Introduction

The space plasma and radiation form a hazardous environment to Earth orbiting spacecraft. Spacecraft problems are regularly experienced that in extreme cases lead to a failure or loss of the spacecraft. The anomalous behaviour of the spacecraft can sometimes be explained by some technological problem or the human factor. In other cases it is the space environment that causes the anomaly. When the space environment is responsible for a spacecraft anomaly it is important to both be able to make a post analysis of the situation that lead to the anomaly, and to make predictions of coming anomalies. The post analysis is made in order to understand the cause of the anomaly. The predictions are useful to spacecraft operators to be prepared to take the necessary action.

The space environment is determined by the current space weather which ultimately is driven by the Sun. As the understanding of the Sun, the solar wind, and the interaction with the Earth is far from complete a number of observed and derived parameters are used to describe the system. By analysing the correlation between different parameters it is possible to partly understand the system. A crucial point here is that the parameters should be available in long uninterrupted time series. From statistical studies it is then possible to find basic correlations that can be used to create simple models of the system. The models can describe general trends and overall situation that lead to anomalies. However, in a more detailed analysis the situation quickly becomes more complicated and other approaches must be considered. One approach is to model the system with artificial intelligence methods, such as neural networks or fuzzy systems, as they are general nonlinear methods. The free parameters of the neural networks are found from training on the observed data, and the network becomes a non-linear model of the system that can be used for analysis and prediction.

The success for empirical models, such as neural networks, depends heavily on the data that are used to define the model. The data base should contain all the relevant parameters, i.e. one needs a physical understanding of the system, and it should contain a large enough number of examples so that the model estimation becomes statistically sound. The data base should contain solar wind data, magnetospheric data, local plasma environment data (when available), and the time of anomalies. The crucial parameter here is the anomaly data as it is often not available other than to the spacecraft operators due to the sensitive industrial information.

The current project will primarily use ESA spacecraft anomaly data. To further extend the anomaly data base and to better satisfy the potential user needs spacecraft operators are welcome to collaborate on these matters. For this purpose there is a questionnaire that any potential user of SAAPS may reply to. It must be pointed out that details of the anomaly data will never be made available to the users of the system other than in term of statistical information. Furthermore, the name of spacecraft and users will never be disclosed. The concept of a satellite anomaly index will also be developed for spacecraft operators use.

This proposal is a contract extension to the Study of Plasma and Energetic Electron Environment Effects (SPEE) (ESTEC/Contract No. 11974/96/NL/JG/(SC)). The aim of this contract extension is to develop an on-line service for spacecraft operators and space environment  analysts which will include:

The data base together with the analysis tools will be an operational software that can be easily updated and operated with new anomaly data and interrogated by external users. The software is called SAAPS for Satellite Anomaly Analysis and Prediction System.
 
 
Index Next