Mitigating Pogo on Liquid-Fueled Rockets

Accelerations induced by pogo were a critical factor in the human-flight rating of the Titan II launch vehicle. Analytical investigations at Aerospace explained pogo occurrences and led to a successful resolution prior to the first Gemini mission.

Mitigating Pogo on Liquid-Fueled Rockets

Kirk Dotson

Interaction of a launch vehicle's propulsion system and structure can be a source of dynamic instability. Since the days of the Gemini program, Aerospace has been finding better ways to model and mitigate this potentially disastrous phenomenon.

Launch vehicles achieve thrust through the combustion of liquid or solid fuel in their rocket engines. In a liquid-fueled vehicle, the engine pumps propellants (fuel and oxidizer) through feed lines from their storage tanks to the engine's thrust chamber. Inevitably, the tanks, feed lines, and engine vibrate during liftoff and ascent (see sidebar, Resonance). This vibration causes the flow of the propellants in the feed lines and engine to oscillate, leading to thrust oscillation. The resulting thrust oscillation can cause the structure to vibrate even more, which increases the fluid oscillations, which causes greater vibration, and so on in a progressive feedback loop. This represents a system instability, and the resulting oscillations can become extreme.

This dynamic interaction between the vehicle structure and the liquid propellants was first recognized during development of the Titan II in 1962. It had occurred on previous launch vehicles as well, but the phenomenon was not yet understood. The engineering community nicknamed the phenomenon "pogo" because it caused the launch vehicle to stretch and compress like a pogo stick (see sidebar, What is Pogo?). Pogo presented serious challenges for the developers of Titan II and remains a prime consideration in the design of launch vehicles today. Then, as now, Aerospace work on the pogo phenomenon has helped prevent potential mission failures.

Pogo and Gemini

The Gemini program followed the Mercury orbital missions and preceded the Apollo lunar expeditions. The primary goal of Gemini was to demonstrate the feasibility of a rendezvous of two or more spacecraft in orbit.

The Gemini capsule, designed to carry two astronauts, was to be launched on a modified Titan II ballistic missile. During its first development flight, the Titan II experienced pogo oscillation going from 10 to 13 hertz over a 30-second period during mid-burn of the first stage. At 11 hertz, this shaking reached a maximum of 2.5 g's at the payload. Superimposed on the steady acceleration, the force of this motion was excessive for military use and clearly unacceptable for an onboard crew.

NASA wanted to keep vibration levels below 0.25 g's to ensure the operational capability of the astronauts, although 1 g was tolerable in terms of the structural integrity of the Titan II. An engineering analysis suggested that the pogo interaction could be minimized by equipping the oxidizer feed line on each engine with an accumulator—essentially a container of gas that acts like a soft spring to reduce the fluid frequency well below the structural frequency and weaken the feedback. After achieving what appeared to be an adequate mathematical model of the phenomenon, the Titan development team installed the two accumulators on the eighth Titan II flight (N-11). This was the first attempt to mitigate pogo interactions for Gemini.

Depiction of pogo occurrence. Due to the time-varying structural dynamic properties of a launch vehicle, the structure-propulsion feedback is not sustained, but rather leads to a "blossom" in the launch vehicle's longitudinal response. This "blossom" occurs over a frequency range. Natural frequency is inversely proportional to the square root of mass; therefore, as propellants are consumed during flight, the natural frequency of the launch vehicle mode increases with respect to time. The maximum pogo response corresponds to close tuning of the structural and hydraulic frequencies.

The result was unexpected and dramatic: peak vibration levels reached 5 g's—much worse than prior launches without pogo mitigation! Rather than suppress the pogo oscillations, the accumulator on the oxidizer line actually made them worse, triggering a premature shutdown of both engines that resulted in mission loss.

Pogo now became the top concern for the Gemini program. Clearly, the understanding of pogo was inadequate, and human-flight rating was in jeopardy.

The Air Force asked Aerospace to step in as part of a Titan II improvement plan. The Aerospace team was led by Sheldon Rubin. He examined pressure recordings from static engine firings conducted a year earlier and identified the key missing element in the pogo model: cavitation—the formation of bubbles in the fluid at the inlets to both the oxidizer and fuel pumps. Like the gas in an accumulator, these cavitation bubbles served to lower the vibration frequency of the fluids in the feed lines. Because the bubbles at the fuel pumps were not recognized, the analyses had shown the fuel frequency to be well above the structural frequency. The oxidizer frequency appeared to be closer to the structural frequency, so that's where the accumulators had been installed. In fact, the cavitation bubbles caused the fuel frequency to fall close to the structural frequency as well. Without the oxidizer accumulators, the oxidizer feedback partially canceled the fuel feedback through phasing of their thrust contributions. When the oxidizer feedback was weakened, the net effect was a greater instability.

