Prediction and analysis of global launch vehicle success rates based on grey system models

Purpose The purpose of this paper is to reveal the bottleneck of reliability growth for carrier rockets by establishing different grey system models to simulate and predict the success rate of rocket launches worldwide. And to raise awareness of the importance of enhancing rocket launch reliability. Design/methodology/approach An overview of the development history of launch vehicles and their reliability challenges was presented at first. Then, a briefly introductions to the launch mission assurance process of the US National Security Space Launch (NSSL) program as well as the reliability management system of China’s Long March 7 launch vehicle was given. Based on global, the US and Chinese launch success rate data from 2018 to 2024, four grey prediction models of the EGM, ODGM, EDGM and DGM were established, yielding high-precision simulation and forecasting results. The study reveals that the improvement of launch vehicle reliability is facing significant bottleneck constraints. Finally, an innovative approach to address the existing challenges in launch vehicle reliability analysis and evaluation is proposed in this paper. Findings The simulation and prediction outcome showed that the global launch success rate had a slow upward trend. The reliability growth of launch vehicles is severely constrained by bottlenecks. Big data technologies, along with uncertainty system analysis methods based on diverse perspectives – such as probability statistics, fuzzy mathematics, grey system theory and rough set theory – as well as innovations in sequence operators, spectral analysis and intelligent algorithms, have laid a solid foundation for effectively integrating complex uncertain data and breaking through the bottlenecks of reliability modeling. The conditions are increasingly ripe for exploring new approaches and methodologies in launch vehicle reliability analysis and evaluation by comprehensively leveraging big data technologies, multiple uncertainty system analysis methods, sequence operators, spectral analysis and intelligent algorithms. Research limitations/implications The limitation of this research is that it revealed the bottleneck of reliability growth for carrier rockets and proposed a novel approach to overcome the bottleneck constraints in launch vehicle reliability analysis and evaluation, which integrates big data technologies with diverse uncertainty system analysis methods (including probability statistics, fuzzy mathematics, grey system theory and rough set theory) as well as complex uncertain data fusion techniques such as sequence operators, spectral analysis and intelligent algorithms. However, the specific methodological approaches to break through the bottleneck constraints in launch vehicle reliability growth remain to be further investigated. Practical implications Manufacturers use reliability growth tests to iteratively improve launch vehicle reliability and performance to predetermined levels through cycles such as “exposing defects—analysing causes—improving designs, processes, or operations.” However, when the data does not meet the modeling conditions of traditional reliability growth models, people often adopt some “flexible” approach, such as using simulated data or borrowing relevant data from similar equipment to “piece together” data, which may bury hidden dangers in launch vehicle reliability. There is an urgent need to explore new models and methods. The novel idea proposed in this paper has the potential to significantly improve the quality and reliability of launch vehicle in smart manufacturing. Originality/value This paper proposes a novel approach to overcome the bottleneck constraints in launch vehicle reliability analysis and evaluation, which integrates big data technologies with diverse uncertainty system analysis methods (including probability statistics, fuzzy mathematics, grey system theory and rough set theory) as well as complex uncertain data fusion techniques such as sequence operators, spectral analysis and intelligent algorithms.