Liquid rocket engine test real-time data diagnosis strategy

Abstract

The invention provides a liquid rocket engine test data diagnosis strategy which has good universality and feasibility. A hierarchical test data real-time interpretation method can be provided for testers, more test data information can be integrated, a quantified fault mode can be identified, a coping strategy can be pushed, and the safety of human and property in the test process can be guaranteed. The method comprises the following steps: S1, establishing a quantitative fault mode standardization library; S2, executing a sensor and converter failure detection logic; S3, executing an engineparameter anomaly detection logic; and S4, executing an engine fault mode detection logic.

Description

technical field [0001] The invention is mainly used in the field of liquid rocket engine test real-time data diagnosis technology. Background technique [0002] The real-time data diagnosis of liquid rocket engine tests has always been valued. It is an essential means to protect engine products, test benches, and test personnel. At the same time, it is also an important basis for evaluating engine performance, analyzing fault causes, and fault location. At present, the following problems exist in the monitoring link of liquid rocket engine test data in my country: (1) The real-time performance is not strong: because the engine test failure is characterized by strong explosion power, rapid occurrence, and large expansion area, high real-time diagnosis of its data Strategy is an essential part of the experiment. However, the existing data diagnosis strategies mostly focus on the research of machine learning based on data-driven neural networks or model-based reasoning and diagn...

AI Technical Summary

Problems solved by technology

At present, the following problems exist in the monitoring link of liquid rocket engine test data in our country: (1) The real-time performance is not strong: because the characteristics of the engine test failure are large explosion power, rapid occurrence, and large expansion area, high real-time diagnosis of its data is implemented Strategy is an essential part of the experiment. However, the existing data diagnosis strategies mostly focus on the research of machine learning based on data-driven neural networks or model-based reasoning and diagnosis algorithms. Although the diagnosis accuracy has been greatly improved, the real-time The performance is greatly reduced. Although relevant research has been carried out since the 1990s, it still cannot be applied to the actual test bed; (2) Insufficient diagnostic coverage: the emergency shutdown data diagnosis strategy of the existing test bed is relatively single, and each type The engine judges the threshold of key parameters according to the empirical thresholds of different parameters such as speed and pressure, and then implements the control strategy. In terms of information adoption, there is a problem of insufficient coverage, which leads to a greatly reduced implementation effect of the diagnosis strategy.

Images