| 摘 要: 预测性维护是工业互联网应用的重点，实现预测性维护的关键是对设备系统或核心部件的寿命进行有效预测。随着近年来机器学习的发展，机械设备海量数据已成为工业互联网分析核心部件剩余寿命的关键指标，也成为设备健康管理决策性数据。基于工程机械设备大数据，结合XGBoost、随机森林、LightGBM等多种机器学习模型，多维度探究影响机械核心部件寿命的机器学习模型效果，建立Stacking算法模型融合的部件寿命预测模型，并在核心部件数据上验证模型预测有效性，从而减少设备非计划停机时间，推进智能制造和预测性维护的进步。 |
| 关键词: 工程机械；寿命预测；机器学习；Stacking |
|
中图分类号: TP181
文献标识码: A
|
|
| Life Prediction of Construction Machinery Core Components Based on Stacking Model Fusion |
|
LIANG Chao
|
( School of Information Science and Technology, Zhejiang SCI-TECH University, Hangzhou 310018, China)
|
| Abstract: Predictive maintenance is the focus of industrial Internet application.The key to predictive maintenance is to effectively predict the life of equipment system or core components.With the development of machine learning in recent years,the massive data of mechanical equipment has become the key index of the residual life of core components in industrial Internet analysis and the decision-making data of equipment health management as well.Based on big data of construction machinery,multi-dimensional search of sample characteristics affecting the life of mechanical core components is conducted by combining multiple machine learning models,such as XGBoost,RandomForest and LightGBM,where Stacking algorithm is used to construct the component life prediction model,and the validity of model prediction is verified on the core component data.This reduces unplanned downtime and advances the intelligent manufacturing and predictive maintenance. |
| Keywords: construction machinery;life prediction;machine learning;Stacking |
