Chris Diduch

Recent developments in islanding detection for distributed power generation

The aim of this paper is to review recent developments in anti-islanding techniques for distributed resources interconnected with electric power systems. The techniques may be classified as remote and local. Remote techniques are associated with island detection on the utility side while local techniques are associated with island detection on the distributed power generation side. Local techniques may further be classified as passive and active.

Islanding Detection Using Proportional Power Spectral Density

Islanding detection and prevention is a mandatory requirement for grid-connected distributed-generation systems. In this paper, we propose an antiislanding technique based on the proportional power spectral density of the voltage period at the point of common coupling (PCC). The concept of proportional power spectral density (PPSD) is introduced as a normalized measure that can be used for islanding detection. We show that if the voltage period measured at the PCC is filtered and used to set the command period of the inverter, then an islanding condition results in a distinct PPSD. The technique is implemented and demonstrated using a 4-kW inverter and verified in simulation.

A New Adaptive Logic Phase-Shift Algorithm for Anti-Islanding Protections in Inverter-Based DG Systems

Recent developments in anti-islanding techniques have demonstrated that phase shift techniques are very effective for anti-islanding protections in inverter-based distributed generation (DG) systems. Several approaches have been proposed in the previous researches such as slide-mode frequency shift (SMS), active frequency drift (AFD), and active frequency drift with positive feedback (AFDPF) etc. The automatic phase shift (APS) method is actually a modified SMS method. It can effectively reduce the non-detection zone (NDZ) of the SMS technique by introducing an additional phase shift increment each time the frequency of the terminal voltage stabilizes. However, it is very difficult to determine a stable islanding frequency, and the APS algorithm sometimes acts slowly, even fails in certain load conditions. A new adaptive logic phase-shift (ALPS) algorithm is proposed in this paper to regulate the additional phase shift at a suspicious islanding situation and evaluate the effects of the phase shift. This algorithm can yield a quick phase shift in an islanding situation yet only produce a very small phase shift when the grid is available for inverter-based DG systems. Both simulation and experiment results have proved the robustness and effectiveness of the newly proposed anti-islanding algorithm

A New Diagnostic Model for Identifying Parametric Faults

This paper presents a new approach to failure detection and isolation (FDI) for systems modeled as an interconnection of subsystems that are each subject to parametric faults. This paper develops the concept of a diagnostic model and the concept of a fault emulator which are used to model and parameterize subsystem faults. There are two stages to the FDI scheme. In the first stage there is a requirement to identify the diagnostic model. Once identified, the diagnostic model is used in the second stage to generate a residual. Artifacts within the measured residual are then used as a basis for identifying parametric faults. The scheme is distinct from others as it does not require an online recursive least squares type identifier.

A New Hybrid Anti-Islanding Algorithm in Grid Connected Three-Phase Inverter System

Passive and active methods are two major categories of anti-islanding approaches which are being widely used in grid-connected distributed generation (DG) systems. Passive anti-islanding techniques have no negative impact on the inverters performance however large non-detection zone is the shortcoming of these techniques. On the other hand, active approaches have smaller non-detection zone, but these active approaches inevitably have negative impact on inverters performance. To solve these problems, a new hybrid of both passive and active anti-islanding algorithm is proposed in this paper. A covariance index is used as a passive indicator to activate an active anti-islanding action, adaptive reactive power shift action, which can intelligently vary the output reactive power of the DG system to realize the anti-islanding protections. Both simulation and experimental results have demonstrated that this new algorithm can provide a fast anti-islanding protection while assure the zero or the least perturbation in inverters grid-connected operation.

Autonomous control systems: Monitoring, diagnosis, and tuning

A systematic and unified approach which accomplishes performance monitoring, performance improvement and fault prediction in control systems is proposed. The feature vector which is a vector formed of the coefficients of the estimate of the sensitivity function and the influence matrix which is the Jacobian of the feature vector with respect to the physical parameter are shown to contain the relevant information to realize an autonomous control system. The feature vector is estimated using a robust, accurate and reliable linear predictive coding algorithm. The influence matrix is computed by perturbing the physical parameters one at a time and estimating the feature vectors for each case. The proposed scheme is evaluated both on simulated as well as on actual control systems.

