Several researches have focused on increasing the accuracy of load forecasting techniques in the last few decades. Among thousands of load forecasting literature, some are mentioned in this section. Statistical models that have been used for the short-term load forecasting (STLF) include multiple regression, exponential smoothing, iterative re-weighted least squares, autoregressive moving average (ARMA), kalman filtering, the Box and Jenkins method, spectral expansion technique, and time-series methods are found in several literature [2, 3]. Soft computing models (SCMs) are well known for their capabilities when dealing with non-linear systems and have garnered significant attention in the area of load forecasting. Thus, SCMs, evolutionary programming, and hybrid intelligent algorithm show improved load forecasting accuracy . Among SCMs, the backpropagation neural network (BPNN) is widely used for STLF due to its high forecastingperformance . Radial basis function neural network (RBFNN) also shows good STLF performance as it is easy to train, computationally fast, and more general approximator compared to other NNs . Support vector machines (SVMs) are also widely used in load forecasting . In [8, 9], a daily load forecasting model was developed using a chaotic time-series derived from power load demand curves. Furthermore, the combination of NN and fuzzy, such as adaptive neuro-fuzzy inference system (ANFIS) shows significant improved forecasting accuracy. Since ANFIS has a capability of using the expert knowledge of fuzzy system, one can model the complicated relationship between social/environmental factors with the hourly load pattern in an area, which is difficult to find in only NNs . Hybrid methods, such as a combination of RBF and genetic algorithm (GA), wavelet transform (WT) and autoregressive (AR), WT and NNs are also applied for alleviating STLF accuracy [9-11]. Another SCM model based on fuzzy ARTMAP (FA) is a relatively new concept for forecastingapplications including load forecasting  and wind speed forecasting .
Real-time microgrid control could take the form of the architecture outlined in , where customer preferences, measurements, and forecasting information is used to set microgrid operation to achieve maximum economic optimization. This architecture takes measurements at fixed time steps, runs an optimization algorithm to determine the ideal settings, and sends setpoints and commands back to the local controllers for customer loads and distributed generation resources. The smaller the time step used, the better the overall optimization, at the expense of higher computation resources . If such a system used a TSN based communication system, communication latency could be virtually eliminated as a bottleneck to microgrid control and state estimation. The limit of microgrid control would then be how fast the optimization algorithm could calculate the ideal state for the microgrid at any moment. This would allow microgrid operation to approach ideal real-time operation. Pairing this communication capability with Advanced Metering Infrastructure (AMI) would allow real-time monitoring and control of the microgrid on a house-by-house basis , which becomes more important as residential renewables continue to become adopted. AMI also potentially allows the real-time control of residential/industrial loads to manage energy consumption within the microgrid.
Smart grids with an intensive penetration of distributed energy resources will play an important role in future power system scenarios. The intermittent nature of renewable energy sources brings new chal- lenges, requiring an efﬁcient management of those sources. Additional storage resources can be beneﬁ- cially used to address this problem; the massive use of electric vehicles, particularly of vehicle-to-grid (usually referred as gridable vehicles or V2G), becomes a very relevant issue. This paper addresses the impact of Electric Vehicles (EVs) in system operation costs and in powerdemand curve for a distribution network with large penetration of Distributed Generation (DG) units. An efﬁcient management method- ology for EVs charging and discharging is proposed, considering a multi-objective optimization problem. The main goals of the proposed methodology are: to minimize the system operation costs and to mini- mize the difference between the minimum and maximum system demand (leveling the powerdemand curve). The proposed methodology perform the day-ahead scheduling of distributed energy resources in a distribution network with high penetration of DG and a large number of electric vehicles. It is used a 32- bus distribution network in the case study section considering different scenarios of EVs penetration to analyze their impact in the network and in the other energy resources management.
Remark 1. While studying real-life examples of observation functions it is clear most of them have strong seasonality behaviour, like with one described in Example 1. Nevertheless the pikes might occur at the neighbouring time segments in different years (as discussed in Example 2 part 1 below), which makes them far from being periodic in the classical sense.
Voltage fluctuations are changes or swings of the voltage envelope in a systematic manner or a series of random voltage variations and are always referred to as voltage flicker. In addition to its effect on light, it is responsible for reduced life of electronic, incandescent, fluorescent and cathode ray tubes, malfunction of phase locked–loops PLLs, mis-operation of electronic controllers and protection devices. Even under deregulation, voltage fluctuations and flicker will increase due to use of nonlinear and vulnerable devices, and in such environment, the control of the voltage fluctuation should be the responsibility of the Transmission utilities (Transco.) and Distribution Companies (Disco.) such as switch-mode power supplies, television sets, light dimmers, and adjustable-speed drives can also inject provided that industrial customers, especially those utilizing large arc furnaces, control the amount of fluctuation of their load current. .
