This paper proposes a reduced-order small-signal closed-loop transfer function model based on Jordan continued fraction expansion to assess the dynamic characteristics of the droop-controlled inverter and provide the preprocessing method for the real-time simulation of power systems. Firstly, dynamic phasors, time delay and zero-order hold are embedded into the small-signal model at the same time, then the closed-loop transfer function of the droop-controlled inverter is built. Compared with the existing closed-loop transfer function approaches, the accuracy of the built transfer function model is dramatically enhanced. Meanwhile, the inner cascaded voltage/current controller parameters are also designed. In order to directly obtain and preserve the maximum overshoot and settling time, which are main features to evaluate the system input-output dynamic response characteristics, the reduced second order closed-loop transfer function is proposed through the continued-fraction expansion regarding arbitrary points on the real frequency axis. Therein, this second order closed-loop transfer function with dynamic response of the original inverter is reduced to the lowest order. Furthermore, combined with the impedance-based approach, the proposed stability assessment approach is utilized to analyze the stability of the microgrid with multiple converters. Finally, simulations and experimental results demonstrate the convenience and accuracy of the proposed approach.

In this article, an online adaptive optimal control algorithm based on adaptive dynamic programming is developed to solve the multiplayer nonzero-sum game (MP-NZSG) for discrete-time unknown nonlinear systems. First, a model-free coupled globalized dual-heuristic dynamic programming (GDHP) structure is designed to solve the MP-NZSG problem, in which there is no model network or identifier. Second, in order to relax the requirement of systems dynamics, an online adaptive learning algorithm is developed to solve the Hamilton-Jacobi equation using the system states of two adjacent time steps. Third, a series of critic networks and action networks are used to approximate value functions and optimal policies for all players. All the neural network (NN) weights are updated online based on real-time system states. Fourth, the uniformly ultimate boundedness analysis of the NN approximation errors is proved based on the Lyapunov approach. Finally, simulation results are given to demonstrate the effectiveness of the developed scheme.

Bond wire degradation is one of the most common failure modes for wire-welded packaging insulated-gate bipolar transistor (IGBT) modules. This article proposes a novel bond wire detection method based on IGBT turn-on gate voltage overshoot. The degree of bond wires lift-off will change the stray inductance of the gate charge loop circuit, and therefore has a strong influence on turn-on gate voltage overshoot before miller platform, which can be used as an effective bond wire fault indicator. A double pulse test platform is built to verify the resolution and sensitivity of the proposed precursor for bond wire degradation. The dependences on bus voltages, load currents, junction temperatures are discussed and tested. The experimental results show that the gate turn-on voltage overshoot of IGBT increases with the increasing severity of bond wire faults, which agrees with the theoretical analysis. The proposed turn-on gate voltage overshoot-based method has a high sensitivity for identifying bond wire fault at the incipient stage. It offers an effective technique for detecting bond wire degradation for practical applications.

Insulated gate bipolar transistor (IGBT) module is the most widely used power electronic device in converters. Condition monitoring of IGBT is critical for avoiding sudden failures. Health management of converters based on accurate IGBT junction temperature is of great importance to reliable operation. However, the existing monitoring methods have some drawbacks, including low feasibility of online implementation, intrusiveness, and slow response. IGBT junction temperature has a strong influence on IGBT on-state voltage drop. A novel online IGBT junction temperature measurement method based on the on-state voltage drop is proposed in this paper. This technique can monitor the on-state voltage drop of IGBT and extract IGBT junction temperature online. Besides, the influences of the measurement circuit temperature variations and IGBT load current variations are considered and compensated based on the off-state stage of the IGBT. Simulation and experimental results validate the feasibility of the proposed technique. The proposed method has advantages of fast response, low cost, and simple circuit structure for non-invasive online IGBT junction temperature monitoring.

In this paper, the optimal control scheme for ice-storage air conditioning (IAC) system is solved via a data-based adaptive dynamic programming (ADP) method. It is the first time that adaptive dynamic programming is employed to design a self-learning scheme which obtains the optimal control policy of ice-storage air conditioning system. First, based on the data of the temperature, irradiance and cooling load in an actual project, a prediction model of cooling load is built by a three-layer neural network with the performance verification. Second, the operation of the ice-storage air conditioning system is analyzed. Third, a data-based adaptive dynamic programming method is designed to realize a self-learning optimal control for the ice-storage air conditioning system. Then, numerical results show that using the data-based optimal control method can reduce the operation costs. Finally, the comparison results show that the developed adaptive dynamic programming method improves the system efficiency, minimizing the overall cost. Thus, the superiority of the developed algorithm is verified.

Aged cable segments in underground power distribution systems will gradually develop into severe faults. The location of aged cable segment is essential for ensuring system reliability and reducing maintenance cost, but is rarely studied by the existing research. In this article, a novel aged cable segment detection and location method based on wide frequency range transfer function (TF) measurement and deep learning approaches is proposed to fill the gap. The relationship between the measured TF and cable ageing parameters is analyzed, and two models with combined sparse autoencoder and convolutional neural network are trained to estimate the aging location and severity. A transfer learning approach is further developed to improve the model performance for various online situations. Experiments are conducted which prove the effectiveness and generalization ability of the proposed method. Compared with existing approaches, the proposed method shows better performance under different ageing situations, even smooth ageing scenarios.