Neural Network Interpolation Parameters of a Multi-Mode Dynamic Model of the Aircraft Engine

Abstract

The foundations of the concept of creation of intelligent aircraft engine control systems based on the decomposition of control processes within the architecture of open information systems are considered. Unlike well-known approaches, the suggested approach allows achieving the management goal based on the principle of minimum entropy by redistributing system resources in conditions of their shortage, as well as adapting system characteristics when changing the management situation based on self-learning and self-organization of intelligent control systems. Based on an analysis of the development trends of aircraft engines, as well as development trends of production and technological systems, including the creation of new composite materials and new technologies for the manufacture and control of parts and components of aircraft engines, the intellectualization of their automatic control systems is discussed. Moreover, the development trends of aircraft engine control systems are considered from the development of their structures, functions, properties, and abilities for new qualitative changes. The article gives the general characteristics and the main directions of the design of intelligent control systems for aircraft engines as complex technical objects. The problem of designing nonlinear dynamic models of aircraft engines using artificial neural networks is discussed. The statement of this problem and possible approaches to its solution are being formed. The results of the neural network identification of an aircraft engine are compared using the least-squares method. Such a technique for designing a model of aircraft engines makes it possible to indirectly calculate engine coordinates inaccessible to measurement - traction, fuel consumption, etc. The suggested approach allows calculation of the design of neural networks simulating aircraft engines at each step using standard procedures, which makes it possible to automate the creation of neural networks. To reduce the computation time, it is suggested using the optimization algorithms taking into account changes in the state entropy. This simplifies the implementation of the neural network model of an aircraft engine in real time as part of an onboard computer complex.

Method for QOE monitoring and increasing in cellular networks based on QOE-to-QOS mapping using spline approximation

“…This approach works in the following way: First, the control system data are gathered, where the provider has some service traffic, as the QoE metrics of service. Next, the machine training algorithm is used to learn the dependency function between these rates [19]. The given method fits as the provider for quality support, as the infiniteness of network service providers allows them to analyze different anomalies at furnished services.…”