The void in solder layer is one of the most important aging form of Insulated Gate Bipolar Transistor (IGBT) module. In this paper, the influence mechanism of different void fractions on the thermal conductivity and specific heat capacity of the solder layer is studied. An improved Cauer model considering the void damage of the solder layer is established, and its model parameter extraction method is proposed. In addition, the temperature dependence of Cauer model and its parameter is discussed in this paper. In order to facilitate the experimental study of void damage in solder layer, we customized IGBT modules with different void fractions. The influence of void morphology on the thermal properties of solder layer is studied by finite element method, and the research purpose is to verify that void fraction can represent the effect of the voids on IGBT thermal characteristics. Considering the thermal resistance and thermal capacitance affected by the voids in solder layer, their calculation methods are given in detail. Then, under different loads, the IGBT modules with different void fractions are tested. The robustness of the improved Cauer model is verified by the experimental and numerical simulation results. The results show that the improved Cauer model has higher accuracy than the traditional Cauer model.

Bond wire aging is a universal failure form of silicon carbide metal-oxide-semiconductor field-effect transistor (SiC MOSFET) module. Real-time condition monitoring of bond wire is an important guarantee for the stable operation of power electronics system. In this paper, a method of monitoring the aging state of bond wire in SiC MOSFET module using on-state drain-source voltage (OSDSV) separation technique is proposed. Firstly, the temperature dependence of the conduction circuit resistance from drain terminal to source terminal in SiC MOSFET module is analyzed to obtain the temperature characteristics of the OSDSV. Secondly, the resistance of the package structure in SiC MOSFET module is extracted, and the OSDSV is divided into two parts: the package structure voltage and the chip OSDSV, and the analytical model for estimating the junction temperature of SiC MOSFET module in the healthy state is acquired. Finally, the package structure voltage is employed to monitor the aging degree of bond wire in SiC MOSFET module, and the flow chart of aging diagnosis of bond wire is established. The proposed method can accurately identify the failure state of bond wire lift-off. Theoretical and experimental results prove that the method is feasible.

Silicon carbide metal-oxide-semiconductor field-effect transistor (SiC MOSFET) has become a promising device due to its excellent material properties. Junction temperature is an important parameter and a significant health index. In this article, drain current during turn-on transient is taken as the thermo-sensitive electrical parameter (TSEP) to monitor the junction temperature of SiC MOSFET. Based on the physical properties of wide bandgap semiconductor materials, the switching behaviors of SiC MOSFET and the variation of drain current with temperature were studied. The influence of temperature dependence of threshold voltage and carrier mobility on drain current is analyzed. It is proven that the drain current has a positive temperature coefficient during turn-on transient. Finally, the validity of the proposed method is verified by the theoretical analysis and experiments.

A model for the online estimation of junction temperature based on the gate pre-threshold voltage in high-power insulated-gate bipolar transistors (IGBTs) was proposed. Firstly, the turn-on behavior and temperature dependence of the gate-emitter voltage during the turn-on delay period were studied. It was revealed that the turn-on waveform of the gate-emitter voltage exhibits a point of inflection after which the voltage between the gate and emitter increases significantly; at this point, the relationship between the voltage and the temperature changes from a positive correlation to a negative correlation. Meanwhile, Analyses on the basis of semiconductor physics revealed that the negative correlation between gate-emitter voltage and temperature is caused by the temperature dependence of the flatband voltage. Finally, a model for estimating the junction temperature was established based on a combination of the turn-on delay time and the gate pre-threshold voltage which is negatively correlated with the temperature, to eliminate the effect of the bus voltage.

Non-intrusive load monitoring (NILM) is an important research direction and development goal on the distribution side of smart grid, which can significantly improve the timeliness of demand side response and users’ awareness of load. Due to rapid development, deep learning becomes an effective way to optimize NILM. In this paper, we propose a novel load identification method based on long short term memory (LSTM) on deep learning. Sequence-to-point (seq2point) learning is introduced into LSTM. The innovative combination of the LSTM and the seq2point brings their respective advantages together, so that the proposed model can accurately identify the load in process of time series data. In this paper, we proved the feature of reducing identification error in the experimental data, from three datasets, UK-DALE dataset, REDD dataset, and REFIT dataset. In terms of mean absolute error (MAE), the three datasets have increased by 15%, 14%, and 18% respectively; in terms of normalized signal aggregate error (SAE), the three datasets have increased by 21%, 24%, and 30% respectively. Compared with the existing models, the proposed model has better accuracy and generalization in identifying three open source datasets.

Bond wire failure is a common failure that seriously affects the reliability of insulated-gate bipolar transistor (IGBT) modules. Real-time health monitoring for bond wires is very important to avoid catastrophic failures. In this article, we introduce a new health monitoring method for bond wires in IGBT modules based on the voltage ringing characteristics. We applied a double pulse circuit to generate voltage ringing in IGBTs and freewheeling diodes (FWDs) during turn off and reverse recovery phases. To monitor bond wire failure, we theoretically and experimentally validated the overshoot and oscillation frequency as indicators. In particular, we investigated the effects of the junction temperature and operating conditions on the overshoot and proposed their influence functions. We designed a measurement circuit to identify small changes in the overshoot. Finally, the health baselines for the overshoot under different conditions were experimentally verified. The novelty of the research lies in the accurate online monitoring for bond wire failure in IGBT modules without any external injection; the measurement circuit adopts the interrupt mode to save the controller resources. Our experimental results showed that after half the bond wire of the IGBT and the FWD failed, their voltage overshoot increased by nearly 5%.