To the question of forecasting the technical condition of low-thrust liquid rocket engines

In the rapidly developing space and rocket industry, spacecrafts are being equipped with low-thrust liquid rocket engines. High requirements are imposed on the reliability, efficiency and economy of fuel use for this type of rocket engine. To ensure monitoring of the characteristics of spacecrafts, a functional diagnostic system is used, which includes telemetry and analytical data processing. Telemetry performs the functions of receiving and transmitting information. Information processing is carried out in computer centers located on the spacecraft and the Earth. The most promising computing tool capable of predicting time series and classifying a large amount of interconnected data is considered an artificial neural network. In this regard, the subject of research in the work is data processing methods based on an artificial neural network. The purpose of the work is to develop a method for forecasting the technical condition of low-thrust liquid rocket engines using an artificial neural network.

The relevance of research on the use of a neural network in the system of functional diagnostics of low-thrust liquid rocket engines for spacecraft is explained in the introduction. In the main part, an analysis of many telemetric data of the rocket engine is carried out and their strength in the forecast of the main diagnostic parameters is determined. It is proposed to use traction, specific impulse, and temperature of the structure as diagnostic parameters. The prognostic capabilities of the neural network were investigated and a schematic diagram of a method for predicting the technical condition of a low-thrust liquid rocket engine was developed. In the developed method, at the first stage, the neural network performs the approximation of the function and extrapolates the time series of telemetric data; the second stage determines the probable class of the technical condition of the engine.

The conclusion outlines a plan for further experimental research in the study area and provides recommendations on the development and improvement of algorithms for functioning of artificial neural networks as part of the functional diagnostics system of the spacecraft. Due to the generalized nature of the methodological schemes, the results of the work can be applied to any type of rocket engines and used at all enterprises of the rocket and space industry of the corresponding profile.