On-line monitoring of power curves

被引：188

作者：

Kusiak, Andrew ^{[1
]}

Zheng, Haiyang ^{[1
]}

Song, Zhe ^{[1
]}

机构：

[1] Univ Iowa, Dept Mech & Ind Engn, Iowa City, IA 52242 USA

来源：

RENEWABLE ENERGY | 2009年 / 34卷 / 06期

关键词：

Power curve; Turbine monitoring; Data mining; Evolutionary computation; Least squares method; Maximium likelihood estimation; CONTROL CHARTS; TREES;

D O I：

10.1016/j.renene.2008.10.022

中图分类号：

X [环境科学、安全科学];

学科分类号：

08 ; 0830 ;

摘要：

A data-driven approach to the performance analysis of wind turbines is presented. Turbine performance is captured with a power curve. The power curves are constructed using historical wind turbine data. Three power curve models are developed, one by the least squares method and the other by the maximum likelihood estimation method. The models are solved by an evolutionary strategy algorithm. The power curve model constructed by the least squares method outperforms the one built by the maximum likelihood approach. The third model is non-parametric and is built with the k-nearest neighbor (k-NN) algorithm. The least squares (parametric) model and the non-parametric model are used for on-line monitoring of the power curve and their performance is analyzed. (c) 2008 Elsevier Ltd. All rights reserved.

引用

页码：1487 / 1493

页数：7

共 33 条

[1]

[Anonymous], 2006, Introduction to Data Mining

[2] Factory cycle-time prediction with a data-mining approach [J].

Backus, Phillip ;

Janakiram, Mani ;

Mowzoon, Shahin ;

Runger, George C. ;

Bhargava, Arnit .

IEEE TRANSACTIONS ON SEMICONDUCTOR MANUFACTURING, 2006, 19 (02) :252-258

[3]

BELL B, 2008, INDIVIDUAL WIND TURB

[4]

Berry M.J. A., 2004, DATA MINING TECHNIQU, V2nd

[5] Random forests [J].

Breiman, L .

MACHINE LEARNING, 2001, 45 (01) :5-32

[6]

CAL Z, 2006, IEEE T EVOLUTIONARY, V10, P658

[7] Analysis of two-component mixture Weibull statistics for estimation of wind speed distributions [J].

Carta, J. A. ;

Ramirez, P. .

RENEWABLE ENERGY, 2007, 32 (03) :518-531

[8]

Casella George, 2021, Statistical inference

[9]

DAVID AS, 1994, WIND TURBINE TECHNOL

[10] An experimental comparison of three methods for constructing ensembles of decision trees: Bagging, boosting, and randomization [J].

Dietterich, TG .

MACHINE LEARNING, 2000, 40 (02) :139-157

← 1 2 3 4 →