As one of the key components of smart grid, advanced metering infrastructure (AMI) provides an immense number of data, making technologies such as data mining more suitable for electricity theft detection. However, due to the unbalanced dataset in the field of electricity theft, many AI-based methods such as deep learning are prone to under-fitting. To evade this problem and to detect as many types of theft attacks as possible, an outlier detection method based on clustering and local outlier factor (LOF) is proposed in this study. We firstly analyze the load profiles with $k$ -means. Then, customers whose load profiles are far from the cluster centers are selected as outlier candidates. After that, the LOF is utilized to calculate the anomaly degrees of outlier candidates. Corresponding framework for practical application is then designed. Finally, numerical experiments based on realistic dataset show the good performance of the presented method.

With the emerging power electronic technologies, numerous electrical devices use high-frequency and nonsinusoidal supplies. Such sources generate complex harmonic fields that overburden the accurate prediction of core losses using the conventional finite element method (FEM). Thereby, to predict core losses under those high-frequency and nonsinusoidal supplies accurately, this study first built a test system for core losses measurement of ferromagnetic materials (e.g., soft ferrite, laminated steels, and nanocrystalline alloy) fed by the square and pulsewidth modulated (PWM) supplies. Then, the loss prediction problems of the classical core loss model under high-frequency and nonsinusoidal supplies are revealed. Afterward, an improved FEM-based core loss model is proposed, which includes a piecewise variable parameters based hysteresis loss model, an eddy current loss model considering different cross sections of magnetic circuit and supply waveforms, and improved excess loss coefficients. Experimental validations on different materials, the nanocrystalline alloy in an 11-kVA intermediate frequency transformer, the ultrathin silicon sheets in a 10-kW high-speed permanent magnet motor, and the nanocrystalline alloy in a 3-kW wireless power transfer system, are conducted systematically. The effectiveness and generality of the improved core loss model were verified.

This article proposes a rotor temperature estimation method for in-service induction machine (IM) based on parameter identification, which combines the advantage of recursive least squares (RLS) and model reference adaptive system (MRAS). The RLS with forgetting factor is firstly adopted to identify the parameters of motor inductances. Then, the online identification of rotor resistance can be realized via the MRAS based on the instantaneous reactive power (IRP) and the proportional integral (PI) regulation adaptive law designed by the Popov hyper-stable theory. Thereby, the rotor temperature is computed by the resistance-temperature relationship of metals. To achieve a fast convergence of the parameter identification, rotor slot harmonics are extracted from the stator current and used to determine the rotor speed in real time. Furthermore, to obtain a more accurate initial value of the stator and rotor leakage inductances and resistances, the first 5-15 cycles of the IM starting process are used to mimic the locked-rotor test condition. A merit of the proposed method is that it requires significantly less time to estimate the rotor resistance than the traditional methods. With a novel test bench, experimental validation is performed on a 22-kW IM. In the test, the rotor temperature is measured in three different ways: 1) wireless sensors inserted in the rotating rotor core and rotor end rings, 2) infrared sensor for rotor end ring temperature, and 3) PT100 installed in stator end winding. With this test bench, the real rotor temperature was revealed, and the effectiveness of the presented method is also verified.

For an inductive power transfer (IPT) system used in electric vehicle (EV) charging, the receiving coil is usually equipped with a shielding structure to reduce iron loss and electromagnetic interference in the chassis. However, the addition of a shielding structure may significantly reduce self-inductance and mutual inductance, and the eddy current on the shielding pad can also produce power loss. To maximize the efficiency of the IPT system used for EV charging and sufficient shielding, this study first analyzed the equivalent circuit model of the IPT system. A 3-D model of an IPT system, with a steel plate and aluminum shielding, is established and investigated in the finite-element analysis (FEA) domain, in which the shielding structure is modeled with some parametric sizes to simplify the modification of the parameters on a shielding structure. The simulation result shows how the ohmic loss on coils and the iron loss on steel plate changes after adding the shielding structure. Afterward, a practical method for optimizing aluminum shielding structure's parameters is proposed to maximize the efficiency of the IPT system, taking iron losses in a steel part and aluminum shielding into account. Furthermore, considering that the optimization procedure becomes complex since the different resonant capacitances are required due to the variable inductances, a new efficiency calculation method by FEA is presented to simplify the optimization procedure. Experimental validation is finally performed on a 3-kW wireless charging prototype; thus, the effectiveness of the proposed method is verified.

Renewable energy units are required to have reliable fault ride-through capability to prevent large-scale disconnection of renewable energy plants due to grid faults. Since synchronous generators have better voltage and current resistance capabilities, one of the advantages of renewable energy connected to the grid through the motor-generator Pair (MGP) system is the isolation of grid faults, which is totally different from existing fault ride-through methods. First, the structure and mathematical model of the MGP system are introduced in this paper. Then, a novel control method for the MGP system based on DC link voltage feedback is given, which is very important for fault ride-through in the control power balance on both sides of the DC bus. The isolation mechanism, damping and reactive power support functions of the MGP system on grid faults are analyzed by means of the rotor motion equation and phasor diagram of synchronous machine. A 5-kW MGP simulation and experimental platforms are built to verify the isolation and protection effects of the MGP system on a renewable energy generation system during the transient processes of grid voltage sag, overvoltage and multiple voltage sags. The results demonstrate that the MGP system can isolate grid faults on the generator side, while the oscillation amplitude on the motor side is much smaller, and the generator can provide strong reactive power support to the grid during three grid faults.