V.K. Sood

Fault identification in an AC-DC transmission system using neural networks

The authors explore the possibility of using neural networks to identify faults that can occur in an AC-DC power system. Three types of neural network models have been studied and are compared. These networks can sense AC bus voltages either as root mean square (RMS) values (with or without phase angle information) or as sampled instantaneous values of sine waves. Depending on which method is used, some confusion can occur in distinguishing a line to line fault from a remote AC fault. A delay of 1-2 cycles in detection of faults when using RMS values is expected due to the algorithm required for determining the RMS value. This may not be too critical in practice. However, where this delay is unacceptable, instantaneous values may be used. Based on the ability of these networks to distinguish reliably between different types of faults, appropriate control measures can be taken to improve the dynamic performance of the AC-DC power system. >

Application of a radial basis function (RBF) neural network for fault diagnosis in a HVDC system

The application of a radial basis function (RBF) neural network (NN) for fault diagnosis in an HVDC power system is presented in this paper. To provide a reliable pre-processed input to the RBF NN, a new pre-classifier is proposed. This pre-classifier consists of an adaptive filter (to track the proportional values of the fundamental and average components of the sensed system variables), and a signal conditioner which uses an expert knowledge base (KB) to aid the pre-classification of the signal. The proposed method of fault diagnosis is evaluated using simulations performed with the EMTP package.

A static compensator model for use with electromagnetic transients simulation programs

A static VAr compensator (SVC) model based on state variable techniques is presented. This model is capable of being interfaced to a parent (or host) electromagnetic transients program, and, in particular a stable method of interfacing to the EMTDC program is described. The model is primarily that of a thyristor-controlled reactor. (TCR) and a thyristor-switched capacitor (TSC). Capacitor switchings within the TSC have been handled in a novel way to simplify storage and computation time requirements. During thyristor switching, the child SVC model is capable of using a smaller timestep than the one used by the parent electromagnetic transients program; after the switching, the SVC model is capable of reverting back to a (larger) timestep compatible with the one used by the parent program. Other features considered include the modeling of a phase-locked-loop-based, valve firing system. An application of this model to the simulation of a SVC controlling the AC voltage of the inverter bus of a back-to-back HVDC (high-voltage direct current) tie is presented. >

Dynamic interactions between HVDC systems connected to AC buses in close proximity

As AC/DC systems become further integrated, multiple-infeed converter (MIC) systems can be expected to become more common. Different categories of MIC systems are defined, and: (1) the impact of the AC system strength; and (2) the coordination of the reactive power demand of the converters on the dynamic performance of such systems are discussed. Analog or physical simulator study results are presented which provide some insight into the dynamic behavior of one category of such systems. In relation to MIC systems which have a common AC bus or are connected to AC buses which are electrically close to each other, the following conclusions can be drawn: (1) careful evaluation of interactions between converters is necessary since even a remote fault affecting one DC system may affect other DC systems within the MIC system; (2) coordination of reactive powers at the common AC bus is necessary to avoid unnecessary mode shifts; and (3) coordination of the operating and control characteristics (i.e. VDCL characteristics) of converters within a MIC system is essential for satisfactory dynamic behavior of the total system. >

Monitoring HVDC systems using wavelet multi-resolution analysis

The paper presents a disturbance classification technique based on wavelet multiresolution analysis. The wavelet multiresolution transform is introduced as a tool for providing discriminative, translation-invariant features with small dimensions to classify different disturbances in an HVDC transmission system. The proposed method extracts features from signals monitored on both dc and ac sides of the HVDC system. It is shown that monitored signals show promising features that can classify different disturbances that may occur anywhere in the HVDC system.

Recurrent neural networks for phasor detection and adaptive identification in power system control and protection

A multi-input multi-output (MIMO) recurrent neural network (RNN) is used as a versatile tool for the high-speed phasor detection and the adaptive identification of control and protection signals in power systems. For the application as a phasor detector, a fast pseudo-gradient training is performed off-line to estimate the time-invariant weights of the RNN. This network is then operated in real-time, in recall mode only, to behave as a nonlinear fixed-coefficient filter. For the application as an adaptive identifier of nonlinear components, training is performed off-line for initializing the connection weights, but subsequently, they are continuously updated in real time. This results in an adaptive identifier suitable for detecting abrupt changes in complex nonlinear systems. Following an initial evaluation on synthetic signals, these two proposed RNNs are then validated using realistic waveforms generated from a series-compensated power system model.

Comparative evaluation of neural network based and PI current controllers for HVDC transmission

An investigation into a neural-network (NN) -based controller, comprising an NN trained offline in parallel with an NN trained online, is described. This NN controller has the potential of replacing the proportional-plus-integral (PI) controller traditionally used for HVDC (high-voltage direct-current) transmission systems. A simplified theoretical basis for the operational behavior of the individual NN controllers is presented. Comparisons between the responses obtained with the NN and PI controllers for the rectifier of an HDVC transmission system are made under typical system perturbation and faults. It is shown that the combined NN controller can adapt its weights online to provide improved or similar performance, when compared to traditional PI controllers, for small- and large-signal disturbances. The response of this simple NN controller is somewhat slower for very fast transients, perhaps due to the inadequate training. >

Investigation into an artificial neural network based on-line current controller for an HVDC transmission link

An artificial neural network (ANN) based current controller for a HVDC transmission link is described in this paper. Different ANN architectures and activation functions (AFs) are investigated for this ANN controller. Small (set current change) and large (DC-line fault) signal perturbations are applied to optimize the learning parameters for the controller. Performance evaluation of the ANN controller under noise conditions is studied. A comparison between a traditional PI and the proposed ANN controller is made for various system contingencies and it is shown that the latter has many attractive features.

Intelligent current controller for an HVDC transmission link

This paper describes an intelligent current controller for the fast and flexible control of an HVDC transmission link using artificial neural network (ANN) and fuzzy logic (FL) paradigms. A simple yet effective ANN architecture is presented with on-line adaptation of the activation function and learning parameters. Two methods of adapting the learning parameters are presented. In the first method, a heuristic approach to evaluate the learning rate as a polynomial of an energy function is considered. In the second method, a FL based on-line adaptation of the learning parameters is discussed. Performance of ANN, ANN-FL based and PI controllers are compared. A feasibility analysis is carried out to implement the proposed neural controller algorithm in real-time.

EMTP modelling of CIGRE benchmark based HVDC transmission system operating with weak AC systems

An EMTP based study of a CIGRE benchmark based HVDC system operating with weak AC systems is carried out. The modelled system provides a starting point for (a) educators teaching HVDC transmission courses, and (b) for utility planners to develop their own low-cost dedicated digital simulators for training purposes. In this paper, modelling details of the AC-DC system, DC converters and controls are presented. To validate the control schemes presented, the HVDC system is tested under AC-DC fault conditions. Results obtained from an EMTP-based study under these fault conditions are also presented in this paper.