Rohit K. Belapurkar

Design of Distributed Engine Control Systems for Stability Under Communication Packet Dropouts

In this paper, we address the issue of the stability of distributed engine control systems under communication constraints and, in particular, for packet dropouts. We propose a control design procedure, labeled decentralized distributedfull-authority digitalenginecontrol andbasedonatwo-leveldecentralized controlframework. Weshow that the packet dropping margin, which is a measure of stability robustness under packet dropouts, is largely dependent on the closed-loop controller structure and that, in particular, a block-diagonal structure is more desirable. Thus, we design a controller in a decentralized framework to improve the packet dropping margin. The effectofdifferentmathematicalpartitioningonthepacketdroppingmarginisstudied.Theproposedmethodologyis applied to an F100 gas turbine engine model, which clearly demonstrates the usefulness of decentralization in improving the stability of distributed control under packet dropouts.

Stability Analysis of Distributed Engine Control Systems Under Communication Packet Drop

Abstract : Currently, Full Authority Digital Engine Control (FADEC), based on a centralized architecture framework is being widely used for gas turbine engine control. However, current FADEC is not able to meet the increased burden imposed by the advanced intelligent propulsion system concepts. This has necessitated development of the Distributed Engine Control system (DEC). FADEC based on Distributed Control Systems (DCS) offers modularity, improved control systems prognostics and fault tolerance along with reducing the impact of hardware obsolescence. Some of the challenges to be dealt with are selection of communication architecture, high temperature electronics and logical functional partitioning of centralized controller.

Stability of Fiber Optic Networked Decentralized Distributed Engine Control under Time Delays

The importance of distributed architecture for turbine engine control is well discussed in literature. Distributed turbine engine control architecture enables the use of new performance optimization methods along with achieving weight reduction. Communication constraints like time delays and packet dropouts can limit the performance of distributed engine control. In this paper, we propose a controller which will stabilize the time delay system based on delay independent condition. The controller under decentralized framework is also studied for stability under both time delays and packet dropouts. The proposed Decentralized Distributed Full Authority Digital Engine Control (D 2 FADEC) is implemented using Fiber Optics and is validated using a gas turbine engine model.

Stability Analysis of ARINC 825-based Partially Distributed Aircraft Engine Control with Transmission Delays and Packet Dropouts

The importance of distributed aircraft engine control architecture is well discussed in literature. The performance of distributed engine control systems is predominantly dependent on the performance of the communication network. Communication constraints such as time delays and packet dropouts can lead to inconsistent data latency, which may limit the performance of distributed engine control or in the worst case scenario, can even destabilize the control system. In this paper, we review three different partially distributed engine control architectures based on ARINC 825, a general standardization of Controller Area Network (CAN) for aerospace applications. The first architecture employs the legacy sensors & legacy actuators and uses a data concentrator to transmit the data to distributed FADEC. Second architecture, which has smart sensors and legacy actuators, makes use of state observers to provide compensation for sensor-controller time delays. In the third architecture, smart actuators allows the use of network predictive controller, which provides compensation for time delays and packet dropouts in both sensor-controller and controlleractuator communication channels. Stability conditions for each of the three architectures are presented and the three architectures are compared on the basis of performance and robustness to time delays and packet dropouts.

LQR Control Design of Discrete-Time Networked Cascade Control Systems With Time Delay

Series cascade control systems, in which, the output of one process drives a second process are studied extensively in literature. Traditional control design methods based on transfer function approach are used for design of cascade control systems with disturbances in inner loop and time delays in outer loop process. Design of current turboshaft engine control systems are based on cascade control system framework. Next generation aircraft engine control systems are based on distributed architecture, in which, communication constraints like time delays can degrade control system performance. Stability of networked cascade control systems for turboshaft engines in a state space framework is analyzed in the presence of time delays. Two architectures of networked cascade control systems are presented. Stability conditions for discrete-time cascade control systems are presented for each of the architecture with time delays which are more than the sampling time.© 2011 ASME

Design of Set-point Controller for Partially Distributed Turboshaft Engine with Network Faults

Future aircraft engine control systems will be based on a distributed architecture, in which, the sensors and actuators will be connected to the controller through an engine area network. Distributed engine control architecture enables the use of advanced control techniques along with achieving weight reduction, improvement in performance and lower life cycle cost. The performance of distributed engine control system is predominantly dependent on the performance of communication network. Due to serial data transmission, time delays are introduced between the sensor/actuator nodes and distributed controller. Although these random transmission time delays are less than the sampling time, they may degrade the performance of the control system. Network faults may result in data dropouts, which may cause the time delays to exceed the sampling time. Design of current turboshaft engine set-point controller is based on cascade control system framework, with an inner loop controlling the gas generator and outer loop to regulate the power turbine speed. This paper analyses the stability of set-point controller for partially distributed turboshaft engine control systems with time delays. A partially distributed turboshaft engine control system is designed based on ARINC 825. Transmission delays are estimated for the proposed design and it is shown that the bandwidth utilization of ARINC 825 is within limits. Stability of networked cascade control systems for turboshaft engines in a state space framework is analyzed in the presence of time delays.

Study of Model-based Fault Detection of Distributed Aircraft Engine Control Systems with Transmission Delays

The next generation aircraft engine control systems will be based on a distributed architecture, in which, the sensors and actuators will be connected to the controller through an engine area network. The advantages and challenges for implementing distributed engine control system are well discussed in public literature. Addition of an engine area network will introduce additional issues like network faults, network-induced delays and data loss. Although the communication protocol of the selected engine area network will have inherent fault diagnostic techniques to limit the data loss and time delay, these communication constraints will affect the performance of distributed engine control systems. Inflight fault diagnostics, onboard health management, fault-tolerant control and model-based control techniques are well studied for the current engine control systems. These diagnostic and control methodologies use an onboard engine model and depend on the accuracy of the tracking filter to estimate the measured and unmeasured parameters. The tracking filter is also used to update the engine model to match the actual engine characteristics. This paper studies the effect of network-induced time delays on model-based fault diagnostics. Design methodology for a parity relation based residual generator is presented. The proposed residual generator captures the time delay information and hence is more accurate than the conventional residual generator for detecting faults under a distributed framework. This paper also highlights the importance of including time delay information for model-based fault diagnostics and control.