非线性预测控制器的嵌入式设计与实现

发布时间：2018-01-09 04:05

本文关键词：非线性预测控制器的嵌入式设计与实现　出处：《上海交通大学》2012年硕士论文　论文类型：学位论文

【摘要】：预测控制是一类利用被控对象的模型来预测对象未来输出的优化控制算法,已在工业领域得到了广泛应用,尤其对线性系统来说是一种有效的控制策略。然而,许多工业过程存在着一定程度的非线性因素,其中包括非线性被控对象和非线性约束,如果采用线性模型预测控制,难以得到令人满意的控制效果。因此,非线性模型预测(Nonlinear Predictive Control,NMPC)控制近年来受到了广泛的关注。随着社会的发展和科技的进步,各种基于嵌入式平台的控制器应运而生。FPGA是近年来发展起来的一类可编程逻辑器件,它吸收并结合了传统软件和硬件实现方案的优点,既保证了系统的高性能,又改善了系统的可设计性和适应性。因其具有灵活性和较短的设计周期,近年来开始成为实现MPC控制的一种有效硬件平台。正因为具有上述优势,本文采用FPGA来设计非线性预测控制器。本论文主要完成以下工作: 1)嵌入式非线性预测控制器的算法设计。分析NMPC优化问题及现有求解算法的特点,提出适合在嵌入式平台上应用的可行算法。即使受到计算资源有限的制约,该算法仍然可以保证在每个采样间隔内都求得较优的可行解。 2)嵌入式非线性预测控制器的软硬件设计与实现。选择合适的嵌入式系统作为实现平台,通过分析控制流程的各个步骤和优化算法不同部分在嵌入式系统实现所需的时间和对应资源代价,进行合理的软硬件分工,尽可能利用FPGA提供的硬件资源和并行计算能力,实现嵌入式控制器。 3)嵌入式非线性预测控制器的仿真研究。通过和MATLAB闭环测试验证该控制器的实时性和有效性。 4)嵌入式预测控制器的现场调试支持软件设计。针对模型预测控制算法在现场应用中难以找到最优参数和难以更新控制器数据的问题,设计了现场调试支持软件。该软件支持多种预测控制器,包括控制参数设置、控制效果仿真以及数据传输下载三大功能。
[Abstract]:Predictive control is a kind of optimal control algorithm for predicting the future output of objects by using the model of controlled objects , which has been widely used in the field of industry , especially for linear systems . However , many industrial processes have some nonlinear factors , including nonlinear controlled objects and nonlinear constraints . With the development of society and the progress of science and technology , various controllers based on embedded platform are born . The FPGA is a kind of programmable logic device developed in recent years . It absorbs and combines the advantages of traditional software and hardware implementation scheme . It has guaranteed the high performance of the system and improved the designability and adaptability of the system . Because of its flexibility and short design cycle , it has become an effective hardware platform for MPC control in recent years . Because of the above advantages , the FPGA is used to design the nonlinear predictive controller . This thesis mainly accomplishes the following tasks : 1 ) The algorithm design of the embedded nonlinear predictive controller . The optimization problem of the NMPC and the characteristics of the existing solution algorithm are analyzed , and a feasible algorithm suitable for the application in the embedded platform is put forward . Even if the computational resources are limited , the algorithm can guarantee a better feasible solution in every sampling interval . 2 ) The hardware and software design and implementation of the embedded nonlinear predictive controller . Select the appropriate embedded system as the implementation platform . By analyzing the different steps of the control flow and the different parts of the optimization algorithm , the necessary time and corresponding resource cost are realized in the embedded system . The hardware resources and the parallel computing power provided by the FPGA are utilized as much as possible to realize the embedded controller . 3 ) Simulation research of embedded nonlinear predictive controller . The real - time performance and effectiveness of the controller are verified through closed loop test with MATLAB . 4 ) On - site debugging support software design of embedded predictive controller . On - site debugging support software is designed aiming at the difficulty of finding the optimal parameters and difficult updating of the controller data in the field application of the model predictive control algorithm . The software supports a variety of predictive controllers , including the control parameter setting , the control effect simulation and the three functions of data transmission download .

【学位授予单位】：上海交通大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：TP368.1;TP273

【相似文献】