【作者】 张生；

【导师】 姜万录；

【作者基本信息】 燕山大学 ， 机械电子工程， 2014， 博士

【摘要】 冷轧带钢生产在我国国民经济中占有重要的地位。目前，虽然我国钢铁产量巨大，但是整体生产水平偏低，尤其是冷轧板带生产的核心技术多掌握在国外公司手中。因此，建造拥有自主知识产权的现代化冷连轧生产线，生产出高质量的冷轧带钢，增强产品在市场上的竞争力是国内科研人员的共同目标。轧制规程设定是冷连轧工艺的核心技术之一，建立更合理的轧制规程是提高冷轧带钢质量的有效途径之一，在这方面我国与国际先进水平尚有一定的差距，必须进行更全面、更深入的研究，才能赶超国际先进水平。本文结合某厂1450mm全连续五机架冷连轧机工艺优化计算机系统研发的工程实践，对冷连轧机工艺优化计算机系统进行了基于量子进化算法和支持向量机的轧制规程优化研究与实践。轧制工艺数学模型是进行轧制规程优化的基础，针对不同的工艺优化目标可以建立不同的目标函数及约束条件。本文对量子旋转门角度更新策略进行了研究，通过改进量子旋转门角度更新查询表提升了量子进化算法的收敛性能；以能耗最低为目标，应用改进的量子进化算法对轧制规程进行优化，使轧制总功率降低3%以上。将其应用于某厂五机架冷连轧机组的规程计算中，自2011年3月投产以来，节约了大量的能源，创造了巨大的经济效益。在传统量子进化算法中，应用查询表对量子旋转门角度进行更新时必须针对具体问题具体设计查询表，通用性较差。为了克服这一缺点，引入粒子群算法和微分进化算法，通过其在角度空间进行启发式搜索的方式进行量子旋转门角度更新，提出了两种混合量子进化算法。通过标准函数测试表明，混合量子进化算法增强了量子进化算法的全局收敛性能，提高了通用性，将其应用于轧制规程优化，结果表明，应用该方法达到了均衡轧制力和轧制功率的目的，在工程应用中具有很高的实用价值。在轧制规程优化过程中，处理多个优化目标的常用办法是将多个目标加权聚合为单个目标进行优化。为了避免权重赋值的人为因素，对应用多目标进化算法实现轧制规程优化进行了探讨。为了提高算法的执行效率，将量子计算以及混沌与多目标进化算法融合，提出了量子混沌多目标进化算法。标准函数测试表明，量子混沌多目标进化算法的收敛速度比NSGA-II高将近30%，最后将之应用于轧制规程优化，获得了合理的结果，为其在冷连轧机轧制规程优化中的应用提供了理论依据，多目标进化算法是规程优化的未来发展方向。现代化轧机中安装了数以百计的高精度传感器，工艺优化计算机系统的数据库中存储了海量的设备状态数据和轧制过程数据。对数据挖掘工具——支持向量机在轧制力预报中的应用进行了研究和实践。首先整理数据库中的海量数据，建立支持向量机进行轧制力偏差预报的样本库。然后利用样本库中样本对支持向量机进行训练并预报轧制力偏差。最后据此对模型计算所得的轧制力设定值进行修正,冷连轧机的轧制力预报精度提高到5%以内，是轧制规程优化的有效手段。在某厂全连续五机架冷连轧机工艺优化计算机系统开发实践中，基于WinCC组态软件和Microsoft SQL Server2005数据库平台，运用WinCC组件和ANSI-C和VBScript脚本语言，完成了轧制规程计算与优化程序，开发了完善的人机界面监视系统，实现了轧制过程数据和设备状态数据的归档、查询等功能。更多还原

