| Energy optimization driving strategy of high-speed electric trains is a significant factor for sustainable development of a country due to its impact on CO2 emissions, economic growth, and so on.The inter-relationship of railway sub-systems is highly complex, especially with regard to assessing their energy consumption.Many mathematical models and algorithms were developed to optimize the driving strategy, but there are still alot of issues have not been resolved.In this research, new algorithms are proposed to achieve the target of "providing accurateoffline & real-time algorithms with considerable flexibility and robustness to get energy optimization driving strategies for single and multiple high-speed trains".The main works and researches of this dissertation include:1. Research on the key impact factors and mathematical modelsfor accurate calculation of energy consumptionof high-speed train driving. First, we proposed a multiple-particle train model considering the train length. Then, the pantograph-catenaries’relationship simulation and calculation model were improved to get much more accurate simulation results.This paper also studies the impact of phase insulator to the speed and energy consumption, which is an important constraint to the optimal strategy.After that, an optimal research platform and simulation system for high-speed trains was designed and implemented.2. We proposed a Multiple-Group Parallel Genetic Algorithm (MGPGA) to optimize minimal-energy driving strategy for a single train driving. Phase insulator was considered as constrain conditions. New characteristic like regenerative braking mechanism was introduced to save the energy. A variable length chromosome was developed accompanied with the crossover operator taking into account the specific characteristics of high-speed train with gear control. In order to improve the calculation time and speed up converge, adaptive crossover and mutation probability were introduced. Also the migration between sub-groups of the MPGA was designed to achieve efficient computing. Cased study was implemented and comparison shows that the proposed approach can produce ahigher accuracy and percentage of energy saving than previously achieved in other works.3. We proposed a real-time regulation algorithm (RMPGA) of high-speed trains recalculating the energy efficient driving strategy when facing unexpected condition or significant delays arise. The stopping criterion was important for real-time application so the maximum generation evolution was discardingbut fixed time limit was selected. An original Repair Operator was proposed for RMPGA to speed up converge to achieve the best solution within a short time limit. Finally, a simulation study is conducted to verify the efficient tracking driving in a MBS system and real time driving strategy regulation for energy saving in unexpected condition. Compared with typical driving style, the new approach achieves a better result when a train has to get back on schedule aftera delay.4. Optimal control strategy for multiple trains in a successive driving under MBS was studied. Model was developped and formulated to this complicated and large scale problem. To solve this problem, we proposed a hybrid algorithm (Tabu Search-Multi Population Genetic Algorithm,TSMPGA) combining Tabu Search (TS) and MPGA, TS was applied as the mutation operator of MPGA to improve convergence rate and the solution quality. The new TSMPGA not only prevents premature and guarantee convergence to global optimum. A Pseudo-Crossover Operator was designed according to the characteristics of multiple trains and MBS system.Case study was implemented and the result reveals that the proposed algorithm was effective and efficient.5. AMinimal-energy driving platform was designed and implemented for high-speed trains with simulation and optimization modules, which can be applied to high-speed trains’real-time optimization. |
PDF Full Text Request