The general aim of this article is to present genetic algorithms as a tool, that can be used in de- mand forecasting in internet shops. First part of article identities factors, which have to be taken into consideration during analysing demand in internet shops, e.g. dispersion of demand, delivery time in- fluence and different e-marketing factors. Specific form of used demand function is shown in the next section of the article. Then genetic algorithm is defined by its genetic operators acting on bit strings (examples of the operators are: crossover, inversion, and mutation) and its method of credit allocation (fitness evaluation and selection). Next the method of identification of the function parameters using genetic algorithms is shown. The next part of article shows appliance of presented genetic algorithm. The advantages and disadvantages of proposed method are shortly discussed in summary.
The effectiveness of active power filter depends on accurate extraction of fundamental component of current waveform and fastness of control strategy. The SAPF consists of a DC-bus capacitor, power electronic devices and coupling inductors (L). Shunt APF acts as a current source for compensating the harmonic currents due to nonlinear loads. SAPF draws current in such a way that the source current which is sum of load current and active filter current becomes sinusoidal i.e.
Predicting energy consumption in buildings plays an important part in the process of digital transformation of the built environment, and for understanding the potential for energy savings. This also contributes to reducing the impact of climate change, where buildings need to increase their adaptability and resilience while reduc- ing energy consumption and maintain user comfort. The use of Internet of Things devices for monitoring and control of energy consumption in buildings can take into account user preferences, event monitoring and building optimization. Detect- ing peak energy demand from historical building data can enable users to manage their energy use more efficiently, while also enabling real-time response strategies (including control and actuation) to known or future scenarios. Several statistical, time series, and machine learning techniques are proposed in this work to predict electricity consumption for five different building types, by using peak demand fore- casting to achieve energy efficiency. We have used several indigenous and exogenous variables with a view to test different energy forecasting scenarios. The suggested techniques are evaluated for creating predictive models, including linear Regression, dynamic regression, ARIMA time series, exponential smoothing time series, artifi- cial neural network, and deep neural network. We conduct the analysis on an energy consumption dataset of five buildings from 2014 until 2019. Our results show that for a day ahead prediction, the ARIMA model outperforms the other approaches with an accuracy of 98.91% when executed over a 168 h (1 week) of uninterrupted data for five government buildings.
FIG 4 describes hybrid feature extraction. The features selected by hybrid feature selector can be considered that have no irrelevant features, but also have redundant features. In power price forecasting data requires non-linear mapping to find an appropriate low dimensional value. Kernel Principle Component Analysis (KPCA) uses kernel function to deal with high dimensional and low dimensional data.
Therefore, to address these challenges, we focus on the real time utility model for revenue optimization since this model can truly reflect the integrated information towards revenue, rather than only optimizes users’ adoption rate as previous works did. Specifically, we first define the problem of utility-recommendation towards revenue optimization. Then we set out a new research challenge: real-time utility-based recommendation. The major difference between our recommendation problem and the problems tackled in prior studies is that we take into account the retailer’s overall revenue, and we do this on streaming data. Also, since in dynamic e-commerce environments, different potential recommendation itemsets could earn different revenues for the retailer, and even the same itemsets could generate different levels of revenue at different times of year (e.g., during the holiday season), it is very difficult for traditional RS to tackle these problems. Therefore, we propose an Adaptive Online Top-K high utility itemsets mining model (RAOTK) that extracts the utility information just using real-time transaction flows, and monitors Top-K high utility itemsets to guide the recommendations towards optimal revenue. Furthermore, we observe that only considering utility may reduce the customers’ product adoption likelihood in some circumstances (since high-utility purchases may be less common than everyday purchases). As such, we derive three algorithm variants, which we call Adaptive Online Top-K HUIs mining with Ratings (RAOTK-R), Adaptive Online Top-K HUIs mining with Frequency (RAOTK-F) and Hybrid Adaptive Online Top-K HUIs mining with ratings and frequency (RHAOTK). Together, these give us a comprehensive set of utility-based recommendation algorithms. Finally, to improve accuracy and make our model more personalized to individual customers, we take buying power of customers into consideration and propose a simple but effective method to estimate the Online consumers’ Willingness to Pay (OWP). By adopting OWP, the performance of the proposed framework is substantially enhanced.