The Aerospace model incorporated the effect of bubbles at both the oxidizer and fuel pumps and showed that the addition of an accumulator in the fuel line of each engine was essential to eliminate pogo. With both fuel and oxidizer accumulators installed, a flight on November 1, 1963, showed a reduction of vibration levels to 0.11 g's. After two subsequent launches with accumulators also met the NASA limit, the Titan II was declared suitable for human flight, and the Gemini program went on to achieve its mission objectives.

Pogo After Gemini

Since then, Aerospace has been intimately involved with pogo mitigation for numerous other programs. In 1963, for example, Rubin's team described the Gemini experience in a joint technical panel held on Thor-Agena pogo. Until that time, the Thor program sought not to suppress pogo but to strengthen the payloads to endure the vibration. Years later, when analysis predicted an increased pogo for an extended version of the Thor-Delta (predecessor of the Delta launch vehicle family), Aerospace recommended the installation of an accumulator, which succeeded in suppressing pogo.

Schematics of accumulators that successfully suppressed pogo on various vehicles. The concept of introducing bubbles near the tank outlet (panel f) was proposed for the Saturn V first stage, but this approach was rejected in favor of the one shown in panel e. Inadvertent effervescing of nitrogen gas from the oxidizer exiting the first-stage tank on Titan IIIE-2 had previously led to pogo instability (view larger image).

In 1964, Aerospace recommended a close-coupled configuration for oxidizer and fuel accumulators to improve their capability for the Air Force's Titan III. A new toroidal fuel accumulator was developed for Titan IIIB and used on all subsequent Titans. By 1967, new metal bellows accumulators were developed for Titan IIIM, as a result of extensive Aerospace involvement; these were first used on the third Titan IIIE and were standard on all subsequent Titan vehicles.

Apollo 6, the last unpiloted Apollo mission, exhibited a strong pogo oscillation. This craft was launched atop a Saturn V, the vehicle that would later carry the first astronauts to the surface of the moon. The pogo appeared during first-stage operation. Aerospace began an analysis for NASA, and concurred with a proposal to use trapped gas in the oxidizer prevalve to serve as an accumulator. Aerospace also recommended against an alternative proposal in which bubbles introduced near the tank outlet would be carried downstream to reduce the feed-line frequency. The accumulator approach was implemented on the first piloted flight, and pogo was permanently eliminated for the Saturn V first stage.

The five-engine second stage of Saturn V also experienced pogo, but the oscillations were concentrated at the center engine, so they were not felt by the astronauts. But on Apollo XII, the vibration at the center engine reached 8 g's and caused concern for the vehicle's structural integrity. Analysis predicted that the 15-g structural limit would not be exceeded, so no fix for pogo was implemented for Apollo XIII. But, as with the N-11 Gemini flight, the unexpected happened: Vibration levels reached 34 g's, causing premature shutdown of the center engine. The structure held together, and the mission was able to proceed using the four remaining engines. Again, NASA asked Aerospace to assess various prevention strategies. Aerospace supported the installation of a liquid-oxygen accumulator, which succeeded in suppressing pogo on all future Apollo flights.

Schematic of propulsion system. The accumulator volume must be carefully selected to ensure that the hydraulic and structural frequencies are well separated during flight.

In 1970, NASA published a monograph on pogo written by Rubin to be used for development of the space shuttle and subsequent vehicles. From 1971 to 1981, Aerospace conducted studies on space shuttle pogo suppression for NASA. The original space shuttle design called for an accumulator at the usual location: upstream from the engine's liquid-oxygen inlet. Preliminary studies at Aerospace sought to understand the complexities of potential interaction and uncertainties, particularly in terms of predicting the degree of pump cavitation. These studies indicated that the optimal location for the accumulator was deep within the engine itself, near the high-pressure oxidizer pump inlet. This represented a new approach to pogo mitigation, and the proposal met with considerable resistance because of the major impact on the engine development test program and the difficulties in implementing an accumulator in a region of such high pressure. Nonetheless, the engine accumulator was implemented, and pogo was eliminated for the shuttle. This was the first vehicle cured of pogo prior to a need shown by flight.

