| 引用本文: | 王文杰,彭文杰,裴 吉,等.基于LabVIEW和Python的水泵数字孪生平台设计[J].灌溉排水学报,2026,45(5):29-39. |
| WANG Wenjie,PENG Wenjie,PEI Ji,et al.基于LabVIEW和Python的水泵数字孪生平台设计[J].灌溉排水学报,2026,45(5):29-39. |
|
| 摘要: |
| 【目的】采用数字孪生技术，结合机器学习和计算流体力学引入模态分解技术，实现对高维流场降阶。【方法】基于LabVIEW和Python开发水泵数字孪生平台，包括数据采集、数值仿真、内流分析与反馈调控4个模块，平台通过传感器将仿真边界条件参数定时采集、自动导入仿真软件进行计算，内流可视化与反馈调控，优化水泵运行维护。与传统系统相比，该平台可实时监控水泵、主动调控，减少停机时间，降低维修成本，提升数字孪生技术在水泵中的“四预”能力。为降低计算资源，以叶轮为研究对象，采用CFX仿真软件进行瞬态计算，以流场压力数据作为模态分解的支撑，采用本征正交分解对三种不同工况下的压力流场进行模态分解，评估流场重构精度并分析不同模态下的主导频率。【结果】数字孪生平台应用于水泵试验台，实现自动采集与仿真计算，且模态分解结果显示前5阶模态占流场大部分能量，均超70%，但呈不同的误差特征，相较于设计工况与大流量工况，小流量工况内流更加不稳定，主要归因于叶轮流道中部区域的流动分离以及叶轮出口区域的动静干涉作用。此外，3种工况下，第1阶与第2阶模态的主导频率为轴频（48.33 Hz），而高阶模态对应于主频的倍频，可能与叶轮流动的不对称性相关。【结论】基于水泵试验台的数字孪生平台成功解决了水泵实时工况与仿真计算之间的脱节问题，显著提升了系统的智能化水平。 |
| 关键词: 水泵；数字孪生；虚拟仿真；状态监测；模型 |
| DOI：10.13522/j.cnki.ggps.2025146 |
| 分类号: |
| 基金项目: |
|
| Design and implementation of a pump digital twin platform |
|
WANG Wenjie, PENG Wenjie, PEI Ji, YUAN Shouqi
|
|
Research Center of Fluid Machinery Engineering and Technology, Jiangsu University, Zhenjiang 212013, China
|
| Abstract: |
| 【Background】Data acquisition and numerical simulation are widely used for monitoring and evaluating pump performance, but they are often disconnected in practical application. This paper presents a digital twin platform for pumps to bridge this gap, in which machine learning, computational fluid dynamics (CFD), and modal decomposition techniques are applied in an integrated manner to represent high-dimensional flow fields using low-dimensional representations, thereby improving computational efficiency.【Method】The digital twin platform was developed based on LabVIEW and Python, consisting of four modules: data acquisition, numerical simulation, internal flow analysis, and feedback control. It collects boundary-condition data from sensors at regular temporal intervals and automatically feeds them into the simulation software for real-time computation. The platform also provides a graphical user interface to visualize internal flow fields, enabling feedback control and optimization of pump operation and maintenance. Compared with conventional monitoring systems, the platform supports real-time monitoring, efficient operation, and active regulation, while reducing maintenance costs and providing capabilities for prediction, early warning, and preventive control. As a demonstration example, the impeller was selected for transient flow simulation, in which proper orthogonal decomposition (POD) was applied to perform modal decomposition of pressure flow fields, evaluate reconstruction accuracy, and identify dominant frequencies of different modes.【Result】The pump digital twin platform was successfully applied to a pump test rig, enabling automated data acquisition and simulation. Modal decomposition results showed that the first five modes captured more than 70% of the total flow field energy, each exhibiting distinct error characteristics. Compared with high-flow conditions, low-flow conditions exhibited greater instability in the internal flow, due to flow separation in the mid-region of the impeller passage and rotor-stator interaction near the outlet. Under all operating conditions, the dominant frequencies of the first and second modes corresponded to the shaft frequency of 48.33 Hz, while high-order modes were associated with integer multiples of this frequency, likely resulting from asymmetric flow structures within the impeller.【Conclusion】The developed digital twin platform effectively integrates real-time operational data with numerical simulation. It overcomes the disconnection between physical operation and computational analysis, and significantly improves the intelligence and operational efficiency of pump systems. |
| Key words: pump; digital twin; virtual simulation; condition monitoring; model |
