基于自适应模糊方法的电站锅炉燃烧优化研究

发布时间：2018-05-26 10:44

本文选题：电站锅炉 + 变量选择　；参考：《华北电力大学(北京)》2017年硕士论文

【摘要】：电站锅炉燃烧优化技术是一种能提高机组能源利用率和降低污染物排放的有效方法和途径,在节能减排要求严格的当下,备受火电厂青睐。本文针对国内燃煤机组的运行特点,主要研究了该技术中燃烧模型的建立和优化等关键部分。论文研究内容和研究成果概括为以下几个方面:(1)分析了现在已有的建模方法,针对算法在建模过程中的不足,介绍了一种新的高效的建模方法,即基于树结构的自适应模糊方法,简称模糊树(FT),进而介绍了鲁棒性较强的ε-模糊树(ε-FT)方法。(2)针对热工过程变量之间的强相关性和强耦合性,采用偏最小二乘法(PLS)对燃烧模型的输入变量进行重要信息提取和变量选择,将得到的最优变量子集作为ε-FT模型的输入,建立了燃烧系统的PLS-ε-FT模型,并与其他方法进行了对比。仿真结果表明,该模型通过PLS进行变量选择,消除了变量的相关性,降低了模型的维数和复杂程度,提高了模型的预测精度和泛化能力。(3)针对电站锅炉现场采集数据常带有噪声而FT又对噪声特别敏感的问题,提出了基于局部异常因子(LOF)的加权模糊树(W-FT)算法,并采用两个典型的非线性例子验证了算法的鲁棒性。进而,基于W-FT建立了鲁棒性较强的电站锅炉燃烧系统模型,并与其他方法所建立的模型进行了对比。仿真结果表明,所提的W-FT方法能有效地辨识噪声和异常值,具有较强的鲁棒性;所建立的模型预测精度较高,泛化能力较强。(4)为了提高机组能源利用率和降低污染物排放,基于ε-FT建立了排烟温度和NOx排放特性模型,提出了两种优化策略,采用改进的果蝇优化算法(MFOA)对电站锅炉的可调参数在一定的范围内进行寻优。仿真结果表明,所建立模型具有较高的预测精度和较强的泛化能力;所提的优化策略均能实现降低排烟温度和NO_x排放量;所提的燃烧优化方案,耗时较少,适合在线应用。这些为电厂的实际运行提供了重要的参考。
[Abstract]:Power plant boiler combustion optimization technology is an effective way to improve unit energy utilization and reduce pollutant emissions, in the energy saving and emission reduction requirements are strict, is favored by thermal power plants. According to the operating characteristics of domestic coal-fired units, this paper mainly studies the key parts of the technology, such as the establishment and optimization of combustion models. In this paper, the research contents and research results are summarized as follows: (1) the existing modeling methods are analyzed, and a new and efficient modeling method is introduced in view of the shortcomings of the algorithm in the modeling process. That is, the adaptive fuzzy method based on tree structure, referred to as the fuzzy tree, then introduces the 蔚 -fuzzy tree (蔚 -FT) method with strong robustness, aiming at the strong correlation and strong coupling between variables in the thermal process. The input variables of combustion model are extracted and selected by partial least square method (PLS). The optimal subset of variables is taken as the input of 蔚 -FT model. The PLS- 蔚 -FT model of combustion system is established and compared with other methods. The simulation results show that the model selects variables through PLS, eliminates the correlation of variables, and reduces the dimension and complexity of the model. The prediction accuracy and generalization ability of the model are improved. (3) aiming at the problem that the data collected from the boiler in power station often have noise and FT is especially sensitive to noise, a weighted fuzzy tree W-FTG algorithm based on local anomaly factor (LOF) is proposed. Two typical nonlinear examples are used to verify the robustness of the algorithm. Furthermore, based on W-FT, the model of boiler combustion system with strong robustness is established and compared with the model established by other methods. The simulation results show that the proposed W-FT method can identify noise and outliers effectively and has strong robustness. Based on 蔚 -FT model of exhaust gas temperature and NOx emission characteristics, two optimization strategies are proposed. The improved Drosophila optimization algorithm (MFOAA) is used to optimize the adjustable parameters of utility boilers in a certain range. The simulation results show that the proposed model has high prediction accuracy and strong generalization ability, the proposed optimization strategies can reduce smoke exhaust temperature and NO_x emissions, and the proposed combustion optimization scheme takes less time and is suitable for on-line application. These provide important reference for practical operation of power plant.
【学位授予单位】：华北电力大学(北京)
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TM621.2

【参考文献】