Adaptive Model Based Control of Aircraft Propulsion Systems: Status and Outlook for Naval Aviation Applications

Abstract

The control of military aircraft propulsion and associated aircraft systems continue to become more demanding, in response to the operational needs of new and existing aircraft and missions. High performance aircraft operate in multiple modes. They are complex and require complex propulsion systems that provide precise and repeatable performance: safely, dependably, and cost effectively. To support these requirements, propulsion control systems must manage multiple effectors based on multiple operating parameters through interactive processes. The scopes of control extends beyond the gas turbine engine to the inlet, exhaust, power and bleed extraction, electrical power systems, thermal & environmental management, fuel systems, starting, accessories, and often propellers, rotors or lift fans. Modern propulsion control systems are increasingly integrated with the aircraft flight controls and the distinction is becoming less & less meaningful. Within the gas turbine, variable geometry and active control of turbo-machinery and auxiliary systems proliferate to relax mechanical design constraints and enable designs with increased thrust to weight ratios, reduced fuel burn and increased durability. Digital controls provide crisp and repeatable responses and improve aircraft reliability and availability, but further enhancements are needed as military aircraft become more capable and versatile (e.g., V-22 and F35). The control system must be aware and appropriately respond to component degradation and damage, optimally managing conflicting constraints and goals. Modern propulsion systems are becoming more profoundly multivariable and include multiple effectors to meet multiple goals. They are multivariable because they are cross-coupled, where each effector can affect multiple goals. In addition, these multiple goals, (e.g., performance, life, operating margin) may be conflicting and need to be traded off, and the best trade off will vary with mission. With predictable and rapid increases in computational capability in Full Authority Digital Electronic Controls, the industry is moving forward to address these needs through model based control, control that manages propulsion and aircraft systems with optimal control responses derived from detailed real time models of component behavior. Since the component characteristics change significantly during a service interval, and yet longer time on wing is necessary, these control systems must sense degradation and damage to multiple components and adapt to it. This paper describes current approaches and NAVAIR plans to develop, mature and deploy this technology, while touching on other potential applications.

Aeroengine Multivariable Nonlinear Tracking Control Based on Uncertainty and Disturbance Estimator

“…Traditional aeroengine control design is based on single input single output (SISO) control theory, where it is assumed that various effectors influence only those variables they are meant to control [1], Obviously, control algorithm design on this way becomes more and more difficult to obtain satisfying system response with the increasing input/output variables and coupling effect on modem aeroengine.…”

Toward an intelligent, deterioration accommodating controller for aging turbofan engines

“…Direct thrust control, however, can deal with both of these issues. There is currently significant interest in direct control of thrust using such approaches as model predictive control and modelbased control because of their ability to reduce fuel consumption and EGT, and generally optimise overall engine performance (4)(5)(6) . However, because thrust is not directly measurable in flight, for any of these methodologies to be implemented a reliable thrust estimate is required.…”

“…However, because thrust is not directly measurable in flight, for any of these methodologies to be implemented a reliable thrust estimate is required. This has led to several different approaches to thrust estimation using standard sensor measurements which have been validated in simulation (4)(5)(6)(7)(8)(9) , and progress is being made toward an engine test of model predictive control within the next few years (5) . In this work thrust is controlled directly, with the assumption that a reliable estimate is available.…”

Multiplexed Predictive Control of a Large Commercial Turbofan Engine

“…The availability of fast and cheap real-time processing has motivated work on MPC applied to aircraft engines [20,25]. Onboard processors used in aircraft propulsion control are still not powerful enough to accommodate the execution of the complex algorithm in real time.…”