Li Liu

Power system stabilizers are widely used to damp out the low frequency oscillations in power systems. In power system control literature, there is a lack of stability analysis for proposed controller designs. This paper proposes a Neural Network (NN) based stabilizing controller design based on a sixth order single machine infinite bus power system model. The NN is used to compensate the complex nonlinear dynamics of power system. To speed up the learning process, an adaptive signal is introduced to the NN's weights updating rule. The NN can be directly used online without offline training process. Magnitude constraint of the activators is modeled as saturation nonlinearities and is included in the stability analysis. The proposed controller design is compared with Conventional Power System Stabilizers whose parameters are optimized by Particle Swarm Optimization. Simulation results demonstrate the effectiveness of the proposed controller design.

Power system islanding is an efficient way to avoid catastrophic wide area blackouts, such as the 2003 North American blackout. Islanding of large-scale power systems is a combinatorial explosion problem. Thus, it is very difficult to find an optimal solution within reasonable time using analytical methods. This paper presents a new method to solve this problem. In the proposed method, angle modulated particle swarm optimization (AMPSO) is utilized to find a number of efficient islanding solutions for large-scale power systems due to its computational efficiency. First, desired generators groups is obtained using slow coherency algorithm. AMPSO is then used to optimize a fitness function defined according to both generation/load balance and similarity to the desired generator grouping. In doing so, the resulted islanding solutions provide good static and dynamic stability. Simulations for power systems of different scales demonstrate the effectiveness of the proposed algorithm.

High performance application of permanent magnet synchronous motors (PMSM) is increasing. PMSM models with accurate parameters are significant for precise control system designs. Acquisition of these parameters during motor operations is a challenging task due to the inherent nonlinearity of motor dynamics. This paper proposes an intelligent model parameter identification method using particle swarm optimization (PSO) approach. As an intelligent computational method based on stochastic search, PSO is shown to be a versatile and efficient tool for this complicated engineering problem. Through both simulation and experiment, this paper verifies the effectiveness of the proposed method in identification of PMSM model parameters. Specifically, stator resistance and load torque disturbance are identified in this PMSM application. Though PMSM is discussed, the method is generally applicable to other types of electrical motors, and as well as other dynamic systems with nonlinear model structure.

Power system defensive islanding is an efficient way to avoid catastrophic wide area blackouts, such as the 2003 North American blackout. Finding defensive islands of large scale power systems is a combinatorial explosion problem. Thus, it is very difficult to find an optimal solution, if it exists, within reasonable time using analytical methods. This paper proposes utilizing the computational efficiency property of an angle modulated particle swarm optimization to find some efficient islanding solutions for large scale power systems. The solutions are referred to as optimal according to a fitness function considering the real power balance between generations and loads in islands, the relative importance of customers, and the desired number of islands. The algorithm can also provide necessary load shedding information. Furthermore, the algorithm can provide a number of candidate solutions for the check of transmission system capacity constraint. Simulations with power systems of different scales demonstrate the accuracy and effectiveness of the proposed algorithm.

It is envisioned that in the future a naval electric ship should be operated with minimum number of human operators, while having the intelligence to automatically reconfigure its power plant under various scenarios. This should be accomplished with less response time and adequate configurations when compared to the systems operated by humans. Multi agent systems have been shown as a promising technology to achieve this goal. In this paper, the authors present a study of various agent systems being developed for Navy shipboard power systems. A particular focus is given on agent based solutions for fault detection and reconfiguration of a shipboard power system. Test scenarios are presented for each agent based solution.

This paper presents a suite of neural network (NN) based decentralized controller designs for large scale power systems' generators, one is for the excitation control and the other is for the steam valve control. Though the control inputs are calculated using local signals, the transient and overall system stability can be guaranteed. NNs are used to approximate the unknown and/or imprecise dynamics of the local power system dynamics and the inter-connection terms, thus the requirements for exact system parameters are relaxed. Simulation studies with a three-machine power system demonstrate the effectiveness of the proposed controller designs.

This paper introduces a reconfigurable and flexible experimental footprint for validation of new development in power electronics and power systems research. The system contains six drives with six diode rectifiers, and six IGBT converters, which can be reconfigured as back to back converter or diode rectifier drive systems. The induction motor (IM) and the permanent magnetic synchronous motor (PMSM) in this system provide a platform for research on motor control, fault detection, and diagnosis (FDD), etc. The two thyrisor dc drives with two dc motors work as loads for IM and PMSM. The dc drives also work as nonlinear loads to support research on power quality. Together with the configured diode rectifier drives, one adjustable nonlinear load offers a platform for development of power quality control algorithms. Additionally, the system interacts with real time digital simulator (RTDSreg) for control algorithm verification in controller hardware in the loop (HIL) or power HIL tests. Hence, the entire system is flexible and reconfigurable for different research areas, e.g. motor control, FDD, active filtering, STATCOM and other power electronics or power system research. Three examples are presented in the area of FDD, STATCOM, and power quality to demonstrate the capability of this experimental footprint.

Permanent magnet synchronous motors (PMSM) are frequently used to high performance applications. Accurate diagnosis of incipient faults can significantly improve system availability and reliability. This paper proposes a new scheme for the automatic diagnosis of turn-to-turn short circuit faults in PMSM stator windings. Both the fault location and fault severity are diagnosed using a particle swarm optimization (PSO) algorithm. The performance of the motor under the fault conditions is simulated through lumped-parameter models. Waveforms of the machine phase currents are monitored, based on which a fitness function is formulated and PSO is used to identify the fault location and fault size. Simulation results in MATLAB provide preliminary verification of the diagnosis scheme.

Permanent magnet synchronous motors (PMSM) are frequently used to high performance applications. Accurate diagnosis of incipient faults can significantly improve system availability and reliability. This paper proposes a new scheme for the automatic diagnosis of turn-to-turn short circuit faults in PMSM stator windings. Both the fault location and fault severity are diagnosed using a particle swarm optimization (PSO) algorithm. The performance of the motor under the fault conditions is simulated through lumped-parameter models. Waveforms of the machine phase currents are monitored, based on which a fitness function is formulated and PSO is used to identify the fault location and fault size. Simulation results in MATLAB provide preliminary verification of the diagnosis scheme.