Learn while flying! China is developing launch vehicles that will learn from them

 Learn while flying! China is developing launch vehicles that will learn from them

In the history of human space, the mission of space launch has high risk. To improve the safety and reliability of launch vehicles has been the goal of Aerospace Science and technology workers. In this regard, the rocket flight control system plays a crucial role.

According to statistics, from 1990 to 2000, about 40% of the failure cases of launch vehicles in Europe, the United States, Japan and Russia may be remedied by using advanced navigation guidance and control technology, so as to continue to complete the mission or complete the mission degraded. In recent years, falcon-9 and delta-4 launch vehicles have experienced engine thrust drop in flight, but the implementation of remedial measures to varying degrees has not affected the main mission.

Picture source: visual China

The reporter of science and technology daily learned from Beijing Institute of Aerospace Automatic control that China is developing launch vehicles for learning. Yu Chunmei, deputy director of the Institute, said that the technology is to introduce intelligent technology into various tasks such as navigation, guidance and control, so as to make the launch vehicle more intelligent and autonomous, and have stronger active adaptability to complex environment and emergency situations, so as to ensure the completion of the mission to a greater extent.

Learning while flying and lifelong learning

For a rocket, the flight control system is its brain. According to the research results of the scientific research team of Beijing Institute of Aerospace Automatic Control, this brain has two major characteristics in learning.

Yu Chunmei said that the flight strategy and orbit of the traditional rocket were designed in advance. If the rocket encounters a complex space environment that is not expected in advance, the flight trajectory of the rocket cannot be adjusted online. In addition, if the rocket has a sudden situation, such as power system failure, and can not take timely measures, it is likely to lead to mission failure.

Another characteristic of learning launch vehicles is lifelong learning. The control system can make full use of the data generated in the whole life cycle, and realize the functions of model intelligent correction, model intelligent establishment, scheme and parameter intelligent optimization based on big data and intelligent analysis technology, so as to continuously carry out self-learning and improvement.

Most of the active launch vehicles are disposable. Each rocket can transmit its learning content back to the ground during its flight, which can be used for reference by later comers, so as to realize the transitive learning between rockets and make subsequent rockets more and more smart and more experienced.

On July 23, 2020, the long March-5 remote-4 launch vehicle was ignited and lifted off from Wenchang space launch site to carry out Chinas first Mars exploration mission (tianwen-1 mission). Photo by reporter Cai Yang of Xinhua News Agency

The reporter learned that the above two learning characteristics complement and promote each other. The accumulated experience and data of learning while flying support lifelong learning; the training and optimization of lifelong learning and the self evolution of algorithm can promote the intelligence of learning while flying.

Rocket learning has a threshold

The control system should be designed once for life

The prospect of learning carrier rocket is bright, but to realize the ability to learn, the rocket must have enough ability.

First of all, higher requirements are put forward for the control system to have the ability of distributed heterogeneous cross core high-speed information exchange and multi heterogeneous memory sharing scheduling management. Yu Chunmei said. In short, rockets have to have brains that are smart enough.

It is inevitable that there are errors in the structure installation of launch vehicles. Elastic vibration, liquid sloshing, unknown environmental disturbance and other factors will affect the control system. Therefore, the rocket is required to have real-time perception of its own health status and flight status. For example, typical non fatal faults of rocket power system occur from time to time, so it is necessary to have the ability of rapid and accurate identification for the main engine thrust drop in the ascending phase, the normally open, normally closed and polarity of the attitude control nozzle in the high altitude flight phase.

Based on the results of on-line identification, perception and self-assessment, in order to realize the functions of online target change, trajectory planning and control online reconstruction, the rocket must have strong adaptability to the needs of environment mutation, uncertainty of its own structural parameters, typical power system failure and mission change, so as to ensure the flight stability under complex flight conditions and fault conditions, and be able to enter the semi autonomous region Long axis maximum elliptical orbit or safe berthing orbit.

Up to now, the long march series launch vehicles have carried out 349 missions, and a large amount of data have been accumulated in each mission. These data become the knowledge base of rocket after overall planning and standardized management. They can learn from the launch vehicle to extract information, acquire experience and knowledge, and combine it with intelligent control technology to realize self-learning ability, so as to continuously grow.

In addition, if the rocket wants to achieve lifelong learning, it also puts forward the requirement of one design, life-long applicability for the control system, which requires the ability to cover the whole life cycle of the model with one design.

For the learning launch vehicle, the intelligent control system architecture is the foundation, the intelligent control algorithm based on the whole life cycle data is the core, and the strong computing power based on the intelligent algorithm is the carrier.

China has made some progress

Strive to have the ability to learn in 2025

China has made some progress in learning carrier rocket related technology exploration and application. The reporter learned that at present, the team of Beijing Institute of Aerospace Automatic control has started to improve Chinas in-service launch vehicles and popularize their learning ability.

At 0:57 on October 12, 2020, China successfully launched the gf-13 satellite with the long march 3B carrier rocket at Xichang Satellite Launch Center, and the satellite successfully entered the scheduled orbit. Xinhua News Agency (Photo by Guo Wenbin)

On July 9 and October 12, 2020, China launched the long march 3B carrier rocket from Xichang Satellite Launch Center. In addition to sending the satellite into the predetermined orbit, the research team of Beijing Institute of Aerospace Automatic control also carried the typical power fault identification and guidance and control reconstruction technology on the rocket, and successfully completed the flight test verification.

For the development of learning launch vehicles, the research team has formulated a two-step plan: to have strong adaptability in 2020, which means that the rocket will have stronger adaptability under some fault conditions, and can continue to fly; and strive to have the learning ability by 2025, that is, to make the main launch vehicles in service in China initially have the learning ability.

In addition, the concept of learning will also be integrated into the research and development of launch vehicles in China.

Source: Qiao JunJing, editor in charge of science and Technology Daily_ NBJ11279