You are here

Innovative Interconnectivity and Scheduling of Smart Sensors and Actuators for Reliable Propulsion Systems Controls

Description:

OBJECTIVE: Develop an optimized distributed digital communication scheme with smart components, demonstrating the use of model-based controls for optimal performance and health monitoring of aircraft propulsion system. 

DESCRIPTION: Bridging the gap between model-based design and platform based implementation is one of the critical challenges for embedded software systems operating over a single communication bus along with sensors and actuators. Current engine control systems rely mostly on analog signals, circumventing many communication and timing considerations. In the context of embedded control systems that interact with an environment, a variety of errors due to quantization, delays, and scheduling policies may generate executable code that does not faithfully implement the model-based design. The performance gap between the model-level semantics of controllers and their implementation-level semantics can be rigorously quantified if the distributed controller implementation and its interaction with the communication bus are well-understood. The core characteristic of distributed engine control is the presence of a single digital data bus through which all information of interest must transit. In a perfect, distributed environment, no analog signal reaches a centralized processor anymore. Rather, the sensors and the control computers all take turns to transmit information over the data bus. Since they all communicate over a single bus, the frequency at which each sensor, actuator, and control computer is allowed to communicate over the data bus must be traded off against that of the other communicating elements. It is easy to figure that differing priority levels can result in profoundly different levels of performance for the closed-loop control system, independent of the particular communication protocol used. The concerns here are mostly related to the implementation of a distributed engine control architecture: assume all digital sensing components as simple as possible to reflect the condition of high temperature electronics components when they become initially available. Such components are assumed to perform only core functions, such as collecting and digitizing sensed information, and sending this information over the communication network. In this context, it becomes important to decide how the sensors, actuators and control computers must be orchestrated to perform as efficiently as possible for the overall engine system. On the one hand, rapid sensor updates are necessary for the control computer to build a proper estimate of the engine state and health. On the other hand, the faster control computer command are updated and sent over the communication network, the better the engine can be managed. Beyond these basic considerations, more subtle trade-offs arise from the fact that not all sensors carry equal value, and that the messages broadcast by some of them may be much more important than others. There is a need to examine the issue of actuator/sensor message scheduling and to figure out what kind of performance trade-offs can be achieved by varying the fraction of time that each actuator and sensor is given to communicate across the network. Deterministic protocols, such as time-triggered architectures, have much better worst-case characteristics than other architectures, and they are much more attractive for certification purposes for that reason. One of the potential beneficial features is to develop the communication system such that its results would be independent from the particular communication protocol used, and would therefore be applicable to a broad range of options. The proposed distributed control policy and communication architecture must be able to guarantee thrust response and stall margin of the engine while considering aspects such as communication bandwidth and overhead, dynamic prioritization and network allocation, availability of computing resources, determinism, delays, and packet dropouts. Additionally, the design should be applicable to a variety of gas turbine engines. Provided that appropriate certification requirements are met, the technology is applicable to both military and civilian markets. Eventually this technology could be applied to a turbine engine in conjunction with a controller consisting of networked smart nodes. 

PHASE I: Assess design tradeoffs in the embedded system communication architecture and protocol. Define a turbine engine control message scheduling scheme. Develop conceptual design of a networked control system taking into consideration the impacts on controller and engine performance. Use modeling and simulation to demonstrate improvement over the SOA. 

PHASE II: Investigate the hardware necessary to realize the controller designed in Phase I. Implement the controller and test it within a controls hardware-in-the-loop setup. Integrate with relevant turbine engine simulations. Based on the results of this experiment, update the controller architecture as necessary. 

PHASE III: Demonstrate the proposed networked control system in a test cell or engine environment. 

REFERENCES: 

1: Distributed Engine Control Working Group (DECWG), "Transition in Gas Turbine Engine Control System Architecture: Modular, Distributed, and Embedded," NASA Propulsion Controls and Diagnostics Workshop, Dec 2009.

2:  Le Ny, J., Feron, E., and Dahleh, M., "Scheduling Kalman Filters in Continuous Time," IEEE Transactions on Automatic Control, Jul 2011.

3:  Nghiem,T., Papas, G.J., Girard, A., and Alur, R., "Time-Triggered Implementations of Dynamic Controllers," EMSOFT '06, Proceedings of the 6th ACM & IEEE International Conference on Embedded Software, Seoul, Korea, Oct 2006.

4:  Alur, R. and Weiss, G., "Interfaces for Control Components," Invited Talk, Workshop on Verifiable Robotics, Computed Aided Verification Conference, Snowbird, UT, Jul 2011.

KEYWORDS: Distributed Control, Time-triggered Architecture, Sensor/actuator Scheduling, Turbine Engine, Communication Bus, Embedded Control Systems 

CONTACT(S): 

Dr. Alireza Behbahani (AFRL/RQTE) 

(937) 255-5637 

alireza.behbahani@us.af.mil 

US Flag An Official Website of the United States Government