Font Size: a A A

Modeling And Optimization Of The Half-bridge Slr Converter

Posted on:2008-07-24Degree:MasterType:Thesis
Country:ChinaCandidate:X C LiuFull Text:PDF
GTID:2208360215986037Subject:Pattern Recognition and Intelligent Systems
Abstract/Summary:PDF Full Text Request
Recently, the research on the non-linear modeling and dynamicsin the soft-switching circuit system has become the hotspot on thepower electronic field. Because they are not only helpful to reveal thereason of the system's irregular phenomenon, and ensure the stabilityof the system's operating performance, but also urge to produce thecorresponding critical point control method and enhance the system'sdynamic response ability. However, the discrete-map modeling methodbased on fixed period is limited. This paper mainly contains threeaspects, stability analysis, modeling with a new method and parameteroptimization of half-bridge SLR converter.The resonance switching system is a non-linear and discretesystem. It is difficult to establish its mathematical model and gain thesystem's transient and steady state solution theoretically. We can't usethe classical control theory to design controller and study its stability.Proceeding from practical problems that is difficult to design resonanceinductance and capacitance values, and on the foundation of analyzingthe working mode and modes of half-bridge SLR converter, this paperdeduces relational expression between output voltage and circuitparameters in the discontinuous current mode, and obtains the stabilityintervals of resonant parameters, then validates the correctness withsimulink. The studied and simulation results are consistent.Support vector machine is an important part of the statisticallearning theory, and displays an extremely good performance studied inthe small sample situation. Support vector machine regression is a newalgorithm, and has a very good promoted characteristic and non-linearmodeling ability. Therefore, it can solve the control problem ofnon-linear, big time-lapse as well as the not easy modeling system. Thispaper introduces support vector machine theory and regressionalgorithm, and then establishes an input-output black box model ofhalf-bridge SLR converter with the data sample picked from formersimulation based on SVMR algorithm for the first time. Modelingprocess is simple, and the equivalent circuit model is concise, and it is convenient for circuit analysis and the control system design in theactual project. The SVM test results validate that the modeling basedon SVMR is accurate and feasible.Finally the paper introduces the optimized concepts and methodsof the switching power supply, and then optimized the half-bridge SLRconverter parameter. The results of the optimization are that the outputvoltage ripple is very small, and satisfy the requests of the engineeringdesign.
Keywords/Search Tags:half-bridge, series load resonant converter, support vector machine regression, simulation and modeling, parameter optimization
PDF Full Text Request
Related items