Damage forecasting based on multi-factor fuzzy time series and cloud model

Timely and effective fault forecasting has great significance to guarantee the security of an aircraft, in view of the characteristics of harsh work environment of a flight control system. Based on the forecasting results, we can prevent damages or benefit from the forecasting activities. Fuzzy time series (FTS) forecast which provides a powerful and useful framework to deal with imprecision or ambiguity problems has been widely used in computer science. Many FTS-based forecasting models have been proposed in recent years, and thus the main problems are how to determine the useful interval length and the appropriate window basis size. In this paper, a new method based on multi-factor FTS and a cloud model was presented to predict the trend of aircraft control surface damage (ACSD). The proposed method constructs multi-factor fuzzy logical relationships based on the historical data of ACSD. To handle the uncertainty and vagueness of the ACSD historical data more appropriately, the cloud model is applied to partition the universe of discourse and to build membership functions. Furthermore, a variation forecasting method improved by the cloud model was proposed to compute the forecasting results. The experimental results prove the feasibility and its high forecasting accuracy of the proposed method.