Machine-Learning-Based Condition Monitoring of Power Electronics Modules in Modern Electric Drives

Li, Dinan; Kakosimos, Panagiotis; Peretti, Luca

doi:10.1109/mpel.2023.3236462

“…or thermal imaging cameras [6], [7] for measurement can be prohibitively expensive. On the contrary, the use of thermal dynamics [8], [9] or artificial intelligence (AI) technology [10] to indirectly measure the temperature of PMSMs represents a more efficient and cost-effective approach. Therefore, the modeling and analysis of the thermal dynamics of the PMSM have received extensive attention [11]- [14].…”

Section: Introductionmentioning

confidence: 99%

Explainable Neural Dynamics Models for Motor Temperature Prediction

Liao,

Chen,

Long

et al. 2023

Preprint

0

View full text Add to dashboard Cite

<p>The permanent magnet synchronous motor finds extensive use in industrial applications, and the development of effective thermal management solutions is crucial to enhance its power density. Accurate temperature prediction of the permanent magnet synchronous motor serves as the fundamental basis for designing effective thermal management strategies. Model-based prediction methods exhibit superior real-time performance, but the intricate modeling process requires substantial expert knowledge guidance and lacks versatility. Conversely, data-driven prediction methods, while offering flexibility, often lack physical implications in terms of system dynamics. This paper proposed a structured linear neural dynamics model for motor temperature prediction. This model is data-driven, with prior knowledge integrated into its structure, which preserves flexibility while guaranteeing system stability through the Perron-Frobenius theorem. Additionally, this paper achieves the decoupling of control input from state transitions and the embedded deployment of this model. The method is validated with a real dataset. The lightweight feature is demonstrated by the implementation of an STM32 Microcontroller with 1.808 KB and 27 mW. The paper is accompanied with the open source data and code at GitHub: https://github.com/ms140429/Explainable-Neural-Dynamics-Model</p>

show abstract

“…or thermal imaging cameras [6], [7] for measurement can be prohibitively expensive. On the contrary, the use of thermal dynamics [8], [9] or artificial intelligence (AI) technology [10] to indirectly measure the temperature of PMSMs represents a more efficient and cost-effective approach. Therefore, the modeling and analysis of the thermal dynamics of the PMSM have received extensive attention [11]- [14].…”

Section: Introductionmentioning

confidence: 99%

Explainable Neural Dynamics Models for Motor Temperature Prediction

Liao,

Chen,

Long

et al. 2023

Preprint

0

View full text Add to dashboard Cite

<p>The permanent magnet synchronous motor finds extensive use in industrial applications, and the development of effective thermal management solutions is crucial to enhance its power density. Accurate temperature prediction of the permanent magnet synchronous motor serves as the fundamental basis for designing effective thermal management strategies. Model-based prediction methods exhibit superior real-time performance, but the intricate modeling process requires substantial expert knowledge guidance and lacks versatility. Conversely, data-driven prediction methods, while offering flexibility, often lack physical implications in terms of system dynamics. This paper proposed a structured linear neural dynamics model for motor temperature prediction. This model is data-driven, with prior knowledge integrated into its structure, which preserves flexibility while guaranteeing system stability through the Perron-Frobenius theorem. Additionally, this paper achieves the decoupling of control input from state transitions and the embedded deployment of this model. The method is validated with a real dataset. The lightweight feature is demonstrated by the implementation of an STM32 Microcontroller with 1.808 KB and 27 mW. The paper is accompanied with the open source data and code at GitHub: https://github.com/ms140429/Explainable-Neural-Dynamics-Model</p>

show abstract

“…It highlights the challenges of accurate estimation and computational time, referencing prior research and various models. The table also provides quantitative data on simulation time and accuracy of 𝑇 𝑗 estimation[7],[15],[33]-[36],[18]-[22], [30]-[32].…”

mentioning

confidence: 99%

X-in-the-Loop Validation of Deep Learning-Based Virtual Sensing for Lifetime Estimation of Automotive Power Electronics Converters

Chakraborty,

Bhoi,

Hosseinabadi

et al. 2024

IEEE J. Emerg. Sel. Topics Power Electron.

0

View full text Add to dashboard Cite

Health degradation issues in automotive power electronics converter systems (PECs) are present due to repetitive thermomechanical stress endured while the vehicle is in real-field operation. This stress results from heat generation, a byproduct of semiconductor operation within PECs, leading to degradation in semiconductor operating life. The best practice in academia and industry is to rely on detailed Physics-of-Failure (PoF) based models for lifetime estimation. However, the PoF-based model of PECs requires substantial computational time and robust devices to estimate lifetime accurately. According to literature surveys, the computational time of the PoF-based models could be reduced further by using a low-fidelity and/or reduced-order model (ROM) that may result in unacceptable accuracy. To fulfill this research gap, this paper proposes a real-time executable, deep learningbased virtual sensing method that enables vehicle manufacturers to estimate the lifetime of the PECs onboard. This computationally efficient virtual sensing method has been integrated into an onboard vehicle validator edge (VVE). At the same time, multiple DL configurations are being explored, and optimization is performed on compositions, hyper-parameters, training, and testing datasets to obtain the best DL model. Finally, to demonstrate the feasibility and accuracy of the proposed method before its implementation within the complex VVE, an X-in-the-Loop (XiL) test is performed with vehicle frontloading. Index Terms-Virtual sensing, electric vehicles, power electronics converter, SiC power module, electro-thermal model, system-level lifetime, vehicle edge, and X-in-the-loop.

show abstract

Machine-Learning-Based Condition Monitoring of Power Electronics Modules in Modern Electric Drives

Cited by 5 publications

References 12 publications

Explainable Neural Dynamics Models for Motor Temperature Prediction

Explainable Neural Dynamics Models for Motor Temperature Prediction

Explainable Neural Dynamics Models for Motor Temperature Prediction

X-in-the-Loop Validation of Deep Learning-Based Virtual Sensing for Lifetime Estimation of Automotive Power Electronics Converters

Contact Info

Product

Resources

About