Seismic Reliability Optimization Of Urban Gas Supply Network System

Lifelines, such as the power transmission, transportation, and the urban water,heat and gas supply pipeline systems and the like, are fundamental infrastructuresystems to support modern cities' functions and maintain regional economicdevelopment. As a needful component of lifelines, the urban gas supply pipelinesystem plays an important role in industry and daily life. The earthquake-induceddestroy would directly influence the system's function and incur severe secondarydisasters, for example, fire. Thus, to ensure the safety of urban gas supply systemsubject to earthquake, it is of great necessity to study its seismic reliabilities.The earthquake resistance of urban gas supply system can be investigated interms of the system reliability. In general, both system reliability analysis andoptimization are included in this research. System reliability analysis is afundamental tool for its optimization design, while the latter is the ultimate object tothe former. This paper investigates the optimization of urban gas supply systembased on its corresponding reliability analysis.The optimization of gas supply system is loosely defined as the selection of thelowest cost combination of potential pipelines with some design constraints such asthe lowest nodal connectivity reliabilities of network satisfied. In fact, this problemis a typical 0-1 combinatorial optimization one which belongs to "NP-Hard"problem. Therefore, the meta-heuristics instead of some traditional optimizationalgorithms are employed to tackle such optimization problem.In this thesis, four Benchmark instances of the above optimization problem,which can be used to verify the meta-heuristics applied to gas supply system, arefirst established. Then, the ant colony optimization algorithm, a new combinatorialoptimization algorithm inspired by the foraging behavior of natural ants, isdeveloped successfully to the seismic optimization problem of gas supply network.Two ant optimization models, i.e. the ant system and the Max-Min ant system, arefurther studied as well by comparison in a small scale network case. Furthermore, a hybrid algorithm combining the ant colony optimization algorithm with thesimulated annealing algorithm is proposed. Also, this thesis investigates the searchperformance of several meta-heuristics such as the genetic algorithm, simulatedannealing, ant colony optimization algorithm and the hybrid algorithm. Cases studyshow that for medium and small-sized network, the ant colony optimizationalgorithm performs as well as the genetic algorithm, while the simulated annealing isthe worst among all algorithms. Finally, the seismic topology optimization ofShenyang main gas supply network is performed using the ant colony optimizationalgorithm.