Search

When discussing the development of MAS for Smart Grid and the advantages of distributed MAS over centralized control in the text of this article [1] and describing MAS tools in Table 2, we inadvertently used text from a previous article by Merabet and colleagues [2] and omitted a citation

With the increasing integration of Distributed Energy Resources (DER) in the power grid, a decentralized approach becomes essential for scheduling and allocation of resources in a smart grid. Economic Dispatch (ED) and Unit Commitment (UC) are the two major resource allocation problems that play critical role in the safe and stable operation of a grid system. The uncertainty associated with renewable energy sources have made the resource allocation problems even more challenging for grid operators. The future grid will have a higher generation mix of renewable energy sources and a large load of Electrical vehicles, with the possibility of bi-directional power flow. This complex smart grid system necessitates the development of a decentralized approach to resource allocation problem, which allows inter-node communication and decision making. Multi-agent systems (MAS) is a promising platform to decentralize the traditional centralized resource allocation aspects of smart grid. This paper presents a comprehensive literature review on the application of MAS to Economic Dispatch (ED) and Unit Commitment (UC) in smart grids.

Forecasting of consumer electricity usages plays an important role to make total smart grid system more reliable. As the activities of individual residential consumers has many uncertain variables, it is hard to accurately forecast the residential load levels. For planning of the electrical resources and to balance demand and supply, accurate forecasting tasks are critical. This paper presents Deep Neural Network (DNN) based short term load forecasting for Residential consumers. In this work, we compare the Mean Absolute Percentage Error (MAPE) value for residential electricity dataset using different types recurrent neural network (RNN). Our preliminary results indicate that Long short-term memory (LSTM) based RNN performed better compared with simple RNN and gated recurrent unit (GRU) RNN for a single user with 1-minute resolution based on one year of historical data sets.

The paper optimizes the placement of Phasor Measurement Units (PMUs) in IEEE 13, 37, and 123 node test feeders. Considering its large cost of placing these PMU units, an optimal placement and evaluation of system observability is necessary. The PMU locations were derived using a code written in AMPL, which is based on a linear programming formulation. Indices such as System Observability Redundancy Index (SORI) and Optimal Redundancy Criterion (ORC) were used to evaluate the full-observability in three systems: IEEE 13, 37, and 123 node test feeders.

The placement of Phasor Measurement Units (PMUs) and Smart Meters have become key components for smart grids to monitor system behaviours in real-time. The optimal PMU placement is key for a smart grid system for its secure operation. This paper presents an object-oriented programming (OOP) approach combined with GLPK (GNU Linear Programming Kit), to solve for optimal number of PMUs for any grid size. This is done through an external library called Glpk.js, which is a JavaScript port of the GLPK. The code is made available as opensource for simulation and analysis of bus networks using a Graphical User Interface (GUI). The code is easily scalable to any size network for any sensor placement problem. The graphical approach makes it easier for any researcher or grid operator to analyse the system for optimal sensor placement locations.

The storage space and cost for Smart Grid datasets has been growing exponentially due to its high data-rate of various sensor readings from Automated Metering Infrastructure (AMI), and Phasor Measurement Units (PMUs). The paper focuses on Phasor Data Concentrators (PDCs) that aggregate data from PMUs. PMUs measure real-time voltage, current and frequency parameters across the electrical grid. A typical PDC can process data from anywhere ten to forty PMUs. The paper exploits the need for appropriate security and data compression challenges simultaneously. As a result, an optimal compression method ER1c is investigated for efficient storage of IREG and C37.118 timestamped PDC data sets. We expect that our approach can greatly reduce the storage cost requirements of commercial available PDCs (SEL 3373, GE Multilin P30) by 80%. For example, 2 years of PDC data storage space can be easily replaced with only 10 days of storage space. In addition, our approach in combination with AES 256 encryption can protect PDC data to larger degree as per National Institute of Standards and Technology (NIST) standards.

Synchrophasors are the state-of-the-art measuring devices that sense various parameters such as voltage, current, frequency, and other grid parameters with a high sampling rate. This paper presents an approach to visualize and analyze the smart-grid data generated by synchrophasors using a visualization tool and density based clustering technique. A MATLAB based circle representation tool is utilized to visualize the real-time phasor data generated by a smart-grid model that mimics a synchrophasor. A density based clustering technique is also used to cluster the phasor data with the aim to detect contingency situations such as bad-data classification, various fault types, deviation on frequency, voltage or current values for better situational alertness. The paper uses data from an IEEE fourteen bus system test-bed modeled in MATLAB/SIMULINK to aid system operators in carrying various predictive analytics, and decisions.

This paper explores several reliability indices to quantify the importance of buses in IEEE-118 and IEEE-300 bus test systems and determine which buses are key to operations of a simulated smart grid. A list of Critical Bus Indices (CBI) were formulated using principles of graph theory. Several parameters such as betweenness-centrality (BC), degree, demand, generation, and their combinations for each bus with in the test systems were explored. Efficiency of these indices in quantifying the importance of buses to the operation of the systems was validated by a node (bus) removal process. Node removal measures the disruption of the system under failure scenarios. System disruption was quantified by changes in normalized expected geodesic distance (NEGD) and normalized expected electric distance (NEED). Preliminary results provide promising insight in identifying the utility of betweenness-centrality in quantifying the importance of buses.

