编辑: 人间点评 2019-08-31
―1― 博士学位论文公示材料 学生姓名 赵健 学号

1110353 二级学科 系统工程 导师姓名 刘士新 论文题目 燃煤机组动力配煤及多机组联合调度优化模型与算法 论文研究方向 系统工程,调度优化,智能优化算法 论文关键词 调度优化;

动力配煤;

经济调度;

机组组合;

数学规划;

启发式算法;

布谷鸟 算法 论文摘要(中文) 随着经济的快速增长和全社会用电量的逐年攀升,电力工业已成为支撑国家经济发展的重要行业.

由于我国的能源具有富煤、贫油和少气的特点,因此在未来较长的一段时期内,燃煤机组仍是我国的 主力发电机组.面对日益严峻的能源形势和环境保护压力,促进燃煤发电企业的节能减排,将是我国保 持经济可持续发展的必然要求之一. 本文以电力生产企业的节能减排为研究背景,以燃煤机组的安全经济运行、多机组联合调度优化 为研究目的,对动力配煤问题、经济调度问题和机组组合问题进行了深入的研究.研究成果体现在以 下四个方面: (1)对布谷鸟算法的变体和应用进行了归类和总结,重点研究了布谷鸟算法的参数设置.在六个 基准测试函数上进行了参数灵敏度分析,实验结果表明布谷鸟算法对自身参数值敏感,通过正交实验 给出了一组较好的参数组合. (2)燃煤机组的动力配煤通过科学的混配单煤,解决了锅炉设计时的煤质与实际供应煤质不匹配 的问题,提高了热效率,降低了发电成本, .在分析现有单煤和混煤煤质关系的基础上,建立了基于配 煤成本最低且单煤资源受限的 0-1 混合整数规划模型.为了进一步降低配煤成本并保证混煤在理论上 不结渣,允许混煤部分煤质指标超标,并将超标的煤质作为惩罚项加入到原目标函数,得到了新的配 煤模型.实验结果表明新模型在以较小的代价即牺牲部分混煤煤质指标的前提下,可获得单价更低的 配煤方案. (3)经典电力经济调度问题具有非凸、不连续和非线性的特点,不易求解,本文提出了一种基于 相邻学习的布谷鸟算法可高效的求解该问题.首先,定义了该问题的解的结构.然后,采用自适应步 长因子提高算法的全局探测能力,采用相邻学习策略提高算法局部搜索能力,并使用改进的λ迭代方 法提高新解的生成速度.数值实验在含有

6、

13、20 和40 机组的基准测试系统上进行,实验结果表明 改进的布谷鸟算法能快速有效的找到问题的高质量解,特别在大规模测试系统上,发电成本明显低于 文献中给出的值. (4)提出了求解电力系统机组组合问题的改进二进制布谷鸟算法.首先,用二维数组定义了该问 题的解的结构,并引入一种新的二进制更新策略帮助算法沿着可行的方向搜索.然后,使用基于信息 交换机制和优先列表的启发式搜索策略提高算法的局部寻优能力.最后,使用贪婪式的修复策略提高 解的多样性.数值实验在含有 10-100 机组的基准测试系统上进行,实验结果表明改进的二进制布谷鸟 算法能够高效的求解机组组合问题,与其它算法相比,其收敛速度更快,解的质量更高. 本文通过对燃煤机组相关问题的研究,在动力配煤、传统经济调度和经典机组组合的决策方面均 提出了新的解决方法,实验结果验证了新方法的有效性,同时也为火电厂的节能减排工作和进一步的 工程应用奠定了理论基础. ―2― 论文摘要(英文) With the rapid growing of economic and the increasing of the total social electricity consumption year by year, the power industry has become an important industry to support economic development in China. It is well known that China is rich in coal, but poor in oil and little of gas, so coal-fired units are still the main generation means. Facing the increasingly energy situation and environmental pressure, to promote the energy conservation and emission reduction of coal-fired power generation enterprises will be an inevitable requirement to maintain the sustainable development of China'

s economy. Based on the research of the energy saving and emission reduction of electric power production enterprises, the safe and economic operation of coal-fired units and multi-unit combined scheduling, this dissertation has conducted sufficient research on optimization problems in the process of power generation, such as coal blending, economic dispatching and unit commitment. The main contributions are as follows: (1) The cuckoo search algorithm'

