A smart grid concept has been defined in recent years, which emphasizes the importance on smart protection and measurement devices, reliable data communication and high security, optimal energy management system, and fault detection, location, isolation and service restoration (FLISR) of distribution networks (DNs). The main objectives of the FLISR approach are to achieve fast fault processing time, reduce the minimum number of interrupted customers, and improve the power supply reliability of the distribution. The conventional FLISR approach is to use signals of fault indicators (FIs) with distribution network states. The discrete installation of FIs to switches or reclosers may slow the processing time of fault detection and location, so it is necessary to develop a more efficient FLISR approach for smart distribution networks using functions of feeder terminal units (FTUs). In this paper, pick-up and tripping signals of overcurrent (OC) relays in combination with distribution grid states (e.g., switching status of devices, loss of voltage…) sent from feeder terminal units (FTUs) are used to detect and locate different fault types. Fault isolation and service restoration of black-out areas are then performed by solving an objective function with two main constraints, including (i) restoring the possible maximum number of out-of-service loads; and (ii) limiting the minimum number of switching operation. Thirteen performance factors (PF) are used for the post-fault service restoration process, consisting of: (i) Power Flow Violations (PFV), (ii) Bus Voltage Violations (BVV), (iii) Total Operation Cost (TOC), (iv) Lost Power (LP), (v) Outage Customer (OC), (vi) Number of Switching Steps (NSS), (vii) Power Losses (LOSS); (viii) Customer Minutes Interruption (CMI), (ix) Load Minutes Interruption (LMI), (x) MAIFI, (xi) SAIFI, (xii) SAIDI, and (xiii) Protection Validation (PRV). E-Terra platform of a distribution management system (DMS) is used to implement the proposed FLISR approach. Simulation and experiment results from a real 22 kV distribution network are also analysed to validate this FLISR approach. As a result, the novel FLISR approach has the ability to identify effectively the over-reaching of OC relays, indicate a mis-coordination risk of adjacent protection devices on the same feeder, and get the total processing time of fault detection, location and isolation as well as ranking all possible service restoration plans in distribution network at less than two minutes.

Integration of distributed generators (DGs) into a distribution network (DN) can cause coordination challenges of overcurrent relays (OCRs) because of different fault-current contributions of DGs as well as the directional change in fault currents. Therefore, the OCRs should be properly coordinated to maintain their adaptability and scalability to protect the DG-integrated distribution network. In this study, an adaptive and scalable protection coordination (ASPC) approach has been developed for the OCRs in a DG-contained distribution network based on two implementation stages. At the first stage, the reliability improvement of fault-current calculation results is performed by determining the min-max confidence interval of fault current for each different fault type, which is the basis for properly selecting tripping and pick-up thresholds of definite-time and inverse-time OC functions in the same OCR. At the second stage, optimization algorithms are used for calculating protection-curve coefficients and Time-Dial Setting (TDS) multiplier for the inverse-time OC functions in the OCR. A real 22 kV DG-integrated distribution network which is simulated by ETAP software is considered a reliable test-bed to validate the proposed ASPC system of OCRs in the multiple-DG-contained distribution network. In addition, the coordination results of OCRs can be obtained by three common optimization algorithms, Particle Swarm Optimization (PSO), Gravitational Search Algorithm (GSA), and Genetic Algorithm (GA). These relay coordination results have shown an effective protection combination of the definite-time OC functions (50P and 50G) and the inverse-time OC functions (51P and 51G) in the same OCR to get the adaptable and scalable DN protection system.

Time-series and machine-learning methods are being strongly exploited to improve the accuracy of short-term load forecasting (STLF) results. In developing countries, power consumption behaviors could be suddenly changed by different customers, e.g. industrial customers, residential customers, so the load-demand dataset is often unstable. Therefore, reliability assessment of the load-demand dataset is obviously necessary for STLF models. Hence, this paper proposes a novel and unified statistical data-filtering method with the best confidence interval to eliminate unexpected noises/outliers of the input dataset before performing various short-term load forecasting models. This proposed novel data-filtering method, so-called the data pre-processing method, is also compared to other existing data-filtering methods (e.g. Kalman filter, Density-Based Spatial Clustering of Applications with Noise, Wavelet transform, and Singular Spectrum Analysis). By using an SCADA system-based database of a typical 22kV distribution network in Vietnam, NYISO database, and PJM-RTO database, case studies of short-term load forecasting have been conducted with a conventional ARIMA model, an ANN forecasting model, an LSTM-RNN model, an LSTM-CNN combined model, a deep auto-encoder (DAE) network, a Wavenet-based model, a Wavenet and LSTM hybrid model, and a Wavelet Neural Network (WNN) model, which are to validate the novel and unified statistical data-filtering method proposed. The achieved numerical results demonstrate which the accuracy of the aforementioned STLF models can be significantly improved due to the proposed statistical data-filtering method with the best confidence interval of the input load dataset. The proposed statistical data-filtering method can considerably outperform the existing data-filtering methods.