基于LSSVM的大型造船厂24h用气流量预测建模  

24-hours Gas Flow Prediction Model of Shipyard Based on Least Squares Support Vector Machine

作　　者：桑愈 陆国平[1] 付向恒[2] 

机构地区：[1]南通大学电气工程学院,江苏南通226019 [2]北京航空航天大学自动化科学与电气工程学院,北京100191

出　　处：《液压与气动》2016年第1期66-70,共5页Chinese Hydraulics & Pneumatics

基　　金：国家自然科学基金(61174066)

摘　　要：针对大型造船厂气动系统的用气特点,建立了基于LSSVM 24 h用气流量预测模型。模型以上海某造船厂每天实际用气流量为样本,根据实际预测控制需要进行预测。仿真结果表明,基于LSSVM的24 h用气流量预测模型能够实时反映系统当前状态,预测结果基本上吻合实际用气量,能够满足大型离心空压机组启、停控制需求。According to the characters of pneumatic system of large shipyards, the least squares support vector machine （LSSVM） regression principle is introduced and 24-hours gas flow prediction model is established based on LSSVM. Taking the daily gas flow of a Shanghai shipyard as model sample and making predictions according to the actual control needs. The simulation results show that the 24-hour LSSVM gas flow prediction model can reflect the current state of the system and the results are basically consistent with the actual gas flow to meet the application requirements.

关 键 词：造船厂 流量预测 LSSVM 节能 

分 类 号：TH137[机械工程—机械制造及自动化] TP181[自动化与计算机技术—控制理论与控制工程]