Load forecasting is an important electric utility task for planning resources in Smart grid. This function also aids in predicting the behavior of energy systems in reducing dynamic uncertainties. The efficiency of the entire grid operation depends on accurate load forecasting. This paper proposes and investigates the application of a multi-layered deep neural network to the Iberian electric market (MIBEL) forecasting task. Ninety days of energy demand data are used to train the proposed model. The ninety-day period is treated as a historical dataset to train and predict the demand for day-ahead markets. The network structure is implemented using Google's machine learning Tensor-flow platform. Various combinations of activation functions were tested to achieve a better Mean Absolute percentage error (MAPE) considering the weekday and weekend variations. The tested functions include Sigmoid, Rectifier linear unit (ReLU), and Exponential linear unit (ELU). The preliminary results are promising. and show significant savings in the MAPE values using the ELU function over the other activation functions.

Quantification of uncertainty due to wind-energy production becomes more and more crucial as the penetration of wind into smart grid increases. System operators (TSOs) and planners would be interested to see how wind production varies over different look-ahead hours and estimate the probability of those variations under several uncertain conditions. As wind is a stochastic source of generation, this paper provides a state-of-the-art literature review on the uncertainties related to wind-energy dispatch.

When a malfunction occurs in a smart-grid electricity-provisioning system, it is vitally important to quickly diagnose the problem and to take corrective action. The self-healing problem refers to the need to take action in near real time in order to reallocate power to minimize the disruption. To address this need, we present a collection of integer linear programming (ILP) models that are designed to identify the optimal combinations of supply sources, the demand sites for generators to serve, and the pathways along which the reallocated power should flow. The models explicitly support multiple time periods and the uncertainty associated with alternative sources such as wind power. Model solutions are evaluated using a simulator configured with multiple, intelligent, distributed software agents.

Because electric-grid sensor data originating from several sensors, such as the phasor measurement units (PMUs); intelligent relays; and the new installation of smart meters, Plug-in Hybrid Electric Vehicles (PHEV), or Gridable Vehicles (GV), are exponentially growing, the Smart Grid’s data-analytic platform has huge potential (generation, transmission, or distribution) and can play a significant role in the decision-making process for meaningful data interpretation in order to act promptly or to automate the grid process to avoid any failures or grid instability. This chapter focuses on identifying the variables of interest that are important for the electric grid that is embedded in distributed real-time data engines which will help with the system operators’ decision-support process. More specifically, the applicability and performance of the M5 model and J48 decision-tree machine-learning technique are investigated using real electric-grid data. We have presented how a decision-tree model, such as M5P, can support system operators in making effective decisions in the Smart Grid. Two sets of test data are used in this chapter; the first data set is taken from a 10-unit commitment with a 50,000 Gridable Vehicle, and the latter one analyzes weekly New York City (NYC) demand data from NYISO.

The chapter describes a resource-allocation problem in a smart-grid application that is formulated and solved as a binary integer-programming model. To handle power outages from the main distribution circuit, the Smart grid’s intelligent agents have to utilize and negotiate with distributed-energy resource agents that act on behalf of the grid’s local generators in order to negotiate power-supply purchases to satisfy shortages. We develop a model that can optimally assign these DERs to the available multiple regional utility areas (RUAs) or units that are experiencing power shortages. This type of allocation is a resource-assignment problem. The DERs in our model depict the behavior of power created with a wind turbine, solar generation, or other renewable generation units, and the region or area refers to a centralized distribution unit. The integer-programming approach is called Capacity-Based Iterative Binary Integer Linear Programming (C-IBILP). All simulation results are computed using the optimization tool box in MATLAB. Computation results exhibit very good performance for the problem instances tested and validate the assumptions made.

Creating an autonomous, self-healing electrical grid is one of the most important challenges facing electric-energy providers. Such a system, known as the “smart grid,” must interweave a multitude of systems, both software and hardware, in order to form a complete solution that is capable of meeting the requirements outlined by the United States Department of Energy (DOE). According to the DOE [LSC05], “It is a colossal task. But it is a task that must be done.”

This paper presents an application of clustering algorithms for streaming smart grid datasets. The authors use a framework that includes a combination of an openPDC and statistical software (R) platform to acquire Synchrophasor data to carry out the clustering process. The clustering operations on streaming synchrophasor data enables operators to detect any anomalies or outages for better decision making. This will reduce the outage and failure rates, and enhance situational awareness of the grid. We show the application of hierarchical clustering to organize the sensor data into clusters. Additionally, several distance metrics (cluster linkage methods) are tested for enhanced hierarchical cluster separation. Experimental results on parameters such as frequency, voltage, and current demonstrate the novelty and effectiveness of our application of hierarchical clustering. The results indicate that the hierarchical clustering with single linkage distance metric is a good choice for sudden surge or sag values. On the other hand, the average distance metric is less sensitive to outliers and can detect small deviations in parameters.

