流体机械2026，Vol.54Issue(6)：112-122,11.DOI:10.3969/j.issn.1005-0329.2026.06.014
五缸泵多领域多体数字孪生模型及其应用
Multi-domain multi-body digital twin model of a five-cylinder pump and its application
摘要
Abstract
To address the issues in traditional data-driven fault diagnosis methods for five-cylinder pumps,such as difficulty in obtaining fault data,insufficient diagnostic capability during the early stage of equipment operation,and poor model generalization,a multi-domain digital twin model of the five-cylinder pump was established.The motor system,control system,mechanical system,and hydraulic system of the five-cylinder pump were modeled,the mathematical models and parameter transfer relationships of each subsystem were clarified,and the overall model was consequently constructed.Based on fault mechanism analysis,the injection and response quantitative analysis of four typical faults,namely piston leakage,inlet blockage,bearing damage,and motor open circuit,were realized.A multi-domain simulation model coupled with electromechanical-hydraulic-control was built using MATLAB/Simscape to obtain simulation data after fault injection.The consistency of the digital twin model was verified using experimental data from the XZQ2200 drilling pump.To demonstrate the application of the model,a drilling rig circulation system model was constructed using the five-cylinder pump digital twin model,and a TimesNet-based fault diagnosis method was proposed.The results show that the root mean square error between the twin data and the measured data is 0.039 2.The fault diagnosis model achieves a fault location classification accuracy of 98.33%and a fault classification accuracy of 89.44%on the twin dataset.The established five-cylinder pump digital twin model has high fidelity and can effectively support typical fault mechanism analysis and fault diagnosis model construction.This study provides a new technical approach for predictive maintenance of complex equipment.关键词
五缸泵/数字孪生/故障注入/故障诊断/钻机Key words
five-cylinder pump/digital twin/fault injection/fault diagnosis/drilling rig分类
机械制造引用本文复制引用
张鑫,刘志亮,王晨阳,谢川,饶猛,左明健..五缸泵多领域多体数字孪生模型及其应用[J].流体机械,2026,54(6):112-122,11.基金项目
国家自然科学基金面上项目(52475091) （52475091）
山东省科技型中小企业创新能力提升工程项目(2025TSGCCZZB0141) （2025TSGCCZZB0141）
四川省科技计划项目(2024JDHJ0057) （2024JDHJ0057）
