. The design and fabrication of test articles was presented in Chapter 3. The experiments showed that: (1) In this proof-of-concept stage, a preliminary fin design was developed based on maximizing specific heat rejection using assumed thermal properties of the fin and heat pipe. (2) Protocols were developed for weaving the carbon fiber tows and for joining the fiber fin with the heat pipe using a titanium-copper-silver braze foil. (3) By collaborating with a textile manufacturer, it was proven that the high conductivity carbon fibers can be woven using automated, mass-production looms. This shows that production of the fin material can be readily scaled up. (4) The brazing method was successful for thin fiber fins, but a proper brazing facility is required for consistent brazing with thick fins due to the limitations of the heating element in the test setup available for this work. (5) The carbon fiber fin technology readiness level was increased from 2 to 4, which required validation of the fin in the environment. Measurements of material properties of carbon fibers and fin performance were presented in Chapter 4. The findings included: (1) A measurement of the spectral emissivity showed only a slight dependence on wavelength, which means that the fibers are nearly gray bodies. The emissivity at room temperature was 0.74, and the emissivity at 600°C was predicted to be about 0.78. (2) The measured emissivity was used to extract temperature data from IR images of the fins during operation. By comparing thermal models of the fin with surface temperature profiles extracted from IR images, the thermal conductivity of the bulk carbon fiber was determined to be 707±100W/m-K. With estimates of the bulk carbon fiber emissivity and thermal conductivity, models using these property values were used to predict the fin performance over a broader range of fin configurations. A system-level study of the applicability of the carbon-fiber radiators was presented in Chapter 5. The investigation demonstrated that: (1) An analytical approximation of the dependence of the heat radiated from a fin on fin length, thickness, and root temperature could be fit by a fourth order polynomial with an R2 value of 0.9999. This analytical approximation of the radiated power could then be incorporated into a larger, system-level model of the heat rejection subsystem used to design the radiator based on subsystem requirements. (2) A thermal resistance model of a heat rejection system appropriate for a high-temperature, high-power nuclear-electric powered vehicle was built. Given the heat load and power conversion system coolant temperature, this model predicts many heat rejection subsystem parameters including total mass, area, radiator fin efficiency, and coolant pumping power. (3) A constrained optimization was run to find the minimum mass heat rejection subsystem design for a test case using the thermal resistance model. The selected test case was a 300 kWe, potassium-Rankine power conversion system generating 700 kWt of waste heat. (4) The optimal design point had a coolant flow rate of 1.5 kg/s, coolant pipe diameter of 2.7 cm, fin length of 6.6 cm, and fin thickness of 3.95 mm, which resulted in an HRS mass of 224.5 kg. The fin length was slightly shorter than when the mass was optimized in Chapter 3 because there were other mass tradeoffs when considering all of the subsystem components. In the fully unconstrained problem, the optimal solution had a coolant flow rate of 4.6 kg/s, coolant pipe diameter of 4.5 cm, fin length of 5.2 cm, and fin thickness of 4.9 mm, and resulted in an HRS mass of 194.6 kg. This unconstrained design point could be feasible with advanced liquid metal pumps. (5) The area required to reject 100 kW was compared with published results from a low temperature heat rejection subsystem for a nuclear-electric power conversion system. For the high-temperature carbon fiber radiator, 17 m2 was required compared to the 170 m2 radiator sized for the low-temperature system. (6) This lightweight, high-temperature radiator may be the key to enabling nuclear-electric power for space applications where maximizing efficiency is vital to mission feasibility.