【作者】 刘丽华；

【导师】 杜小泽；

【作者基本信息】 华北电力大学 ， 工程热物理， 2014， 博士

【摘要】 近年来,空冷技术在我国北方火力发电中得到了大力的发展。环境因素与火电直接空冷机组的耦合比较复杂,环境风速与风向、环境温度以及轴流冷却风机的性能等因素都对火电直接空冷机组性能产生影响,并且受轴流风机群运行中的群抽效应和环境风影响下的热风回流的耦合影响,轴流风机群外围的空冷凝汽器单元受环境风影响大。在四台联建火电直接空冷机组上进行了春、夏与秋三个季节不同工况下的机组热力性能实验,并在夏季主导风向下,基于两台联建火电直接空冷机组的轴流风机群分区调节的基础上,进行了不同工况下的机组热力性能实验。然后建立火电直接空冷机组性能的灰关联度模型,定量分析火电直接空冷机组自身运行实验变量与环境条件以及轴流风机群各风机的性能对火电直接空冷系统性能的影响,并基于上述两个热力性能实验数据,分别应用神经网络(ANN)与最小二乘支持向量机(LS-SVM)等算法,建立汽轮机背压预测模型。接着,在选定的基准工况下应用火电直接空冷机组汽轮机背压LS-SVM预测模型,分析出处于夏季主导风向下的轴流风机群不同分区风机转速降低和升高对环境风下游与上游火电直接空冷机组抵御环境风影响的作用规律。建立由太阳辐射引起的直接空冷系统附加热负荷计算模型,对太阳辐射影响空冷凝汽器热负荷的程度进行分析。基于火电直接空冷机组热力性能实验数据,应用建立的火电直接空冷机组性能灰关联度模型进行分析,结果显示出,文中选定的机组自身运行与环境实验变量对火电直接空冷机组的性能影响不同,并且处于不同位置的轴流冷却风机的性能对火电直接空冷机组性能的影响也不同,处于外围的轴流冷却风机的性能对火电直接空冷机组性能的影响最强烈。以北方地区某135MW火电直接空冷机组凝汽器的全年太阳辐射量为主线,假定汽轮机排汽量不变,环境温度为月平均日温度,揭示出由太阳辐射引起的月平均附加热负荷在1-5月期间升高,5月时达到最大值,为371.75kW,而后逐步降低,与日照时数的变化趋势基本一致。假定方位角为零时,凝汽器由太阳辐射引起的月平均附加热负荷增加而导致汽轮机背压升高,1-7月间汽轮机背压变化增加,7月时达到最大值,为0.2097kPa,8-12月相应的影响会减弱,与环境温度的变化趋势大致相同,并且由太阳辐射引起的直接空冷系统附加热负荷平均占空冷凝汽器总热负荷的0.134%。基于LM-BP算法,建立火电直接空冷机组的汽轮机背压ANN预测模型,其预测值与汽轮机背压的真实值基本一致,模型的平均相对误差,MRE,为9.27%；平均均方差,RMSE,为1.827kPa；绝对变异分数,R2,为0.9859。为检测模型的鲁棒性与可靠性,分别将空冷凝汽器的环境输入变量,如自然环境风速、自然环境风向、空冷岛挡风墙上环境风速和空冷岛挡风墙上环境风向等参数,加入±5%的随机波动后,输入到模型中,其MRE分别为14.57%,12.21%,11.22%和11.16%,RMSE分别为2.773,2.130,1.844和1.895kPa,R2分别为0.9641,0.9779,0.9836和0.9824,说明模型的稳定性很好,数据波动产生的相对误差能得到有效的抑制。针对夏季主导风向下的两台联建火电直接空冷机组,分别建立相应的汽轮机背压LS-SVM预测模型,讨论轴流风机群不同分区风机转速降低与升高时,轴流风机群环境风下游与上游火电直接空冷机组汽轮机背压的变化规律：(1)轴流风机群不同分区的轴流冷却风机转速发生变化时,汽轮机背压变化明显不同；(2)改变其中一台机组轴流冷却风机的转速,也会对另一台机组的汽轮机背压产生影响；(3)单纯提高轴流冷却风机转速并非总能够起到降低机组汽轮机背压的效果,适当降低受热风回流影响较大分区的轴流冷却风机转速,反而可以起到降低轴流风机群环境风下游机组汽轮机背压的作用,并且轴流风机群环境风上游机组汽轮机背压也不会升高。更多还原