This paper presents a linear programming based solution for an Optimal Placement of Phasor Measurement Units (PMU) problem (OPP) under contingency situations. With the increasing usage of synchrophasors or PMUs for smart grid monitoring and control, the problem of PMU placement has become a major concern due to high installation costs. This paper uses a Linear Programming (ILP) approach to find optimal solutions. This paper shows that performance of the LP constraints and formulation makes the approach very attractive for smaller utilities with limited budgets. This is investigated with and without zero injection cases in the CPLEX solver. This paper also proposes an index called the Optimal Redundancy Criteria (ORC) that assists utilities in providing redundant observability for critical buses in the system. The proposed approach yields optimal solutions for full and redundant observabilities within 15 milliseconds (ms) for test bed cases such as the IEEE 14, 30, 57, 118, and 300 bus systems, and a 208 bus Southern region Indian power grid (SRIPG).

This paper presents a density based clustering (DBSCAN) technique to visualize and analyze the smart-grid data. The technique will aid in detecting bad-data, various fault types, deviation on frequency, voltage or current values for better situational awareness. Synchrophasors (or a PMU) is a sensor placed on a transmission line that tracks voltage, current, phase and frequency of the line. To improve situational awareness of the smart grid monitoring in real-time, the utility must monitor the phasor data measurement delivered by the sensors. Time-stamped synchronized measurements offer tremendous benefit for pre and post-event analysis. The paper uses data from openPDC framework to aid system operators in carrying various predictive analytics, decisions.

Synchrophasors are the state-of-the-art measuring sensors that sense voltage, current, or frequency with high data rate. This paper presents an approach to analyze the streaming smart-grid data generated by synchrophasors. A novel unit-circle representation is used to visualize the real-time phasor data. A Density based clustering (DBSCAN) method is proposed to cluster the phasor data to detect bad-data for classification, and identifying various fault anomalies. The paper uses datasets from an IEEE 14 bus system test-bed to aid system operators in carrying various predictive analytics and decisions.

Without a reliable metering and communication infrastructure, the smart grid could become a catastrophe to national security and economy. A true smart grid infrastructure should detect all existing and predict future threats through intrusion detection methods. Smart grids are susceptible to various physical and cyber-attack as a result of communication, control and computation vulnerabilities employed in the grid. The paper provides a comprehensive study on types of threats and solutions on smart grid communication and metering infrastructures. As a part of this survey, the smart grid metering infrastructures susceptibilities and recommended remedial actions are identified. In addition, the paper details types of known attacks on existing metering infrastructure and defensive methodologies.

As electric grid sensor data originating from several sensors such as the phasor measurement units (PMUs), intelligent relays, and new installation of smart meters, Plug-in Hybrid Electric Vehicles (PHEV) or Gridable Vehicles (GV), are exponentially growing, the data analytic platform for Smart Grid has huge potential (generation, transmission or distribution) and can play a significant role in the decision making process for meaningful data interpretation to act promptly or automate the grid process to avoid any failures or instability in the grid. This paper focuses on identifying the variables of interest that are important in the electric grid embedded in distributed real time data engines which will help decision support process for system operators. More specifically, the applicability and performance of M5 model and J 48 decision tree machine learning technique is investigated using the real electric grid data. We have presented how decision tree model such as M5P can support system operators in making effective decision in the Smart Grid. Two sets of test data are used in this paper; the first data set is taken from a 10 unit commitment with 50000 Gridable Vehicle and the latter analyzes a weekly New York City (NYC) demand data from NYISO.

When a malfunction occurs in a Smart Grid electricity provisioning system, it is vitally important to quickly diagnose the problem and take corrective action. The self-healing problem refers to the need to take action in near real time to reallocate power to minimize the disruption. To address this need, we present a collection of integer linear programming (ILP) models designed to identify optimal combinations of supply sources, demand sites for them to serve, and the pathways along which the reallocated power should flow. The models explicitly support the uncertainty associated with alternative sources such as wind power. A simulator configured with multiple intelligent distributed software agents has been developed to support the evaluation of the model solutions.

Electrical grids are highly complex and dynamic systems that can be unreliable, insecure, and inefficient in serving end consumers. The promise of Smart Grids lies in the architecting and developing of intelligent distributed and networked systems for automated monitoring and controlling of the grid to improve performance. We have designed an agent-oriented architecture for a simulation which can help in understanding Smart Grid issues and in identifying ways to improve the electrical grid. We focus primarily on the self-healing problem, which concerns methodologies for activating control solutions to take preventative actions or to handle problems after they occur. We present software design issues that must be considered in producing a system that is flexible, adaptable and scalable. Agent-based systems provide a paradigm for conceptualizing, designing, and implementing software systems. Agents are sophisticated computer programs that can act autonomously and communicate with each other across open and distributed environments. We present design issues that are appropriate in developing a Multi-agent System (MAS) for the grid. Our MAS is implemented in the Java Agent Development Framework (JADE). Our Smart Grid Simulation uses many types of agents to acquire and monitor data, support decision making, and represent devices, controls, alternative power sources, the environment, management functions, and user interfaces.