数字控制DC-DC变换器的非线性建模与单输入模糊PID控制算法研究

Abstract: Compared with traditional analog power supply, digitally controlled DC-DC converters have the characteristics of easy realization of complex algorithms, fine programmability, flexible configuration, and well robustness. What's more, they can effectively integrate with SoC by using the mainstream CMOS manufacturing process, which can help to reduce the cost and minimize the volume. Therefore, this new method is now applied in different areas, besides it also has some problems such as limit cycling oscillation and control problems under complex load changes. In this work, we will analyze the mechanism of quantization effects and delay effects of digitally controlled DC-DC converters, taking parasitic parameters into account, and then establish a complete and accurate nonlinear theory model of the system, to reveal the nature of the limit cycling oscillation and to give out the judgments of stability, while to optimize parameters by using geometric programming algorithm, then we can take effective measurements to improve its stability and reliability. Based on above analysis, we give out a new single input PID fuzzy control algorithm through combining their individual advantages of linear and nonlinear control, so that the power system can meet dynamic and static performance requirements in input voltage changes and complex load conditions. Meanwhile, using the signed distance method to simplify the structure and increase real-time controlling. Finally, verifying the correctness of the proposed theory and methods by MPW tapeout.

与传统的模拟电源相比，数字控制DC-DC变换器由于具备易实现复杂算法，配制灵活，鲁棒性好，能够与SoC系统进行无缝整合集成，更有助于降低成本和缩小体积。因此得到越来越多的应用，但由于存在非线性效应引起极限环振荡以及复杂负载变化下的控制问题。本课题将通过深入分析数字控制DC-DC变换器的量化效应与延迟效应产生机理，考虑寄生参数，建立完整精确的系统非线性理论模型，揭示极限环振荡产生的本质，并给出稳定性判据，同时采用几何规划算法进行参数优化，以便采用有效措施提高其稳定性与可靠性。在此基础上给出新的单输入模糊PID控制算法，结合线性与非线性控制各自的优点，使电源系统在输入电压变化以及复杂负载变化下满足动、静态性能要求。同时采用符号距离法，有效简化结构的同时提高控制实时性。最后通过MPW流片来验证所提出的理论与方法的正确性。

与传统的模拟电源相比，数字控制DC-DC变换器由于具备易实现复杂算法、可编程性好、配制灵活、鲁棒性好等特点，以及采用主流CMOS制造工艺，能够有效的与SoC系统进行无缝整合集成，更有助于降低成本和缩小体积。因此可应用在更多的不同领域，但也存在极限环振荡与复杂负载变化下的控制问题。本课题通过深入分析数字控制DC-DC变换器的非线性效应、量化效应与延迟效应产生机理，考虑寄生参数效应，建立了完整精确的系统非线性理论模型，揭示了极限环振荡产生的本质，并给出稳定性判据，同时采用几何规划算法进行参数优化，以便采用有效措施提高其稳定性与可靠性。在此基础上给出优化的单输入模糊PID控制算法，采用线性与非线性控制相结合的控制方法，同时加入符号距离法，有效简化结构的同时提高了控制实时性，使电源系统在输入电压变化的复杂负载变化下满足动、静态性能要求。最后通过FPGA来验证所提出的理论与方法的正确性，为最终成为IP核集成于SoC中打下坚实的基础。同时,开展面向LED绿色照明的反激式开关电源的控制策略研究与控制芯片优化设计，取得了一系列令人满意的研究成果。