基于随机森林算法的船舶柴油机“冒黑烟”现象预测  

Prediction of Black Smoke from Diesel Engine Using Random Forest Algorithm

作　　者：王晓东[1] WANG Xiaodong(Warship Automatic System Division,Shanghai Ship and Shipping Research Institute Co.,Ltd.,Shanghai 200135,China)

机构地区：[1]上海船舶运输科学研究所有限公司舰船自动化系统事业部,上海200135

出　　处：《上海船舶运输科学研究所学报》2022年第4期33-38,共6页Journal of Shanghai Ship and Shipping Research Institute

摘　　要：为有效避免船舶加速航行过程中因柴油机燃烧不充分而引发的“冒黑烟”现象,利用船舶主动力推进系统上位机软件收集船舶主动力数据,将人工智能算法中的随机森林(Random Forest)算法与Python软件相结合,预测船舶加速航行时可能出现“冒黑烟”现象的时间,为柴油机控制器的操作控制提供依据。通过对比试验对随机森林模型算法的有效性进行验证,结果表明,采用该算法能很好地预测和避免船舶加速航行时出现的“冒黑烟”现象,确保柴油机的运行工况良好,从而使船舶更好地满足绿色运营的需求。Black smoke from the diesel engine is an indication of insufficient combustion.It usually happens during ship accelerating.This paper introduces a design which collects the main power data of the ship and predicts acceleration maneuvering with the Random Forest algorithm in Python.The prediction is sent to the diesel engine controller to adapt the engine to the condition change.The effectiveness of Random Forest algorithm in prediction is compared to other algorithm and the advantage of the former is noticeable and meets the demands of green shipping.

关 键 词：随机森林算法 船舶柴油机 不充分燃烧 Python软件 

分 类 号：U664.121[交通运输工程—船舶及航道工程]