高效高功率因数三相电源控制策略研究

发布时间：2018-05-12 15:13

  本文选题：三相电源 + 复合有源箝位　； 参考：《西安理工大学》2016年博士论文

【摘要】：电力电子变换器是电能利用的重要转换装置,三相电源是将三相工频交流电变为直流电的电力电子装置,传统的三相电源采用二极管或晶闸管整流,会引起谐波污染、无功损耗增大、电磁干扰等负面效应。基于软开关技术的PWM有源功率因数校正变换器具有高功率因数和高效率,可以解决传统三相电源存在的问题,因此,在大功率不间断电源、无功补偿、太阳能发电以及交直流传动系统等领域取代传统装置,有着广阔的应用前景。本文结合单晶炉加热电源的技术要求,采用先进的拓扑结构,将功率因数校正,软开关技术与先进控制算法有机地结合,设计了一套高效率、高功率因数、高性能的三相电源样机,并针对该对象的控制策略进行了深入研究,主要研究内容如下(1)在分析现有拓扑的基础上,选择了前级复合有源箝位软开关三相功率因数校正变换器(CACZVS三相PFC变换器)配合后级移相全桥零电压零电流(FB-ZVZCS)DC/DC变换器的两级结构的三相电源拓扑。针对单晶炉用加热电源对于高功率因数和高效率的要求,本文在对比分析了不同方案的拓扑结构,软开关条件及相关调制方法的基础上,选择一种新型复合有源箝位软开关三相功率因数校正变换器(CACZVS三相PFC变换器)实现功率因数校正和输出直流母线电压控制,后级采用移相全桥零电压零电流(FB-ZVZCS)DC/DC变换器实现输出电压和功率的控制。本文分析了三相电源的工作机理、等效电路,在此基础上建立了相应数学模型,为后续控制方法研究提供了条件。(2)针对变换器控制参数多、相互耦合、难以选择的问题,提出多目标混沌粒子群优化算法设计控制器参数以提高变换器性能。本文以前级CACZVS三相PFC变换器基于两相旋转坐标系下的传统前馈解耦的比例积分(PI)控制器参数优化问题为例,给出了系统的参数优化方法和步骤。三相PFC变换器的控制涉及交轴电流、直轴电流和输出电压三个控制闭环,有3个PI控制器的6个控制参数需要设计,控制参数之间存在相互耦合,优化目标需综合考虑输出直流电压快速性及准确性,输入单位功率因数等因素,这是一个复杂的多参数(多维)多目标优化问题,针对这一问题,本文提出(1)基于派瑞托(Pareto)理论的粒子群优化算法,解决多个待优化目标的自动均衡问题。(2)单向耦合映像格子的时空混沌模型产生多维初始粒子群方法,解决了传统单维混沌粒子群方法的问题,提高了粒子群方法的搜索速度,降低陷入局部最优解的概率。(3)设计了三种适合的非线性控制方法,以降低电源参数不确定(或变化)对系统性能的影响,提升了三相电源的控制性能。前级CACZVS三相PFC变换器具有非线性、强耦合、参数不确定和负载等参数随工况时变等特征,本文提出以下控制策略用来提升变换器控制性能,具体方法包括:a)基于负载滑模观测器的反馈线性化控制方法。该方法针对不确定负载采用滑模观测器来在线估计负载参数,同时,针对三相PWM变换器的非线性强耦合特性,设计了对应的反馈线性化控制器,将估计所得的负载值应用到电压环反馈线性化控制器当中,以提升控制器对于负载时变的适应能力;b)考虑负载及输出滤波电容不确定,设计了针对电压环负载和输出滤波电容未知的自适应控制器,电流环采用反馈线性化控制器,该控制方案参数少且易于选取,对负载变化和电容参数不确定具有更好的适应性;c)提出了考虑输入电压波动,输入滤波电感、滤波电感等效电阻、输出滤波电容以及负载参数不确定的鲁棒滑模变结构控制方法。与传统反馈线性化方法相比,本文方法通过增加鲁棒项,提高了系统在参数不确定及参数变化时的鲁棒性,同时边界层渐缩方法的采用减小了控制量的颤振,使输入三相电流变化更为平滑,负载变化时输出波动更小。(4)提出鲁棒定频模型预测控制(RCF-MPC)方法实现前级CACZVS三相PFC变换器的电流环定频模型预测控制,电压环采用鲁棒变结构控制器,在同时存在参数不确定(变化)和输入三相电压不平衡的情况下,提升了前级变换器的性能。单晶炉的实际工况通常是输入三相电压不平衡与参数不确定(变化)并存,传统处理三相不平衡的方法需要分别考虑三相正序电流和负序电流的控制,控制闭环增多,控制结构复杂,本文采用预测控制解决这个问题。应用于传统硬开关模式的三相功率因数校正变换器(三相VSR变换器)的有限状态模型预测控制策略(FS-MPC)由于存在不定频的问题,无法被直接应用于一个开关周期内需要开关状态至少变化一次的软开关变换器。针对上述问题,本文提出鲁棒定频模型预测控制(RCF-MPC)方法,该方法采用时间序列最优定频预测控制框架,在价值函数中增加负序电流指标的同时增加处理预测模型不确定性的鲁棒项,电压环仍然采用鲁棒变结构控制器,在同时存在三相电网不平衡和变换器参数不确定(变化)这样更接近于实际的工况下时,改善了变换器的性能。(5)提出采用论域减缩的模糊控制方法实现后级FB-ZVZCS变换器控制,在未知后级变换器模型的情况下,得到了比传统PI控制方法更好的控制性能。后级FB-ZVZCS变换器传统的控制方法基于小信号线性化模型和线性系统理论设计控制器,然而小信号模型的准确性依赖于工作点,当负载等参数随工作状态变化时,传统方法控制性能变差。本文设计了变论域模糊控制方法提高变换器性能,该方法不依赖于对象精确模型,同时论域收缩因子的使用,解决了模糊控制效果受限于模糊变量和模糊规则数量,响应速度和控制精度无法兼顾的问题,在保证相应快速性的前提下,提高了后级变换器的稳态跟踪精度。(6)设计了基于DSP TMS320F28335控制电路,制作了前级和后级变换器,完成了 1.2kW小功率测试样机,为实验验证控制方法提供了实验平台,同时为加热电源的实用化奠定了基础。在实验平台上对两级变换器进行了调试,得到了期望的控制效果。
[Abstract]:Power electronic converter is an important conversion device for electric energy utilization. The three-phase power is a power electronic device that turns the three-phase power frequency alternating current into DC. The traditional three-phase power supply uses diode or thyristor rectifier, which will cause the harmonic pollution, the increase of reactive power and the negative effects of electromagnetic interference. The active power of PWM based on soft switching technology The factor correction converter has high power factor and high efficiency, which can solve the problems of the traditional three phase power supply. Therefore, it has a wide application foreground in the fields of high-power uninterruptible power supply, reactive power compensation, solar power generation and AC and DC transmission systems. With advanced topology, the power factor correction, soft switching technology and advanced control algorithm are organically combined, a set of high efficiency, high power factor and high performance three-phase power supply prototype is designed, and the control strategy of this object is studied deeply. The main research internal capacity is as follows (1) on the basis of the analysis of the existing topology and the selection of the former The three-phase power factor correction converter (CACZVS three-phase PFC converter) with the active clamped soft switching (CACZVS) is matched with the three-phase power topology of the post stage phase shift full bridge zero voltage zero current (FB-ZVZCS) DC/DC converter. In