s variants and applications are summarized and its parameter settings are studied in detail. The sensitivity analysis of such parameters is carried out on six benchmark test functions. Simulation results show that the cuckoo search algorithm is very sensitive to different values. However, a good combination of parameters is obtained by orthogonal experiment. (2) Coal blending technology can be made use of solving the mismatch coal quality problem between the actual supplied coal and the designed requirement, hence to reduce generation cost and improve production effectiveness. Based on the analysis of the relationship between single coal and the blend, this dissertation formulates a resource-constrained 0-1 mixed integer programming model by using the minimum blending cost and limiting the number of single coal. In order to further reduce the blending cost and keep the blend not slagging in theory, partial of the coal'

s quality indicators of the blend are allowed to exceed the standard value, and the exceeding parts are added into the objective function as a penalty, then a new model is established. Simulation results verify the effectiveness of the new model. A lower blending cost is obtained on the premise of a small sacrifice of some coal'

s quality indicators. (3) The classic economic dispatch problem is studied, which has non-convex, non-continuous and non-linear solution spaces and is difficult to solve. In this dissertation, an improved cuckoo search algorithm is proposed to solve this problem. First, the structure of the solution is defined. Then the self-adaptive step size factor and the neighbor-study strategies are established to improve the algorithm'

s ability of exploration and exploitation. The improved λ iteration method is used to improve the speed of generating a new solution. Simulation experiments are carried out on the benchmark test systems of 6, 13,

20 and 40-unit. The simulation results show that the improved cuckoo search algorithm is effective and better than other algorithms in the literature, especially for large-scale unit systems. (4) An improved binary cuckoo search algorithm is proposed to solve the classic unit commitment problems in power systems. First, a two-dimensional array is used to define the structure of the solution. Next, a new binary updating mechanism is introduced to help the proposed algorithm to choose a feasible direction. A heuristic search algorithm based on information exchange strategy and a new priority list is used to improve the ability of exploitation. A greedy repair strategy is used to increase the diversity of solutions. Simulation experiments are carried out on the benchmark test systems from

10 to 100-unit. The simulation results show that the improved binary cuckoo search algorithm can solve the unit commitment problems efficiently. Compared with other algorithms, it is faster and the solution quality is higher. Through the study of relevant problems of coal-fired units, new methods are proposed in the decision-making of coal blending, traditional economic dispatching and classic unit commitment. The experimental results also verify the effectiveness of the new methods. It lays a theoretical foundation for energy conservation and emission reduction of thermal power plants and further engineering application. ―3― 论文主要创新点 1)对布谷鸟算法的变体和应用进行了归类和总结,重点研究了布谷鸟算法的参数设置.在六个基 准测试函数上进行了参数灵敏度分析,实验结果表明布谷鸟算法对自身参数值敏感,通过正交实验获 得了一组较好的参数组合. 2)为了降低配煤成本并保证配煤方案在理论上不结渣,允许部分混煤煤质指标超标并加以惩罚, 建立了基于配煤成本最低且单煤资源受限的 0-1 混合整数非线性规划模型.得到了以较小的代价牺牲 部分混煤煤质指标的单价更低的配煤方案. 3)提出了求解经济调度问题的相邻学习布谷鸟算法.使用自适应步长因子和相邻学习策略提高算 法的全局探测能力和局部寻优能力.使用改进的λ方法提高新解的生成速度.实验结果表明该算法能 快速有效的找到问题的高质量解. 4)提出了求解机组组合问题的改进二进制布谷鸟算法.引入新的二进制更新策略使算法沿着可行 的方向搜索.使用基于信息交换机制和优先列表的启发式搜索策略提高算法的局部寻优能力.使用贪 婪式的解的修复策略提高解的多样性.实验结果表明该算法收敛速度更快,解的质量更高. ―4― 攻读博士学位期间取得的学术成果 注:按有关规范填写本人为第一作者或导师是第一作者、本人为第二作者,且第一署名单位为东北大 学........

下载(注:源文件不在本站服务器,都将跳转到源网站下载)
备用下载
发帖评论
相关话题
发布一个新话题