| 引用本文: | 牛林,杜至刚,赵建国.基于相关向量机的交流特高压输电线路可听噪声的预测研究[J].电力自动化设备,2009,(6): |
| .Prediction of audible noise of AC UHV transmission lines based on relevance vector machine[J].Electric Power Automation Equipment,2009,(6): |
|
|
| 摘要: |
| 通过对输电线路可听噪声的累积分布特性、频谱特性和横向分布特性的分析,推导出可听噪声声压级和声功率级之间的定量关系式:提出基于相关向量机的特高压输电线路可听噪声的预测模型,该模型由于采用了贝叶斯框架下的学习机制表现出优秀的泛化能力,使其预测具有较高的准确度以及更广泛的适用范围;利用相关向量机回归模型预测了我国1 000 kV交流单回和同塔双回架空输电线路的可听噪声,结果表明:我国交流特高压输电线路的初设结构合理,满足环境噪声标准的要求.特高压单回路采用8×LGJ-500及更大截面的导线可听噪声低于55 dB,采用8×LGJ-630以上的导线低于52 dB;特高压双回线路同相序时采用8×LGJ-630及更大截面的导线可听噪声低于55 dB,逆相序时需采用8×LGJ-800导线才能使噪声低于55 dB,只有同相序时采用8×LGJ-800导线才能低于52 dB. |
| 关键词: 交流 特高压输电线路 可听噪声 相关向量机 预测 |
| DOI： |
| 分类号:TM723 |
| 基金项目: |
|
| Prediction of audible noise of AC UHV transmission lines based on relevance vector machine |
|
NIU Lin1 DU Zhigang2 ZHAO Jianguo1
|
| Abstract: |
| The cumulative distribution,frequency spectrums and lateral profiles of transmission line AN (Audible Noise) are analyzed and the quantitative relationship between the sound pressure and the sound power is deduced. An AN prediction model based on RVM(Relevance Vector Machine) is presented for UHV transmission lines,which adopts a RVM - based probabilistic Bayesian learning framework and has good generalization capability,resulting in higher prediction accuracy and wider application range. With the proposed ... |
| Key words: AC UHV transmission line audible noise relevance vector machine prediction |