this paper, the requirements for high power factor and high efficiency for the heating power supply for single crystal furnace are compared and analyzed in this paper. On the basis of the topology, soft switching conditions and related modulation methods, a new type of complex active clamp soft switching three-phase power factor correction converter (CACZVS three-phase PFC converter) is selected to realize power factor correction and output DC bus voltage control. The post stage phase full bridge zero voltage zero current (FB-ZVZCS) DC/DC converter is used. The current output voltage and power control. This paper analyzes the working mechanism of the three-phase power supply and the equivalent circuit. On this basis, the corresponding mathematical model is established, which provides the conditions for the study of the following control methods. (2) the multi target chaotic particle swarm optimization algorithm is proposed to design and control the multi-objective chaotic particle swarm optimization algorithm. The former stage CACZVS three-phase PFC converter, based on the traditional feedforward decoupling proportional integral (PI) controller parameter optimization problem in the two phase rotating coordinate system, gives the parameter optimization method and steps of the system. The control system of the three-phase PFC converter involves the cross axis current, the direct axis current and the output voltage of three. A control closed loop, the 6 control parameters of 3 PI controllers need to be designed and the control parameters are coupled with each other. The optimization target needs to consider the speed and accuracy of the output DC voltage and the input unit power factor. This is a complex multi parameter (multidimensional) multi-objective optimization problem. In this paper, the problem is proposed (1). Based on Parry (Pareto) theory, the particle swarm optimization (PSO) algorithm is used to solve the automatic equilibrium problem of multiple targets. (2) a multi-dimensional initial particle swarm optimization method is produced in the spatio-temporal chaos model of the unidirectional coupled map lattice, which solves the problem of the traditional single dimensional chaotic particle swarm optimization, raises the search speed of the particle swarm optimization method and reduces the most local maximum. The probability of optimal solution. (3) three kinds of nonlinear control methods are designed to reduce the influence of the power parameter uncertainty (or change) on the system performance and improve the control performance of the three-phase power supply. The pre stage CACZVS three-phase PFC converter has the characteristics of nonlinear, strong coupling, parameter uncertainty and load and other parameters with the working condition. The lower control strategy is used to improve the performance of the converter control. The specific method includes: a) feedback linearization control method based on the load sliding mode observer. This method is used to estimate the load parameters online by using the sliding mode observer for the uncertain load. At the same time, the corresponding feedback linearization is designed for the non linear strong coupling characteristics of the three-phase PWM converter. The controller applies the estimated load value to the voltage loop feedback linearization controller to improve the adaptive capacity of the controller for the load time. B) designs an adaptive controller for the voltage loop load and the unknown output filter capacitor considering the uncertainty of the load and output filter capacitance. The current loop uses feedback linearization control. The control scheme has little parameters and is easy to select. It has better adaptability for load change and uncertainty of capacitance parameters. C) proposed a sliding mode variable structure control method considering input voltage fluctuation, input filter inductance, equivalent resistance of filter inductor, output filter capacitance and