The Genetic-Evaluation-Based Family Tree Model And Its Application On Aerospace Products

Since the reliability of aerospace products is related to many factors such as the cost, life and operation of aerospace products in space, etc, the technological development of the reliability of aerospace products will affect the economic construction and the strategic security of the entire country, it will have a significant impact on the development of a state in such respects as political, economic, military, science and technology, etc. Compared to the general industrial products, aerospace products have some characters of complex systems, harsh environment, high cost and high risk. A little failure occurred in the use of the products could lead to disastrous consequences, and the loss is quite heavy. Since the reliability of aerospace products is related to many factors such as the cost, life and operation of aerospace products in space, etc, the technological development of the reliability of aerospace products will affect the economic construction and the strategic security of the entire country, it will have a significant impact on the development of a state in such respects as political, economic, military, science and technology, etc. Therefore, in order to achieve the goals of high reliability, long life and good performance, we must be stricter, more rigorous, and more scientific with the techniques used in aerospace products, and take advantage of the appropriate measures to deal with the potential and existing failures. We should make sure to be "better, faster, cheaper" in the premise of good control of risk.The existing software for reliability analysis and evaluation commonly used some traditional reliability analysis techniques in aerospace research such as fault tree analysis (FTA), event tree analysis (ETA) and failure mode effects and criticality analysis (FMECA), etc. Traditional reliability analysis methods were usually based on the product's failure or end-state characteristics of life. These methods analyzed some experimental data by calculation in the latter part of product design to analyze and evaluate the reliability of products to prevent future trouble after "remedial measures". These methods are mainly used to analyze and evaluate the historical data and other data in use for existing products qualitatively and quantitatively, in order to take corresponding appropriate preventive and remedial measures, and to prevent the occurrence of failure or danger. Apparently, these methods with some risks can't guarantee the probability of failure will be reduced to a minimum before the design of products. In addition, the limitation of traditional methods of reliability is:these methods in general are intended to analyze the corresponding cause of the failures and other subsequent events after the failure appeared. And most of them can only use part of test data to analyze and calculate during the process of products' using, and a lot of valuable historical data of space can not be used effectively. Therefore, before the design and production of a new product implementation, the problem about how to make full use of the existing historical data to establish product database, and try to utilization the historical data and field data related to the process of assess will be a very critical and urgent issue. We can take both environmental and genetic factors into account to design quality products with higher reliability.The subject on the basis of mastering research results on the reliability at home and abroad integrated the theoretical knowledge and practical application closely, and tried to propose a novel reference program of reliability analysis- the family tree model of aerospace products. The model will analyze and compare the key components of a certain series of space products vertically to make full use of more historical experimental data to provide accurate reference data for the reliability design of aerospace products to reduce the cost and cycle of development. The actual life and reliability of aerospace products are affected not only by the factors on the design and manufacture, also by external stimuli of the physical environment to a large extent. Similar to the biological phenomenon, the phenotype value of quantitative traits will be influenced both by environment and gene. So the subject introduced methods of estimation of breeding value according to the theory of genetic evaluation.The main idea of the family tree model for aerospace products in this subject was to establish a family tree model for these components, which had genetic relations, especially for some key components, and apply the concept of the breeding value in genetic evaluation to family tree model of aerospace products. Then estimated heritability values of individual property, and calculated the individual's breeding value on the property. The properties were including but not limited to the cost, life, designers for individual products, and other factors. Then added breeding value into individual properties and sorted them. And at last we can determine the one which have a greater genetic influence on future generations in a certain properties, and chose the outstanding individual as the central reference object for the product design of next generation in the family tree, by combining the phenotype value of property. So, in the design of aerospace products, we can use more known historical data and breeding value to provide policy options for the design of new unit components' performance in the future.The calculation of the family tree model in this paper taken the data of American communications satellite-Echostar as an example, and analyze the expected life of the product. Firstly establish a family tree and assign the heritability 0.9, calculate the path coefficient between the individual breeding value and its phenotypic values. Then calculate regression coefficient according to partial regression coefficient equation and find out the estimated breeding value of all individuals' life expectancy in the family tree. Finally, we can select excellent parent product for the design optimization of next-generation product by integrating the estimated breeding values and phenotypic values of expectance life.The innovation of this paper was to apply the breeding value in genetic evaluation to the subject of improving the knowledge information base of the family tree model for aerospace products. The core cross point was that the phenotypic value is affected by both the internal genes and the external environment factors in the reliability of aerospace products which is the same as in the biological populations.This paper introduced the background, the significance and three traditional assessment methods of reliability. Followed by the introduction of the theory knowledge of family tree, this paper applied the methods of evaluation of breeding value from genetic evaluation to the family tree model. Finally, the paper focused on the family tree model for aerospace products, and taken communications satellites of Echostar series for example, estimated and analyzed the breeding values of life expectancy for the products in the family tree by using the methods of evaluating the breeding value.This model showed conducive to improve the reliability of aerospace products and the data accuracy, provides a theoretical reference for the later reliability optimization design, and minimizes the failure rate of aerospace products in a theoretical sense in order to save the development costs, shorten the development cycle and reduce the risk.