VEGAdapt stands for "Robust & Adaptable Launcher TVC Control Systems for the VEGA Evolution" and is an ESA-ESTEC Network Partnering Initiative (NPI), jointly funded by ESA/UoB/ELV and with additional EPSRC funding, awarded to investigate linear parameter varying and adaptive control techniques for the VEGA Evolution launcher.
* Mr. Diego Navarro-Tapia (PhD student)
The VEGA configuration seemingly resembles traditional launchers but differs significantly to the other European launchers from a flight controls, actuation and propulsion perspective. The first three stages (respectively P80, Z23, and Z9) are solid while the upper stage's main engine is liquid allowing up to five boosts for injection including a de-orbitation function. All four stages are controlled in attitude using a newly developed Thrust Vector Control (TVC) system driven by a high performance electromechanical actuation (EMA). The upper stage, called AVUM, is complemented with the Roll and Attitude Control System (RACS) to provide fine attitude control functions using six 220 Newton thrusters. The avionics bay is located on the AVUM where its flight computer implements the Flight Program Software (FPS-A) for guidance, navigation and control (GNC).
Launcher TVC flight control systems for atmospheric flight are heavily impacted by several issues such as: perturbations coming from aero-servo-elastic loads, their coupling with mass variations, filtering offsets, aerodynamic instability due to vehicle’s design aspects, nonlinearities in the actuator, and propulsion effects among other. For example, in terms of the evolution of the aerodynamic structural interaction dynamics along the flight, it is noted that they involve varying mode frequencies and damping factors, both of which are badly known and thus difficult to establish with a good level of confidence using computational modelling tools. Indeed, analysis of flight data recordings have evidenced significant modelling mismatch and moreover, the TVC control and filtering function showed mode resonances of the first and second mode. In addition, the aero-servo-elastic interactions when coupled with rapidly varying mass properties are prone to instability. With respect to filtering offsets, even slight deviations have the potential to damage the vehicle by resulting in unnecessarily high aerodynamic loads. This is especially critical since the launch vehicle is aerodynamically unstable due to the center of pressure being located above the center of mass. Compounding the difficulties, nonlinear effects in the EMA such as hysteretic and backlash induce undesirable effects in the closed-loop such as delays, which is another potential contributor to instability in the low frequency range. Finally, and from the solid propulsion system perspective, the motion suffers from roll torque perturbations that spin up the launcher during the atmospheric flight potentially leading to unwanted separation conditions.
With the current design approach it is hard to achieve over the entire flight uniform stability and performance robustness characteristics. Furthermore, the TVC control laws settings need to be updated for each flight to mission specific configurations and trajectories. Thus, globally performing control laws based on the current state of the art PID architecture are not possible, which explains the need to retune the controllers so that each mission has the adequate levels of stability margins and performance.
All in all, the future control design challenges for the next generation TVC control systems are determined by the high level of uncertainty variation across flight, the nonlinearities in the system, as well as the aerodynamic/mass/structural/propulsion properties of the launcher. Thus, the future designs must hold over all mission and payload configurations in the face of a large set of severe external wind and load conditions.
It is obvious that the above has an impact on planning, costs and flight responsiveness. These shortcomings and effects motivate that for the next generation launchers it is necessary to develop envelope-extending control laws with improved mission responsiveness capabilities.
ESA-ESTEC has been working for about a decade in support to the development and qualification of the current VEGA GNC system. During initial developments the use of robust, adaptive and predictive control technologies were suggested and favoured by ESTEC over the currently implemented classical control technology. Without a suitable heritage in understanding the system impacts of adopting new technologies it was not the right time to take risks in adding complexity to the development process against a challenging schedule. In addition, and especially critical, the computational infrastructure was the main limiting factor. The current flight computer allows implementation of only simple classical low dimensional control solutions as the computational resources are already fully exhausted.
Although the current design is sub-optimal in meeting performance and operational requirements, the VEGA control laws have demonstrated via a series of consecutive flights that the design is well mastered given the tight set of industrial constraints. Nonetheless, ESA analyses have shown that no noticeable improvements would be obtainable without fundamentally altering the current design. Thus, in the face of an aggressive launch service market, functional GNC improvements to enhance mission responsiveness must be proposed. It is necessary to progress towards a rationalized GNC mission configuration process with short turn-around time while coping over a large domain of payloads.
Thus, the aim of the activity is to study robust and optimally scaled adaptable control laws that can provide extended launch vehicle missions and safety envelope capabilities.
The technical objectives of the activity can be listed as follows: