【作者】 王竹；

【导师】 盛德仁；

【作者基本信息】 浙江大学 ， 热能工程， 2019， 硕士

【摘要】 随着“上大压小”政策以及“超低排放”战略的实施,大容量、高参数、低消耗、少排放是火电机组未来发展的主要方向,给电力生产过程带来了新的机遇与挑战。汽轮机在经过长期运行、DEH或通流部分改造后,汽轮机调节阀的实际流量特性将偏离设计值,从而影响机组一次调频和负荷控制能力。为保证机组安全、稳定运行,机组DEH控制系统中汽轮机调节阀的流量特性参数应与AGC控制和一次调频功能相匹配,准确获取汽轮机调节阀实际流量特性是实现对其精确控制的前提。针对机组运行数据具有数据量大、耦合性强的特点,本文首先介绍了数据预处理算法,以得到精简的稳态数据库。然后提出了基于试验数据和历史数据挖掘的汽轮机调节阀流量特性辨识方法,其中,基于历史数据挖掘的汽轮机调节阀流量特性辨识方法包括回归分析、迭代计算和变工况计算三种方法。采用改进的K-means聚类算法对结果进行提炼、拟合,分别通过仿真和分段线性化对DEH阀门管理曲线进行优化修正,确定阀门合理的重叠度。最后介绍了调节阀流量特性仿真与优化方法,以及重叠度调整前后的节流损失分析。编写软件以实现上述功能,应用于某660 MW机组,结果表明不同方法均能辨识出汽轮机调节阀真实的流量特性曲线,验证了三种辨识方法的正确性和可行性。同时,优化后的机组在顺序阀运行下,汽轮机调节阀流量特性曲线的线性度有了很大改善,机组AGC变负荷和一次调频的能力得到了提高,为指导机组实际运行优化提供参考。更多还原

【Abstract】 With the implementation of the policy of "developing large units and suppressing small ones" and the strategy of "ultra-low emission",large capacity,high parameters,low consumption and low emission are the main direction for the future development of thermal power units,bringing new opportunities and challenges to the power production process.After long-term operation,DEH or flow path retrofit,the actual flow characteristics of steam turbine’s regulating valves will deviate from the design values,thus affecting the unit primary frequency and load control ability.In order to ensure the safe and stable operation of the unit,the flow characteristics parameters of steam turbine’s regulating valves in the DEH control system of the unit should be matched with the functions of AGC and primary frequency regulation.Accurate acquisition of the actual flow characteristics of steam turbine’s regulating valves is the prerequisite for accurate control.In view of the large data volume and strong coupling of unit operation datas,in this paper,the data preprocessing algorithm is firstly introduced to obtain a compact steady state database.Then,the flow characteristics identification methods of steam turbine’s regulating valves based on experimental data and historical data mining are proposed.The latter includes regression analysis,iterative calculation and variable condition calculation.The improved K-means clustering algorithm is used to refine and fit the results.The DEH valve management curve is optimized and corrected by simulation and piecewise linearization to determine the reasonable overlap of the valves.Finally,the simulation and optimization methods of flow characteristics of regulating valves are introduced,including the throttling losses analysis after the adjustment of overlap.The software is compiled to realize the above functions and applied to a 660 MW unit.The results show that different methods can identify the actual flow characteristics curve of steam turbine’s regulating valves,and verify the correctness and feasibility of the three identification methods.At the same time,the linearity of flow characteristics curve has been greatly improved in sequence valve operation mode,and the capacity of AGC load variation and primary frequency regulation has been enhanced,which provides reference for guiding the actual operation optimization of units.更多还原