In evolutionary algorithms we generalise the area as evolutionary computation. The evo- lution is considered as a foundation term. Evolutionary algorithms are all based on pop- ulation of candidate solutions, where the candidate solutions are randomly generated in a finite search space. All the candidate solutions are evolved to find their possible opti- mal solutions. There are several strategies and methods named as evolutionary process, evolutionary strategies, evolutionary programming etc. These are all different methods, having different processes. However, all of them are originated under the general term of evolutionary computation. The initialization of the population is the first step of each evolutionary computation algorithm. Numerous structures have been adopted to generate and initialize population randomly in the search space. The next step is formulating an ob- jective or fitness evaluation function. This function is used for the evaluation of the entire population throughout the search space. Decisions are made on the basis of the derived function evaluation values. Some candidate solutions are selected as parents to produce new candidate solutions, called offspring or child. The process of evolution is described as two different techniques, which are crossover and mutation. In crossover technique the bits of parent candidates are twisted, exchanged, and recombined with each other to produce an offspring. Mutation changes the bits of one parent and thus produces a new offspring. The new population is generated through crossover and mutation. This new population of offspring is then evaluated on the given evaluation function or fitness function until it reaches its global optimum or maximum number of iterations.
The need to maximize P X C ( i ) , for i = 1, 2, 3, 4, P C ( ) i , the prior probability of each class can be com- puted based on the training data set. The Bayesian classifier more reliably predicts the new student belonging to class “first”. In the same manner any new student can be fitted to their respective class based on the performance. This research work focuses on the development of fuzzy logic and fuzzy C-means based fuzzy expert system to determine academic performance. Also presented is a new method for a new student allocation based on Baye- sian approach. A difference in outcomes is seen between the classical and proposed fuzzy logic based expert systems methods when results are evaluated from fuzzy expert system. While the classical method adheres to a constant mathematical rule, evaluation with fuzzy logic has great flexibility and reliability. The proposed fuzzy C-Means based Fuzzy Expert System automatically converted the crisp data into fuzzy set and also calculates the total marks of a student who appeared in semsester-1, semester-2 and semester-3 examination. A simple and qualitative methodology to compare the predictive power of clustering algorithm and the Euclidean distance is evident as result of this work. The Fuzzy C-Means clustering models have improved on some limitation of the existing traditional methods, such as average method and statistical method. The Fuzzy C-Means algorithm is best model for modeling academic performance in an educational domain. However, due to multiple iterations and various Eigen vectors the FCM method suffers from heavy computational burdens and is time-consuming. Apart from this, it is also highly sensitive to the initialization which usually requires a priori knowledge of the cluster numbers to form the initial cluster centers. Such limitations can be mitigated by the Subtractive cluster- ing based Tskagi-Sugeno (T-S) fuzzy model and combined Subtractive clustering with FCM called hybrid SC-FCM method.
Abstract: Rapid growth of information and communication technology benefits both the consumers and utilities for efficient utilization of the available power. A smarter grid has the potential to make our electricity system more efficient and cleaner. Imbalance in supply and demand can also be effectively taken care with the inclusion of demand response programs such as load based and price based without addition of generation which in turn is a long-term process. If the utilities fail to predict or balance the system during the power deficit it may lead to black outs and disturbs the operation of powergrid. At present, in order to obtain the system stability in a short span of periods and to address the variable demand conditions of the consumers the utilities are mainly concentrating on the meta-heuristic algorithms which provides the solutions very quickly and brings the grid into stable condition. The performance of three well recognized population based meta-heuristic algorithms such as GA, ACO and PSO, to solve load management at the consumer level in the smartgridenvironment were examined in terms of their efficiency, effectiveness and consistency in obtaining the optimal solution.
In the world we are living now have shortened of pure water and it has become responsible for our generation to hand over the water safely to the future generations. It means we have to utilize the water properly and for that IoT can be very useful to give people some knowledge about the conditions of water so a proper use of that water can be happen. With the help of sensors and using its data people can make sure that they make the best use of available water. With the help of sensors we can know the color and smell of sensor from very distance space if we upload the data of sensors in the cloud and manage to view that data in our mobile Phones. Even we can save much amount of water in dams if we know the amount of water present in it by using ultrasonic sensors. If we know the height we can easily know when to release water from the dam without any fear and can produce less damage when the water is released. It is not possible for people to monitor the water flow at all places. But by implementing IoT applications it is possible to monitor the pathways and tunnels, the water flows. And using IoT we can even guide the route of water to reach its destiny by opening gate valves with the help of sensors which is operated bya person at a distance. We can also know turbidity, pH and temperature using sensors.
