Heat-Flux-Based Condition Monitoring of Multichip Power Modules Using a Two-Stage Neural Network

Hu, Borong; Hu, Zedong; Li, Ran; Ng, Chong; Jia, Chun‐Jiang; McKeever, Paul; Tavner, Peter; Zhang, Chi; Jiang, Huaping; Mawby, Philip

doi:10.1109/tpel.2020.3045604

Cited by 21 publications

(11 citation statements)

References 23 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In turn it also implies that some of the heat generated by the device will be transversely diffused to the others. However, given the small thickness of the materials and components inside a wire bonded power module (between the measurement points), lateral heat transfer is minor [24] and will have only secondary effect on the results of condition monitoring.…”

Section: A Generic Modelmentioning

confidence: 99%

“…By exploiting the characteristics of the power loss depending on temperature, a bespoke iterative algorithm is developed to compute the junction temperature under a known thermal resistance increment ∆Rth20. As the solder layer aging can be emulated by inserting thermal conductive pad (thermal interference material, TIM) between the baseplate and heatsink [22]- [24], then a series of ∆Rth20 values can be used in calibration (as detailed in Section IV). Then the bespoke iterative algorithm can find the device junction temperature and power loss to satisfy the measured silicone gel and heatsink temperatures.…”

Section: A Calculation Of Condition Indicatormentioning

confidence: 99%

“…Look-up Tables In calibration, the change of thermal resistance can be known by inserting thermal conductive pads under the baseplate as stated above [22]- [24]. Another way is to use preaged power modules with thermal resistance measured off-line [11].…”

Section: Establishingmentioning

confidence: 99%

“…It is still necessary to assume that the heat dissipation through the case equals the total power loss. As Hu et al stated in [24] that large errors could occur if other heat dissipation paths were ignored. They used a neural network as the multiple-input and multiple-output (MIMO) thermal model based on external measurements.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Monitoring Power Module Solder Degradation From Heat Dissipation in Two Opposite Directions

et al. 2022

IEEE Trans. Power Electron.

Self Cite

View full text Add to dashboard Cite

show abstract

Section: A Generic Modelmentioning

confidence: 99%

Section: A Calculation Of Condition Indicatormentioning

confidence: 99%

Section: Establishingmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Monitoring Power Module Solder Degradation From Heat Dissipation in Two Opposite Directions

et al. 2022

IEEE Trans. Power Electron.

Self Cite

View full text Add to dashboard Cite

show abstract

“…The training process includes an optimization approach to find the optimal weight. In this paper, the network is trained by Levenberg-Marquardt back-propagation algorithm [31]. The training consists of four steps: the feedforward computation, the back-propagation to the output layer, the back-propagation to the hidden layer, and the updating weights.…”

Section: B Training Of the Networkmentioning

confidence: 99%

Artificial Neural Networks-Based Multi-Objective Design Methodology for Wide-Bandgap Power Electronics Converters

Rajamony

Wang

Calderon-Lopez

et al. 2022

IEEE Open J. Power Electron.

View full text Add to dashboard Cite

Design methodology of power electronics converters is critical to fully explore the potential of wide-bandgap power semiconductors at the converter level. However, existing design methods largely rely on complex mathematical models which significantly increases the computational time, complexity and further leads to problems including poor constraint handling capabilities, inaccurate design, difficult parameter tuning and inadequate problem dimension. These all could generate suboptimal designs that make the whole design process meaningless. To overcome the aforementioned problems, in this paper, an artificial neural network (ANN)-based multi-objective design approach is proposed, which offers significant advantages in reducing the repetitive usage of complex mathematical models and hence the computational time of design. The computational time was reduced by up to around 78% and 67% compared to the numerical modeling and geometric program (GP) methods as validated through a hardware design process. The proposed method was implemented in MATLAB/Simulink to design a 1kW single-phase inverter, resulting in a design with an optimized efficiency (98.4%) and power density (4.57kW/dm 3 ). The accuracy of the design is verified through experimental prototyping and the measured efficiency and power density are 98.02% and 4.54kW/dm 3 , respectively, so the errors of efficiency and power density are both less than 1%.

show abstract

A survey on machine and deep learning in semiconductor industry: methods, opportunities, and challenges

Huang,

Meng,

Huang

2023

Cluster Comput

View full text Add to dashboard Cite

Heat-Flux-Based Condition Monitoring of Multichip Power Modules Using a Two-Stage Neural Network

Cited by 21 publications

References 23 publications

Monitoring Power Module Solder Degradation From Heat Dissipation in Two Opposite Directions

Monitoring Power Module Solder Degradation From Heat Dissipation in Two Opposite Directions

Artificial Neural Networks-Based Multi-Objective Design Methodology for Wide-Bandgap Power Electronics Converters

A survey on machine and deep learning in semiconductor industry: methods, opportunities, and challenges

Contact Info

Product

Resources

About