Design And Optimization For The Main Permanent Magnet Of MRI

Many practical engineering problems often require optimizing parameters or certain technical indexes after basic design finished in order to reach special target or to economize cost. The thesis is devoted to get the best combination of the parameters to heighten homogeneity of magnetic flux density in the target region to ensure that the design of permanent magnet for Magnetic Resonance Imaging (MRI) equipment could satisfy the demands of the least volume of permanent magnet and high magnetic flux density. The result can make contribution for heightening the performance of MRI equipment and promoting its popularization in medical field.A method of permanent magnet circuit design is introduced. Validity of the method is validated by the design of main permanent magnet of MRI with 0.35T. Main dimension of electromagnetic equipment can be designed in virtue of magnetic circuit design which provides advantaged approach for predigesting the design process of the practical electromagnetic equipment.In order to find out significant variables referred to optimization design and adjustment direction of each variable, Orthogonal Design (OD) is applied. Simultaneously, combined with Excellent Sequence Method (ESM), the optimization process for muti-variable is achieved and a method for reducing the domain of variables gradually which can enhance the computational efficiency is introduced. Response Surface Method is used to make an appropriate Response Surface Model (RSM) of optimal object and optimal variables. Focus analysis on the method of determining shape parameter in Radial Basis Function (RBF) is useful to improve the precision of RSM. Based on the traditional Genetic Algorithm (GA), a compositive improved (GA) is proposed and the performances of different improved GA are discussed to validate the superiority of the improved GA. Based on all the four methods above mentioned, a combined method for optimization design is proposed. The main idea of the optimization method is to find out significant factors and adjustment direction of each factor based on the OD; meanwhile, original optimization can be carried through ESM; then, RSM can be constructed with the best shape parameter; thus, the best result can be found by the improved GA.The validity of the proposed optimal method is first validated by test functions. And then, the optimization method is successfully applied to the optimal design of electromagnetic equipment. A sort of structure for main magnet of MRI with C shape which meets the requirement of high magnetic flux density, high homogeneity of magnetic flux density and small volume of permanent magnet is obtained. The computation and measure results of test function and practical engineering problem both show that the proposed optimization method is applicable to optimize the problem of multi-variable and to solve large scale magnetic field inverse problems. Furthermore, the method which is able to search the best result in the whole region is prone to programming and can greatly reduce the computational time based on a certain precision.