uncertainty of load parameters. Compared with the method, the robustness of the system is improved by increasing the robust term, while the system is robust when parameters are uncertain and parameters change. At the same time, the use of the boundary layer shrinkage method reduces the chatter of the control quantity, makes the input three-phase current more smooth, and the output fluctuates less when the load changes. (4) the robust constant frequency model predictive control (RCF-MPC) side is proposed. The method realizes the constant frequency model predictive control of the current loop of the pre stage CACZVS three-phase PFC converter, and the voltage loop adopts the robust variable structure controller. The performance of the pre stage converter is improved when the parameters are not determined (change) and the input three-phase voltage is unbalance. The actual working condition of the single crystal furnace is usually the input three-phase voltage imbalance and the reference. The number of uncertainties (changes) coexist, and the traditional methods of three-phase unbalance need to consider the control of the positive sequence current and negative sequence current respectively. The control closed loop is increased and the control structure is complex. This problem is solved by the predictive control in this paper. The three phase power factor correction converter (three phase VSR converter) used in the traditional hard switching mode is used in this paper. The limited state model predictive control strategy (FS-MPC) can not be directly applied to a soft switching converter which needs to change at least once in a switching period due to the existence of uncertain frequency. In this paper, a robust fixed frequency predictive control (RCF-MPC) method is proposed in this paper. This method uses the time series optimal constant frequency prediction. The control framework increases the negative sequence current index in the value function while increasing the robustness of the uncertainty of the prediction model. The voltage loop still adopts the robust variable structure controller, which improves the performance of the converter when there is a three phase grid imbalance and the converter parameter uncertainty (change) which is more close to the actual condition. 5) the fuzzy control method of domain reduction is proposed to control the post stage FB-ZVZCS converter. Under the condition of the unknown post stage converter model, the control performance is better than the traditional PI control method. The traditional control method of the post stage FB-ZVZCS converter is based on the small signal linearization model and the linear system theory design controller, however, the control method is based on the linear model and the linear system theory. The accuracy of the small signal model depends on the working point. When the parameters of the load change with the working state, the traditional method control performance is worse. In this paper, a variable domain fuzzy control method is designed to improve the performance of the converter. This method does not depend on the exact model of the object, and the use of the domain shrinkage factor, the effect of the fuzzy control is limited to the model. The number of fuzzy and fuzzy rules, the response speed and the control precision can not be taken into account. Under the premise of ensuring the corresponding speediness, the steady-state tracking precision of the post stage converter is improved. (6) the DSP TMS320F28335 control circuit is designed, the former and the post stage converters are made, and the 1.2kW small power test prototype is completed, and the experimental verification control is carried out. The method provides the experimental platform and lays the foundation for the utility of the heating power supply. The two stage converter is debugged on the experimental platform, and the desired control effect is obtained.

【学位授予单位】：西安理工大学
【学位级别】：博士
【学位授予年份】：2016
【分类号】：TN86
，

本文编号：1879098