||Wright-Patterson AFB, OH 454337542
The next generations of gas turbine engines for aircrafts and power generation systems are facing new challenging requirements for operational and maintenance capabilities in addition to the ever increasing requirements for performance, reliability, and cost of ownership. For example, future engine systems must develop integrated thermal management (Ref. SAB Report), eliminate or limit the operational impact of fuel cooling, and provide system-wide Prognostics Health Management (PHM) without affecting the level of reliability and the physical/performance attributes. In order to meet these goals, engine control systems must have both the increased knowledge needed to adjust operation and the physical ability to respond to changes and events. In the past, engine control systems, designed around the Full-Authority Digital Engine Control (FADEC) were primarily responsible for control of airflow throughout the engine and the safe delivery of fuel. However, a modern FADEC is a complex controller responsible for many subsystems for optimizing performance and detecting failures. Yet, classical controls technology is inadequate for managing performance under both dynamic and steady-state, as well as input-output constraints and at the same time managing localized events such as engine component deterioration or transient behavior, instabilities, flutter, and flow-separation. To manage these dual responsibilities and provide reliable system operation, use of adaptive controls along with distributed, localized active control strategies may be a viable approach. Implementation of these strategies requires distributed sensors, actuators, and processing devices that can withstand the harsh engine environments.
The transition to a distributed, loosely coupled but information integrated control system represents a new paradigm for the hitherto centralized FADEC based engine control systems. This paradigm represents some unique challenges and opportunities because of the combination of criticality, environmental constraints, and the need for thorough validation before deployment. The distributed system paradigm also represents an opportunity for increased deployment of adaptive/condition based control strategies. Analysis of performance of such loosely coupled distributed systems under dynamic conditions is essential for gaining a fundamental understanding of their behavior and for deployment of new strategies. However, the analysis and validation of these strategies on a real FADEC system in a real environment can be time consuming and expensive. This suggests that a simulation-based research technique would be a viable alternative in this situation. In addition, the ability to use physics based models to synthesize information that may not otherwise be measurable can add new insight and capability to the control system.
The above discussions suggest several potential areas for research and analysis of propulsion systems in the context of a distributed propulsion control system environment. Proposals are solicited in all of these areas including some that are briefly described below.
1. Controls: New control techniques that may include model-based control and model-predictive control are prime candidates for large-scale multivariable plants. Dynamic Inversion (DI) is a good, straightforward methodology for designing multi-variable control laws for nonlinear systems. Application of adaptive nonlinear constrained DI control strategies may result in new control algorithms and/or requirements that influence the design of elements of the distributed architecture. Another area of interest in this context involves techniques/algorithms for efficient partitioning of controls algorithms so that they can be mapped onto the distributed control elements and multi-processors. Other areas of research include, but not limited to, development of control algorithms for distributed sensors, actuators, and other propulsion subsystems; failure management/accommodation strategies based on neural networks or fuzzy logic.
2. Simulation, Integration, Software Verification and Validation (V&V): In the distributed control framework, the stability and performance of the closed loop control system is impacted by various communication constraints such as data latency, bandwidth limitations, non-deterministic operation, and asynchronous operation and/or failure events such as packet dropouts or irregular or uncontrolled communications. The best vehicle for analyzing these effects and developing new solutions is a simulation environment in which all these effects can be modeled, analyzed, and candidate strategies for mitigating them be evaluated. This simulation environment would be useful for evaluating various communication protocols and strategies, and their effect on system stability and software/hardware complexity. This environment would be useful for developing and analyzing open, plug-and-play system hardware/software elements. The end result should be the development of requirements for the communication networks and validated algorithms for managing these events and disturbances. Clearly, research in all of these areas would fill an important need and contribute to the realization of distributed FADEC system architectures. The complexity of FADEC software and consequently the cost of V&V have greatly increased in recent years for a variety of reasons. A distributed architecture presents an opportunity to reduce this complexity and V&V cost via partitioning. It is also highly desirable to develop new methods to modularize the V&V process for the distributed components in order to reduce the complexity and cost of V&V.
3. Diagnostics and PHM: In the past, most FADEC systems utilized a combination of Built In Tests (BIT) and Fault Detection and Isolation (FDI) Algorithms to detect failures of control sensors and actuators for diagnostics and maintenance purposes. The use of PHM sensors solely for the purpose of monitoring and diagnosing engine health is relatively new. The transition to distributed architecture presents an opportunity to partition the PHM functions and yet functionally integrate them with the control algorithms for improving the overall system availability in the presence of degradations. In a distributed FADEC control system with distributed PHM the amount and types of health/degradation information about the distributed components is likely to be significantly more in quality and quantity. The integration of all this information in real-time would lead to development and validation of diagnostics reasoner, which could be an area of research interest because of its potential for extracting the full benefits PHM in a distributed FADEC system.
Distributed Propulsion Control; PHM; Model-based control; Real-time simulation; FADEC; Sensor; Software V&V