Real-Time Deep Learning at the Edge for Scalable Reliability Modeling of Si-MOSFET Power Electronics Converters

Abstract

With the significant growth of advanced high-frequency power converters, on-line monitoring and active reliability assessment of power electronic devices are extremely crucial. This article presents a transformative approach, named Deep Learning Reliability Awareness of Converters at the Edge (Deep RACE), for real-time reliability modeling and prediction of high-frequency MOSFET power electronic converters. Deep RACE offers a holistic solution which comprises algorithm advances, and full system integration (from the cloud down to the edge node) to create a near real-time reliability awareness. On the algorithm side, this paper proposes a deep learning algorithmic solution based on stacked LSTM for collective reliability training and inference across collective MOSFET converters based on device resistance changes. Deep RACE also proposes an integrative edge-to-cloud solution to offer a scalable decentralized devices-specific reliability monitoring, awareness, and modeling. The MOSFET convertors are IoT devices which have been empowered with edge real-time deep learning processing capabilities. The proposed Deep RACE solution has been prototyped and implemented through learning from MOSFET data set provided by NASA. Our experimental results show an average miss prediction of 8.9% over five different devices which is a much higher accuracy compared to well-known classical approaches (Kalman Filter, and Particle Filter). Deep RACE only requires 26ms processing time and 1.87W computing power on Edge IoT device. components in the energy conversion process. Therefore, understanding, modeling and predicting the reliability models of the power converters are crucial for enabling emerging technologies and future applications such as electric vehicles, smart grids, and renewable energy.Mathematically formulating and precise understanding of the physical degradation in high-frequency power converters is notoriously difficult, due to the system sophistication and many unknown non-deterministic variables. To solve this problem, a wide range of stochastical diagnostic and prognostics techniques have been proposed to address the reliability issues of a complex system in the design, fabrication, and maintenance process. The evaluation of these processes is beneficial to enable power convertors health management systems and resiliency for useful life estimation and reducing the risk of failures [2,3]. Kalman filter and Bayesian calibration are two examples of classical time series modeling and prediction techniques. However, these approaches are often bounded to first-order models in isolation and are not able to bring the collective behavior of many devices with the same underlying physic to create an accurate algorithmic construct. Therefore, their prediction accuracy is very limited. Moreover, they have very limited scalability for emerging advanced technologies [4,5].Recent advances in deep learning open a new horizon toward smart and autonomous systems. Deep learning offers a scalable data-driven discriminative paradigm to un...

Digital Twins for Wind Energy Conversion Systems: A Literature Review of Potential Modelling Techniques Focused on Model Fidelity and Computational Load

“…A digital twin with real-time simulation capabilities for health and reliability monitoring of power electronic converters was realized in [139,140] respectively. In [140], a data-driven deep learning procedure was used to develop a Long Short-Term Memory (LSTM) surrogate model.…”

“…A digital twin with real-time simulation capabilities for health and reliability monitoring of power electronic converters was realized in [139,140] respectively. In [140], a data-driven deep learning procedure was used to develop a Long Short-Term Memory (LSTM) surrogate model. In [141], an online condition monitoring technique was presented for wind turbine converters based on a physics-based model of the thermal time constants of the cooling system.…”

Machine Learning Systems for Intelligent Services in the IoT: A Survey

“…In Figure 3b), the simplest hybrid scenario, a cloud-edge architecture trains models in the cloud and downloads trained models onto edge servers which perform local inference with reduced latency, on data offloaded from nearby devices. Baharani et al [14] use this type of architecture to perform model training in the cloud and inference on an edge server. By reducing the data and energy footprint required for inference, computation can be moved downstream onto devices with reduced processing capacity, thus further reducing latency [153] [175].…”

Neural Network and Fuzzy Control Based 11-Level Cascaded Inverter Operation

“…In [3,4], it is found that the total life cycle cost of the standalone microgrid system is decreased by maintaining the PQ with the integration of HES. In [5,6], different MPP techniques are proposed to track maximum power from the HES. However, looking at the complexity, the conventional MPP techniques lag to track maximum power.…”