No abstract

A mixture method based on exponential curve and ANN is presented according to settlement prediction of roadbed with measured data. Based on this method, the rule of roadbed settlement is classified into sure part and uncertain part. Exponential curve is used to model the sure part, and ANN to model the uncertain part, thus the mixture settlement model can be obtained. Prediction results show that the mixture model has advantages of high precision and small network scale; it provides a new method for settlement prediction of roadbed.

Emerging design concepts such as miniaturisation, modularity, and standardisation, have contributed to the rapid development of small and inexpensive platforms, particularly cubesats. This has been stimulating an upcoming revolution in space design and development, leading satellites into the era of "smaller, faster, and cheaper". However, the current requirement-centric design philosophy, focused on bespoke monolithic systems, along with the associated development and production process does not inherently fit with the innovative modular, standardised, and mass-produced technologies. This paper presents a new categorisation, characterisation, and value-centric design architecture to address this need for both traditional and novel system designs. Based on the categorisation of system configurations, a characterisation of space systems, comprised of duplication, fractionation, and derivation, is proposed to capture the overall system configuration characteristics and promote potential hybrid designs. Complying with the definitions of the system characterisation, mathematical mapping relations between the system characterisation and the system properties are described to establish the mathematical foundation of the proposed value-centric design methodology. To illustrate the methodology, subsystem reliability relationships are therefore analysed to explore potential system configurations in the design space. The results of the applications of system characteristic analysis clearly show that the effects of different configuration characteristics on the system properties can be effectively analysed and evaluated, enabling the optimization of system configurations.

The rapid developments in micro-technologies and the introduction of modularity and standardization into system designs, present significant opportunities for cost reduction in the design and development of satellite systems. However, the high cost of space launch has become a major hindrance to capitalizing on these opportunities. Therefore, seeking appropriate launch opportunities and reducing launch costs might contribute to further growth of the space market. This paper focuses on the analysis of dedicated launch costs factoring in the effect of launch reliability, which in return, can enable the optimization of system designs. Applying a value-centric architecture, system characteristic space is introduced as the design space to define the characteristics of different systems. Based on our launch vehicle database, the launch cost and reliability of different families of launch vehicles are investigated, where the reliability is calculated using a modified two-level Bayesian analysis. The factors of launch cost and reliability are subsequently integrated into the expected launch cost, acting as the objective function for the analysis and optimization process associated with the manufacturing cost of satellites. Through reviewing and redesigning a few classical launch cases, the effectiveness and applicability of the design architecture proposed are validated.