In 1989, Aerospace developed an advanced pogo stability analysis code using a building-block formulation and an automated technique for extracting the vibration characteristics of the coupled structure-propulsion system. The code has been used for stability analyses of the Atlas and Titan upper stages, as well as for the Titan IV and Delta IV boost vehicles. The next version of the software is being broadened in analysis capabilities for the Evolved Expendable Launch Vehicle. The effort includes comprehensive review of the characterization of propulsion elements and the elimination of many restrictions and limitations in the existing codes.

Recent History

Even after 40 years, the potential for pogo continues to cause concerns. A recent flight exhibited accelerations near the spacecraft interface that were significantly higher than those seen on a previous flight with a similar upper stage and spacecraft. Thus, as had happened with Saturn V, unexpectedly high responses were observed for similar missions without an apparent cause. This raised a concern for an upcoming Titan IV/Milstar mission, because the engine used on these previous flights would also be used, for the first time, on the Titan IV/Milstar mission.

Aerospace formed a multidisciplinary team to investigate the cause of the flight oscillations and to provide a risk assessment for the Titan IV/Milstar mission. The initial stage of the investigation revealed that a synchronization of the frequency, amplitude, and phase of the engine chamber pressure and structural response occurred during the earlier missions—which raised a concern that a pogo feedback loop existed between the propulsion system and the launch vehicle/spacecraft structure.

Aerospace and the contractor subsequently began an intensive effort to assess the pogo stability of the Titan IV/Milstar mission. Data from ground tests suggested that the engine oscillations could be associated with unsteady cavitation at the inlet of the liquid-oxidizer pump. That is, it was shown that the chamber pressure oscillations were most likely caused by cavitation bubbles at the pump inlet, which periodically formed and collapsed when the pump operated under a particular combination of inlet pressure, speed, and flow.

This Titan IVB rocket successfully launched a Milstar satellite in April 2003. Aerospace research limited the risk that liquid-fueled engine cavitation and dynamics would lead to system instability. The Aerospace and contractor team defined a mission profile that provided high confidence in mission success. (Russ Underwood, Lockheed Martin)

Equations for calculating pressure and flow oscillations across the pump interfaces existed for similar types of unsteady cavitation, but the equation coefficients were not known for the exact phenomenon that existed in this oxidizer pump. Moreover, the coefficients could not be identified from the available tests. Engineers combed through the existing literature and conducted pogo stability analyses to estimate the required parameters. The pogo model with the estimated pump parameters supported the hypothesis that the high accelerations on the earlier missions were caused by interaction of the launch vehicle/space vehicle structure with the propulsion system during periods when the oxidizer pump was undergoing unsteady cavitation.

The pogo model with the best-estimated pump parameters predicted that the Titan IV/Milstar mission had the potential to experience instability if it flew as planned. In the worst case, a pogo response for the Titan IV/Milstar vehicle posed a potential for damage or even mission failure.

While developing the pogo stability model, the Aerospace and contractor team also worked to identify the operating conditions at which the unsteady pump cavitation occurred sufficiently to induce propulsion-structure interaction. From prior flight data, they established that the cavitation phenomenon only existed in a well-defined region of dimensionless pressure and flow parameters, and if these conditions were avoided, the risk of pogo during the Titan IV/Milstar mission could be effectively mitigated. The proposed mitigation procedure, therefore, involved controlling the propulsion system operation to avoid the cavitation-induced engine dynamic behavior. The mitigation was implemented, and the Titan IV/Milstar mission flew on April 8, 2003, without any evidence of pogo, successfully delivering the satellite into orbit.

Conclusion

From the early days of Gemini to the latest Milstar launch, Aerospace work on the pogo phenomenon has been instrumental in preventing catastrophic mission loss. Forty years of experience has shown that pogo is not an isolated phenomenon, but can affect launch systems as diverse as the Delta, Titan IV, and space shuttle. Even launch vehicles with a pogo-free flight history are not always immune. As the launch community transitions to the Evolved Expendable Launch Vehicle and other future systems, Aerospace will no doubt be called upon to use its expertise to help prevent pogo and ensure continued mission success.

Acknowledgement

The author thanks Sheldon Rubin for his assistance in writing this article.