首页|期刊导航|电源学报|兆瓦级非车载充电机的模型预测试验策略

兆瓦级非车载充电机的模型预测试验策略OA

Model Prediction Test Strategy of MW Off-board Charger

中文摘要英文摘要

兆瓦级非车载充电机的大规模应用面临着传统检测方法所存在的高能耗、低动态精度及多模块协同检测能力欠缺等问题,因此提出一种融合模型在环 MIL(model-in-the-loop)仿真与正交优化长短期记忆Orthogonal-LSTM(Orthogonal long short-term memory)网络的模型预测试验方法,旨在突破试验平台功率瓶颈并提升复杂工况下的检测适应性.通过构建模型预测控制MPC(model predictive control)驱动的虚拟充电机模型,实现对待试验充电机的动态特性模拟.然而,面对充电模块的叠加、功率的提升,模型与实际试验结果间的误差增大.为解决这一问题,结合正交试验法优化LSTM模型参数,实现模型的误差预测与补偿.实验结果表明:所提方法为单机多模块结构的兆瓦级充电机提供了一种高精度、低功耗且可扩展的检测解决方案,具有重要的工程推广价值.

The large-scale application of megawatt-level off-board chargers faces problems such as high energy consumption,low dynamic accuracy,and lack of multi-module collaborative detection capabilities existing in traditional detection methods.A model prediction test method integrating model-in-the-loop(MIL)simulation and orthogonally optimized long short-term memory network(Orthogonal-LSTM)is proposed,aiming to break through the power bottleneck of test platform and improve detection adaptability under complex working conditions.By constructing a virtual charger model driven by model predictive control(MPC),the dynamic characteristics of the charger under test are simulated.However,in the face of the superposition of charging modules and the increase in power,the error between the model and the actual test results increases.To solve this problem,the LSTM model parameters are optimized in combination with the orthogonal test method to achieve the error prediction-compensation of the model.Experimental results show that the proposed method provides a high-precision,low-energy-consumption and scalable detection solution for megawatt-class chargers with a single-unit multi-module structure,and has great value for engineering promotion.

王维;金赣;王理想;杨靖宇

南京师范大学电气与自动化工程学院,南京 210023南京师范大学电气与自动化工程学院,南京 210023南京师范大学电气与自动化工程学院,南京 210023南京师范大学电气与自动化工程学院,南京 210023

信息技术与安全科学

非车载充电机试验方法模型预测试验模型预测控制长短期记忆网络模型正交试验法

Off-board charger test methodmodel predictionmodel predictive control(MPC)long short-term memory(LSTM)network modelorthogonal test method

《电源学报》 2026 (9)

65-76,144,13

10.13234/j.issn.2095-2805.2026.9.65

评论