Leiming Geng, Ruihua Zhang and Weihua Liu
It is an indispensable part of airworthiness certification to evaluate the fuel tank flammability exposure time for transport aircraft. There are many factors and complex coupling…
Abstract
Purpose
It is an indispensable part of airworthiness certification to evaluate the fuel tank flammability exposure time for transport aircraft. There are many factors and complex coupling relationships affecting the fuel tank flammability exposure time. The current work not only lacks a comprehensive analysis of these factors but also lacks the significance of each factor, the interaction relationship and the prediction method of flammability exposure time. The lack of research in these aspects seriously restricts the smooth development of the airworthiness forensics work of domestic large aircraft. This paper aims to clarify the internal relationship between user input parameters and predict the flammability exposure time of fuel tanks for transport aircraft.
Design/methodology/approach
Based on the requirements of airworthiness certification for large aircraft, an in-depth analysis of the Monte Carlo flammability evaluation source procedures specified in China Civil Aviation Regulation/FAR25 airworthiness regulations was made, the internal relationship between factors affecting the fuel tank flammability exposure time was clarified and the significant effects and interactions of input parameters in the Monte Carlo evaluation model were studied using the response surface method. And the BP artificial neural network training samples with high significance factors were used to establish the prediction model of flammability exposure time.
Findings
The input parameters in the Monte Carlo program directly or indirectly affect the fuel tank flammability exposure time by means of the influence on the flammability limit or fuel temperature. Among the factors affecting flammability exposure time, the cruising Mach number, balance temperature difference and maximum range are the most significant, and they are all positively correlated with flammability exposure time. Although there are interactions among all factors, the degree of influence on flammability exposure time is not the same. The interaction between maximum range and equilibrium temperature difference is more significant than other factors. The prediction model of flammability exposure time based on multifactor interaction and BP neural network has good accuracy and can be applied to the prediction of fuel tank flammability exposure time.
Originality/value
The flammability exposure time prediction model was established based on multifactor interaction and BP neural network. The limited test results were combined with intelligent algorithm to achieve rapid prediction, which saved the test cost and time.
Details
Keywords
Ruihua Zhang, Leiming Geng and Weihua Liu
To reduce the flammability exposure assessment time and meet the requirements of airworthiness regulations of transport aircraft, inerting system has become the standard…
Abstract
Purpose
To reduce the flammability exposure assessment time and meet the requirements of airworthiness regulations of transport aircraft, inerting system has become the standard configuration of modern civil aircraft. Therefore, airworthiness regulations put forward definite quantitative index requirements for the safety of inerting system, and to obtain the quantitative data of the safety of inerting system, it is necessary to solve the calculation method. As one of the quantitative/qualitative evaluation techniques for system safety, fault tree analysis is recognized by international airworthiness organizations and national airworthiness certification agencies. When fault tree analysis technology is applied to quantitative analysis of the safety of inerted system, there are still some problems, such as heavy margin of constructing fault tree, great difficulty, high requirement for analysts and poor accuracy of solving when there are too many minimum cut sets. However, based on tens of thousands of flight simulation tests, Monte Carlo random number generation method can solve this problem.
Design/methodology/approach
In this paper, the fault tree of airborne inerting system is established, and the top event is airborne inerting system losing air separation function. Monte Carlo method based on random number generation is used to carry out system security analysis. The reliability of this method is verified.
Findings
The static fault tree analysis method based on Monte Carlo random number generation can not only solve the problem of quantitative analysis of inerting system, but can also avoid the defects of complicated solution and inaccurate solution caused by the large number of minimum cut sets, and its calculation results have good reliability.
Practical implications
The research results of this paper can be used as supporting evidence for airworthiness compliance of airborne inerting system.
Originality/value
The research results of this paper can provide practical guidance for the current civil airworthiness certification work.