摘要
针对军方用户使用车辆电源带非线性负载时,因缺乏明确功率匹配指示有预先安全评测的需求,提出运用极限学习机(ELM)的方法,通过在已建立的车辆电源仿真平台上,采集相关有效数据、分析其非线性关系,并建立相应预测模型,最终实现电源带载能力的评估、滤波电容的选取与直流侧电压纹波的预测,以给军方用户更加合理、安全、高效地使用该电源提供安全指导.预测结果分析与实际应用效果表明,所建立的ELM预测模型预测精度较高、泛化能力较强.
Aimed at the need for advanced savety evaluation presented by military user using vehicle power supply to drag nonlinear load without a clear power match indication,the extreme learning machine（ELM）method is suggested to gather relevant data on a vehicle power supply simulation platform established,analyze their nonlinear relationship,and establish corresponding prediction models,so that the evaluation of load bearing capacity of the power supply,the selection of filtering capacitance,and the prediction of DC side voltage ripple are realized at last,providing the user with safety guidance to use this power supply in even more reasonable,safe,and high-efficient way.Prediction results and practical application effects show that the established ELM prediction model will have higher prediction precision and stronger generalization ability.
出处
《兰州理工大学学报》
CAS
北大核心
2018年第1期74-79,共6页
Journal of Lanzhou University of Technology
基金
国家自然科学基金(61364011)
甘肃省青年自然科学基金(1506RJYA108)
甘肃省先进控制重点实验室开放基金(XJK201505)
关键词
车辆电源
极限学习机
预测模型
vehicle power supply
extreme learning machine
prediction model
