【作者】 沈正华；

【导师】 杨春节；

【作者基本信息】 浙江大学 ， 控制理论与控制工程， 2011， 硕士

【摘要】 造纸的蒸煮过程将造纸原料变成纸浆,在很大程度上决定了纸张的质量。但同时,蒸煮过程也具有高能耗、高物耗、高污染的“三高”特点。在大力发展资源节约型、环境友好型社会的今天,如何实现蒸煮过程的节能、减排、降耗成为制浆造纸工业的重要课题。本文立足于节能降耗的观点,研究间歇蒸煮过程的参数优化问题,希望能达到保质、高产、节能、降耗的目的。本文的创新工作及贡献包括：(1)在分析间歇蒸煮过程机理的基础上,建立了关于蒸煮过程的质量、产量和能耗模型。首先,利用支持向量机(SVM)原理,根据造纸工厂的日常数据,建立了关于纸浆kappa值的质量模型。仿真数据表明,该模型准确能地描述蒸煮参数与kappa值之间的数学关系。接着,根据Hatton的经验模型,并利用现场数据回归,建立了有关纸浆得率的产量模型。然后,分析蒸煮过程的能量消耗情况,建立了关于间歇蒸煮过程的蒸汽能耗模型。(2)本文根据优化模型的特点,提出了一种基于改进遗传算法的间歇蒸煮优化方法,能在满足纸浆质量要求的前提下降低蒸煮过程的能耗和物耗。改进的遗传算法在基本遗传算法的基础上,融入小生境技术和自适应交叉技术,从而增强了整个算法的局部寻优和全局搜索能力。用日常生产数据进行的仿真实验表明,该方法优化效果明显,与现场实际情况相比,节约蒸汽3.52%,节约白液用量17.18%,硫化度减少0.37%。(3)为了实现真正意义上的多目标优化,本文提出了第二种优化方法—一基于改进的NSGA-II的间歇蒸煮参数多目标优化方法。该种方法的优化模型在前者的基础上,多加入纸浆得率目标,实现在保证纸浆质量的同时,提高纸浆产量,降低蒸煮过程的能耗和物耗。改进的NSGA-II算法在NSGA-II的基础之上,引入一种适应度评价机制,采用自适应交叉变异算子,并加入了外来群体迁入机制,使得该算法相比NSGA-II,能够更好地保证解的收敛性和分布性。将改进的算法应用于间歇蒸煮多目标优化问题,优化效果明显,而且提供多个最优解,具有指导性。更多还原

【Abstract】 Pulp cooking process turns papermaking feedstock into pulp and determines paper’s quality to a great extent. Meanwhile, Pulp cooking process is characteristic of high energy consumption, high materials consumption and high pollution. Today, we are vigorously developing a resource-conserving and environment-friendly society, it is an important task for paper and pulp industry about how to realize saving energy, lowering consumption and reducing pollutants discharge. Considering the viewpoint of energy efficiency, optimization of batch digester parameters was researched. The innovation and contribution of this thesis are as follows:(1) The mechanism of batch digester was analyzed, and the related batch digester parameter models were established, including quality model, yield model and energy consumption model. Firstly, according to the production data of one papermaking mill, a SVM model about kappa number was developed, which accurately reflect the relationship between kappa number and cooking parameters. Secondly, based on Hatton’s empirical model, a model about pulp yield was developed. Thirdly, after analyzing batch digester’s steam consumption, a model about batch digester’s energy consumption was developed.(2) Because it is difficult to use SVM model in conventional optimization approach, an improved Genetic Algorithm using niching technology and adaptive crossover technology was presented, which realizes energy conservation and consumption reduction on the premise of guaranteeing the pulp’s quality. On the basis of Standard Genetic Algorithm, the improved Genetic Algorithm integrates niching technology and adaptive crossover technology to ensure solution’s convergence and distribution, thereby enhance its ability of local and global search. This method was used on the optimization of one papermaking mill, the simulation results showed good performance. It saved 3.52% of steam,17.18% of white liquid dosage and 0.37% of sulfidity, comparing to real production data.(3) To realize really multi-objective optimization, an improved NSGA-II algorithm for optimization of batch digester parameters method was presented. On the basis of former optimization model, this method adds in one more optimization goal, which is improving the pulp yield. So the purpose of the optimization method is to realize energy conservation and consumption reduction, improve pulp yield on the premise of guaranteeing the pulp’s quality. As to the improved NSGA-II algorithm, a fitness assessment method which combines non-dominated front and crowding distance was introduced to describe the individual’s quality. In order to generate better individuals, an adaptive crossover and mutation mechanism introduced by Srinivas was employed. Moreover, a foreign group immigration strategy was adopted to avoid falling into local Pareto optimal solution. Simulation results show the improved NSGA-II algorithm can better ensure the convergence and diversity of solutions, compared with NSGA-II. The improved NSGA-II algorithm was used on multi-objective optimization of batch digester parameters, and it received obvious good performance.更多还原