共查询到20条相似文献,搜索用时 203 毫秒
1.
2.
风机的性能曲线是风机选型和优化运行的重要依据.通常该曲线通过试验试验数据和性能图表上的数据进行曲线拟合获得.由于该曲线非线性很强,传统方法复杂昂贵,而且拟合精度不高。针对以上不足,提出了一种基于非线性权重自适应粒子群优化(NWAPSO)参数全局寻优的最小二乘支持向量机(LS-SVM)风机性能预测方法。通过最小二乘支持向量机建模,并应用非线性权重自适应粒子群优化算法对模型参数进行全局寻优,得到具有较高精度的风机性能曲线。计算结果表明,根据本文方法建立的模型很简洁,只需要知道少量的训练样本就能建立,可以比较精确的预测风机性能,具有较显著的工程应用价值。 相似文献
3.
风机的性能曲线是风机选型和优化运行的重要依据.通常该曲线通过试验试验数据和性能图表上的数据进行曲线拟合获得.由于该曲线非线性很强,传统方法复杂昂贵,而且拟合精度不高。针对以上不足,提出了一种基于非线性权重自适应粒子群优化(NWAPSO)参数全局寻优的最小二乘支持向量机(LS—SVM)风机性能预测方法。通过最小二乘支持向量机建模,并应用非线性权重自适应粒子群优化算法对模型参数进行全局寻优,得到具有较高精度的风机性能曲线。计算结果表明,根据文中方法建立的模型很简洁,只需要知道少量的训练样本就能建立,可以比较精确的预测风机性能,具有较显著的工程应用价值。 相似文献
4.
5.
6.
基于一步最小二乘法建立航空发动机状态变量模型 总被引:1,自引:0,他引:1
《燃气轮机技术》2017,(4)
研究采用一步最小二乘法建立用于航空发动机控制系统及故障诊断设计中的状态变量模型,结合伪随机二进制激励信号求取模型矩阵系数。在线性最小二乘准则下,算法保证模型动态响应与非线性模型响应一致,过程一步完成无须迭代。利用一步最小二乘法建立了JT9D涡扇发动机稳态工作点的小偏差状态变量模型,仿真表明该方法得到的状态变量模型能够精确拟合发动机的时域和频域特性。相比与传统拟合方法,该方法不仅计算速度快,而且拟合精度大大提高,验证了方法的有效性。 相似文献
7.
改进的PLSR算法在大坝安全监控中的应用 总被引:1,自引:1,他引:0
针对自变量中含有过多与因变量无关的信息时应用偏最小二乘回归算法将出现偏差的问题,引入改进的偏最小二乘回归算法对坝体水平位移建立数学模型,并以安徽陈村混凝土拱坝为例进行分析.结果表明,改进的偏最小二乘回归算法通过正交投影变换解释矩阵能有效改善噪音的影响,建立的统计模型在保证解释能力的前提下其拟合效果和预测能力较偏最小二乘回归算法均有所提高. 相似文献
8.
9.
循环流化床锅炉燃烧系统中床温的模糊辨识 总被引:1,自引:0,他引:1
针对循环流化床锅炉稳定燃烧、经济运行的关键参数--密相区床温进行了动态特性研究,建立了床温模型.通过阶跃扰动试验分析了床温系统的动态特性,给出了某工况点附近的单输入单输出系统的传递函数.为床温设计了实用的基于模糊规则的非线性建模方法:首先利用简化的减法聚类法得到输入空间的模糊划分,并假定模糊规则数目囊括输入模糊集的所有组合,其次利用递推最小二乘算法学习得到输出模糊集的中心,从而构造1个带乘积推理机、单值模糊器和中心平均解模糊器的模糊系统.验证数据表明:设计的模糊系统能有效拟合床温系统的输入输出曲线,而且算法简单、实用. 相似文献
10.
太阳电池直流模型的计算机仿真 总被引:1,自引:3,他引:1
太阳电池特性由材料、设计和工艺决定,太阳电池直流模型则可描述其电特性。现有的测试方法不能直接测出模型参数,但容易获得太阳电池片实测I—V数据,采用最小二乘法拟合出模型参数,就可以建立模型参数与制造工艺的直观联系。本文对提高模型拟合的速度及精度的实现方法进行讨论,在计算机上实现快速拟合。 相似文献
11.
离心泵选型合理与否,与节能及保证供水关系很大.本文通过对水泵性能的差异及水泵运行过程中性能的变化两个主要因素的分析,提出了离心泵的合理选型方法. 相似文献
12.
13.
14.
15.
16.
通常泵组件的调节主要依靠工程师经验,难以保证水泵在所有时段内均高效运行,易造成能源浪费,对此提出了以效率模型为依据的泵站优化调度方法。为减小误差,使用泵站实际运行数据基于BP神经网络训练水泵特性。在此基础上,利用训练后的数据初始化种群,以泵站效率之和最小为目标,建立泵站优化调度数学模型,设定水量、压力和高效区等约束条件,采用改进的遗传算法求解水泵组合方案及各泵运行参数。结果表明,改进遗传算法对泵站调度的优化效果明显,可适用于泵站的优化节能运行。研究成果为泵站优化调度运行提供了一种新方法。 相似文献
17.
18.
Several photovoltaic applications, specially the water pumping systems, are installed to contribute in the socio–economic development in Algerian Sahara. About sixty pumps are installed in remote regions to supply water for drinking and irrigation. The centrifugal pumps present the state-of-the-art for most applications. This paper presents the electrical and hydraulic performance of a surface centrifugal pump versus total water heads and versus a size of PV array. Also the area irrigated by this solar surface pump is calculated under Sahara climate conditions for four crops, namely wheat, potatoes, tomatoes and sunflowers. These were selected to meet with food standards of the sites. Two application sites were chosen to test the developed model. The first is Bechar station and the second is Tamanrasset station. In both, the recorded solar radiation data were used. The results, which are illustrated by tables and curves, are analysed. The analysis of the performance concluded that this surface pump is suitable for installation in the Sahara regions with low water head. 相似文献
19.
Photovoltaic (PV) pumping offers the possibility of supplying water to remote and desert regions for their daily needs. The sizing of the PV pumping systems is a very significant step in order to optimize the power peak of the PV array and to ensure the best choice of the motor, the pump and the inverter. Two mathematical models were proposed in this article to contribute in the studies of PV pumping sizing. These models link directly the operating electrical power to the water flow rate of the pump versus total head. These models are based essentially on the experimentation of pumps on CDER PV pumping test facility. Two pumping systems are tested: the first uses a centrifugal pump and the second uses a positive displacement pump. The results obtained by the models are very satisfactory. Also, the models enabled us to simulate the electrical and hydraulic performances of two tested pumps. The performances are calculated using the measured meteorological data of different sites located in Sahara and coastline regions of Algeria. 相似文献
20.
The I–V characteristic curve is very important for solar cells/modules being a direct indicator of performance. But the reverse derivation of the diode model parameters from the I–V curve is a big challenge due to the strong nonlinear relationship between the model parameters. It seems impossible to solve such a nonlinear problem accurately using linear identification methods, which is proved wrong in this paper. By changing the viewpoint from conventional static curve fitting to dynamic system identification, the integral-based linear least square identification method is proposed to extract all diode model parameters simultaneously from a single I–V curve. No iterative searching or approximation is required in the proposed method. Examples illustrating the accuracy and effectiveness of the proposed method, as compared to the existing approaches, are presented in this paper. The possibility of real-time monitoring of model parameters versus environmental factors (irradiance and/or temperatures) is also discussed. 相似文献