【Abstract】 Recently air-cooling technology has been actively developed in north China. Environmental factors and the thermal direct air-cooled power generating units are coupled complexly, in which the performances of the thermal direct air-cooled power generating units are inluenced by the environmental wind velocity, the enviromnental wind direction, the environmental temperature and the performances of the axial flow cooling fans. In addition, the air-cooled condenser (ACC) cells in the pherimery are influenced strongly by the cluster effect of the operation of the axial flow fan cluster and the exhaust hot air recirculation of the environmental natural wind simultaneously.The thermal performance experiments of the four thermal direct air-cooled units constructed together were carried out during spring, summmer, and autumn. The further thermal performance experiments of the fan cluster portion operation on the two thermal direct air-cooled units constructed together were carried out under the summer dominant wind. The comprehensive effects of the unit experimental variables itself, and the environmental conditions, and the performances of the experimental axial flow cooling fans on the unit performances were obtained quantitatively by the grey relational analysis models, and the models for the turbine back pressure prediction based on the above thermal performance experimental data were established respectively using the artifical neural network (ANN) algorithm and the least squares support vector machine (LS-SVM) algorithm. Secondly, the rules of the effect on the downstream and upstream units against the environmental wind could be obtained based on the turbine back pressure LS-SVM prediction models when the speeds of the different portion fans on the axial flow fan cluster were reduced or increased based on the reference conditions under the summer dominant wind. The heat load from the solar radiation calculation model on the ACC was built, and the related influencing analysis was obtained.Based on the analyzing results of the grey relational models on the experimental data of the thermal performances of the thermal direct air-cooled units, it is found that the unit experimental variables itself and the environmental conditions had different influences on the performances of the thermal direct air-cooled units, and the performances of axial flow cooling fans in different places of the axial flow fan cluster had different influences on the performances of the thermal direct air-cooled units as well, with the strongest effect of the fans in the periphery of the axial flow fan cluster.For the condenser in a135MW thermal direct air-cooled unit, the monthly averaged midday solar radiation over a period of21years was obtained, and its effects on the heat load and turbine back pressure were investigated based on the constant turbine outflow and monthly average day temperature of enviroment. The monthly average heat load from the solar radiation increases in a5-month period with371.75kW in May and then decreases, which is consistent with the sunshine time. For the azimuthal angle of0°, the turbine back pressure increases over a7-month period, with the maximum increment reaching for0.2097kPa in July, and then decreases over a5-month period, which is consistent with the environmental temperature, and the heat load from the solar radiation is the0.134%of the overall heat load of the thermal direct air-cooled condenser averagely.An ANN model based on a LM-BP algorithm was developed, which could be used for predicting the turbine back pressure of a thermal direct air-cooled unit. The prediction of the ANN model usually agreed well with the actual values of the turbine back pressure. Furthermore, when it was used for predicting the turbine back pressure, the model yielded agreeable result with mean relative errors(MRE) of9.273%and root mean square error (RMSE)1.83kPa and absolute fraction of variance (R2) of0.9859. Moreover, when the enviromental experimental variables with±5%random fluctuation, such as environmental natural wind velocity and direction, ACC wind velocity and direction on windwall, were input into the ANN prediction model, the related MRE was14.57%,12.21%,11.22%and11.16%, the related RMSE was2.77,2.13,1.84and1.89kPa, the related R2was0.9641,0.9779,0.9836and0.9824, which demonstrated the robustness and reliability of the model, and the relative errors of the data fluctuation were effectively restrained.The LS-SVM models for the turbine back pressure prediction of the two thermal direct air-cooled units constructed together were established respectively under the summer dominant wind, and the rules of the influences on the turbine back pressure variations of the downstream and upstream thermal direct air-cooled units were obtained when reducing and increasing the fan speed of different portion of the axial flow fan cluster portion:(1) The turbine back pressure variations are significantly different if the fans speed of different portions are changed.(2) If the air cooling fan speeds of one unit are changed, the turbine back pressure of the other unit would be affected either.(3) It does not always play a role to reduce the turbine back pressure of the unit when the speeds of the air cooling fans are increased appropriately, and it plays a role to reduce the turbine back pressure of the unit in the downstream of the environmental natural wind and not to increase the turbine back pressure of the unit in the upstream of the environmental natural wind if the fan speeds of the portions affected by exhaust hot air recirculation obviously are reduced appropriately.更多还原