Apparent Power-Based Anti-Islanding Protection for Distributed Cogeneration Systems

In this paper, the performance of a passive anti-islanding method is experimentally tested for three phase (3φ) cogeneration systems. The tested method is based on determining the wavelet packet transform (WPT) high-frequency subbands present in the d-q-axis components of instantaneous 3q apparent powers (sd and sq), when evaluated at the point of common coupling (PCC). This passive anti-islanding method is founded based on the nature of instantaneous 3φ apparent powers that have components continuously exchanged between both sides of the PCC. An islanding condition can be considered as a transient disturbance that creates nonperiodic and nonstationary high frequency components in sd and sq. These frequency components can be parameterized by WPT high-frequency subbands, which can provide accurate detection of the islanding condition. The d-q WPT-based anti-islanding method is tested for a 3φ cogeneration system under various loading and power delivery conditions. Performance results reveal accurate, fast, and reliable detection and response to the islanding condition.

Aggregated domestic electric water heater control - building on smart grid infrastructure

A pilot project is underway in New Brunswick, Canada, to demonstrate direct control of thermostatically controlled loads as a means of providing ancillary services. This paper will focus on one of the load classes, aggregated control of domestic electric water heaters (DEWHs) as a means of shifting load profiles. The approach not only strives to satisfy utility level objectives by curtailing individual DEWHs in a strategic manner but also takes account of the thermal dynamics of individual heaters and estimates of hot water usage in an attempt to maintain customer comfort and satisfaction. The aggregated load controller is unique in its topology and operation, consisting of an individual load extractor, a water demand model, and a temperature estimator for each DEWH, together with a reserve capacity forecast unit and aggregated load control logic. Details of the aggregated load controller are developed, the relationship to the system operator is described and a summary of preliminary results is provided.

Identification and Estimation for Electric Water Heaters in Direct Load Control Programs

PowerShift-Atlantic (PSA) is a pilot project lead by Canadian Maritime utilities that demonstrates direct load control strategies for up to 20 MW of commercial and residential loads for the purpose of balancing the intermittency of renewable generation and supporting demand-side management programs. On the residential side, domestic electric water heaters (DEWHs) form a significant end use class. The ability to accurately estimate and predict the state of individual end use devices allows aggregated control systems to better ensure end-use performance and comfort levels. This paper presents a methodology for estimating and predicting the state of individual DEWHs from models of their thermodynamics and water consumption that are derived under two scenarios: 1) when measurements of both power consumption and water temperature are available; and 2) when only measurements of power consumption are available. The proposed methodology was implemented as part of the PSA pilot project for the DEWH load class to simulate the behavior of the load in presence of the controller and evaluate the performance of the controller. Experimental results show that the model and water usage profile mimic the actual behavior of DEWHs, and can predict the future power consumption when the thermostatic control of a DEWH is interrupted as part of a load control strategy.

Passive Islanding Detection Approach Based on Tracking the Frequency-Dependent Impedance Change

This paper presents a passive islanding detection approach based on measurements of the frequency-dependent impedance at the point of common coupling. The approach exploits the presence of harmonics injected by the electric power system (EPS) and harmonics injected by the distributed generator (DG). Models of impedance are developed to characterize the interconnection topology among the DG, the EPS, and the local load. The interconnection topology changes when islanding occurs resulting in a frequency-dependent change of the impedance as seen by the DG. Islanding is detected when certain changes in the impedance magnitude are detected at certain harmonics. The approach is new: 1) an impedance-based analytic model is derived from the interconnection topology and this may be used as a basis for feature extraction and threshold selection; 2) a passive algorithm is proposed as an enhancement to the under/overfrequency (UF/OF) method; 3) a grid impedance parameter space is derived to characterize the detection zone; and 4) a local load parameter space is used to compare the nondetection zone (NDZ) of the proposed method with that of the UF/OF method without the enhancement. The proposed approach is shown to reduce the NDZ when combined with the UF/OF method.