Abstract— The authors recently proposed several model predictive control (MPC) approaches to managing residential level energy generation and storage, including centralized, distributed, and decentralized schemes. As expected, the dis- tributed and decentralized schemes result in a loss of perfor- mance but are scalable and more flexible with regards to net- work topology. In this paper we present a distributed optimiza- tion approach which asymptotically recovers the performance of the centralized optimization problem performed in MPC at each time step. Simulations using data from an Australian electricity distribution company, Ausgrid, are provided showing the benefit of a variable step size in the algorithm and the impact of an increasing number of participating residential energy systems. Furthermore, when used in a receding horizon scheme, simulations indicate that terminating the iterative distributed optimization algorithm before convergence does not result in a significant loss of performance.
The aim of the experiment is to measure the performance exhibited by Snort under high traffic. Snort version 188.8.131.52 has been installed as a host based NIDS with the default configurations. We set the output methods for both alerting and logging with writing output to files located in the default log directory. Two experiments have been performed to evaluate the performance of Snort. In first experiment, live malicious traffic is send to the server with Snort and its performance is evaluated for all the five algorithms. In second experiment, the performance is calculated and compared for the CDX-2009 dataset, which
The modern society has become much more dependent on the continuous availability of electric power, and that has made electric power to become one of the human’s fundamental needs of the modern age. However, for one to get access to it, a number of complex stages are involved, from the generation of power to the transmission, using long distance and short distance transmission lines up to where it is distributed to the households and industries. Through each stage there are technical issues experienced by the powergrid. Under-voltage (voltage sags), over-voltage (voltage swells), voltage surges and voltage spikes are amongst the problems experienced in the powergrid, they cause mal-operation of the electric equipment, increase in power loss and over burdening of the power system (Dhomane, et al., 2016). A modernized powergrid (smartgrid) is set to overcome all electrical issues involved in every stage where the electric power is passed, and that will create a power that is sustainable, economical, reliable and efficient to the society as a whole. The term Smartgrid has been defined by many organizations, one organization defined smartgrid as an automated electric powergrid, that performs based on the analogue or digital information it collects using Information and Communication Technology (ICT), such as information about the actions of the suppliers and consumers, in order to improve the generation, transmission and distribution of electricity that is efficient, reliable, economical and sustainable to the society(Subhalakshmipriya & Suganya, 2015). Various researchers from different perspectives are attracted by smart grid’s vision of transforming the traditional powergrid into an integrated state of the art future generation powergrid(KarthiKeyan, et al., 2016). The aim of the smartgrid concept is to provide electric power quality that is environmental free from greenhouse gas, electrical system that is economically evolved and technologically integrated, intelligently integrated communication and control system to the powergrid, to sustain energy for the future generation. However there are still many various goals the smartgrid concept wishes to archive. The top priority of worldwide energy utility companies at the moment is to increase energy efficiency while maintaining a clean environment from greenhouse gases by adopting renewable energy sources and an accelerated development of smartgrid technology(Ceaki, et al., 2017).
Critical infrastructures have gained increased concern as targets from future cyber attacks. The electric powergrid is extremely dependent on cyber infrastructures to per- form automated monitoring and control functions. Smartgrid initiatives, such as AMI, WAMS, and substation automation will significantly expand this dependency. The expo- sure of these ICT assets and their increasing adoption in smartgrid initiatives introduces numerous concerns questioning the adequacy of the grids current security posture. AMI presents both utilities and consumers with improved control over their electricity con- sumption. However, this large infrastructure presents numerous vulnerabilities which could be exploited by attackers to cut off power, falsify billing data, or access sensitive consumer privacy data. WAMS will incorporate PMUs to improve the accuracy of state estimation of the bulk power system. Although if critical PMU data is manipulated by an attacker, utilities may compute incorrect estimations which may result in improper control actions. Additionally, increased substation automation attempts to improve the reliability of critical control and protection applications. However, if an attacker is able to gain access to substation systems, they could trip breakers to cause significant damage to the grid.
ABSTRACT: With 215 GW of installed capacity, the potential demand of electricity in India is estimated to be as high as 900GW. It is expected that demand for electric energy will triple by 2050 all around the globe. Rolling power outages in developing countries, previously just an unwelcome fact of life, have escalated to the level of national emergency. This paper explores about the idea behind SmartGrid and the immense opportunities that lies within the SmartGrid. This paper gives an overview how SmartGrid technology can offer a country the tools needed to engage and overcome the challenges faced by current Grids.