【Abstract】 Cold-rolled sheets played an important role in the national economy. Nowadays,Chinese steel production is very huge, but the level of production is lower, especially thecore of cold rolled steel strip production technology almost completely in the hands offoreign companies. Hence, modern cold rolling production line with independentintellectual property rights is built, and the cold-rolled sheets of high of high quality areproduced so as to enhance the competitiveness in the world market, which is a commontarget of domestic researchers. Rolling schedule calculation is one of the core technologiesof tandem cold rolling process, and a reasonable rolling schedule is an effective way ofimproving the quality of the cold-rolled sheets. However, in this respect, there exist greatgaps between Chinese enterprises and advanced international levels; in order to catch upwith the advanced technology in the world, the Chinese reseachers need to make moreintensive study. In this study, relying on engineering practice of process optimizationcomputer system of five stand tandem cold rolling mills in a cold rolled-sheets plant, theoptimization study and practice of the rolling schedule based on quantum evolutionaryalgorithm (QEA) and support vector machine (SVM) is applied to process optimizationcomputer system.The mathematical models of rolling process are foundation of rolling scheduleoptimization. Aiming at the different technology objectives, different objective functionsand constraints are chosen. The objective functions are optimized by the improvedquantum evolutionary algorithm. Firstly, the lookup table of the quantum rotation gate isanalyzed, which is the most important influence factor of performance of the quantumevolutionary algorithm. The performance is improved by modifying the lookup table.Then, based on the minimize energy consumption objective function, the rolling scheduleis optimized by the improved quantum evolutionary algorithm. The total power reducedmore than3%. This method has already stable operation for two years in a plant, a largenumber of energy is saved and a lot of economic benefits are created.In the traditional quantum evolutionary algorithm, the rotation angle of quantum rotation gate is updated by the lookup table, but the lookup table needs to be designedaccording to the specific questions, therefore its commonality is poor. For overcoming thisshortage, the particle swarm optimization (PSO) and differential evolutionary algorithm(DEA) are introduced to update the rotation angle of the quantum rotation gate by themean of heuristic search in the angle space, and the hybrid quantum evolutionaryalgorithm (HQEA) is constructed. The test results of standard functions show that thehybrid quantum evolutionary algorithm can strengthen the global convergenceperformance of the quantum evolutionary algorithm and improve its commonality. TheHQEA is applied to optimize the rolling schedule and the rolling force and power arebalanced. Its practical value In the engineering application is proved.In the process of schedule optimization, the common method of dealing with multipleobjectives optimization is that a weight is given to every objective, and then they areaggregated into one single objective function. For eliminating the human impact of theweight assignment, it is researched that rolling schedule is optimized with multi-objectiveevolutionary algorithm (MOEA). To improve the execution efficiency of multi-objectiveevolutionary algorithm, the quantum computation and chaos are introduced to the MOEA,and quantum chaotic multi-objective evolutionary algorithm (QCMOEA) is proposed.Some standard functions testing show that the efficiency of QCMOEA is30%higher thanNSGA-II. Finally, the QCMOEA is applied to schedule optimization of tandem coldrolling mill, and the reasonable schedules are obtained. There is a theoretical foundationfor MOEA applying to schedule optimization, which is future direction.At present, hundreds of high accuracy sensors are installed on the modern rolling mill.Therefore, a mass data about equipment status and rolling process are stored in thedatabase of the process optimization computer system. Application of data mining in therolling schedule optimization is discussed and practiced, and support vector machine waschosen as data mining tool to predict the rolling force. Firstly, the mass data in thedatabase is preprocessed and a sample set for rolling force prediction based on supportvector machine is built. Secondly, the support vector machine is trained by the samplesand predicted the rolling force deviations between setting value and measured value.Finally, the result is applied to adjust the setting value of the rolling force, and the accuracy of the rolling force prediction is improved to less than5%. It is an effectivemethod for schedule optimization.The development of process optimization computer system of tandem cold rollingmill in a sheet plant is practiced based on the WinCC configuration tool and MicrosoftSQL Server2005database tool. The WinCC and programming languages such as ANSI-Cand VBScript are used for building human machine interface (HMI) monitoring system,accomplishing rolling schedule calculation and optimization, realizing the function offiling and query rolling process data and equipment status data, and improving the millmanagement function.更多还原