Rahul Saxena

A review of load flow and network reconfiguration techniques with their enhancement for radial distribution network

Load flow analysis of electrical distribution networks either for providing household electricity or in integrated circuits has always been a topic of great interest for researchers from last few decades. Various novel methods and techniques have been proposed for load flow calculations in the network and simulation tools are being used to determine the various characteristics of the network. The paper mainly focusses on these techniques at a single site and the enhancements made to the already existing techniques throughout the use of GPU architecture using CUDA platform in lieu of enhancing the performance of already existing solution with respect to time and algorithmic complexity. The paper also explains the need of enhancing prevailing solutions for load flow analysis for future generation smart grids or real time systems by comparing the performance with serial version for different topologies of network. Further the paper also throws light on some of the network reconfiguration techniques used to remodel the RDN due to high power losses in the network and floats an idea how parallel processing can be beneficial in enhancing already existing genetic algorithm based network reconfiguration technique to support real time load flow calculations and topology construction in smart grids and future generation systems.

Faster Load Flow Analysis

Over the past few decades, load flow algorithms for radial distribution networks have been an area of interest for researches, which has led to improvement in the approach and results for the problem. Different procedures and algorithms have been followed in lieu of performance enhancement in terms of simplicity of implementation, execution time, and memory space requirements. The implementation of load flow algorithm using CUDA parallel programming architecture for a radial distribution network is discussed. The computations involved in serial algorithm for load current, branch impedances, etc., have been parallelized using CUDA programming model. The end result will be an improvement in execution time of the algorithm as compared to the running time of the algorithm over CPU. Finally, a comparison has been drawn between the serial and parallel approaches, where an improvement in execution time has been shown over the functions involved in computations.

GPU-Based Parallelization of Topological Sorting

Topological sort referred to as topo sort or topological ordering is defined as constraint-based ordering of nodes (vertices) of graph G or DAG (Directed Acyclic Graph). In other words, it gives a linearized order of graph nodes describing the relationship between the graph vertices. Many applications of various fields in computer science require a constraint-based ordering of tasks and, thus, topological sorting holds a big place of importance for many applications like semantic analysis in compiler design, Gantt chart generation in software project management and many more. In this paper, a parallel version of this ordering algorithm over CUDA (Compute Unified Device Architecture) has been discussed by identifying an approach to process-independent portions of the graph simultaneously for load flow analysis over radial distribution networks. The serial implementation of topological sort has been first discussed followed by its implementation on thread-block architecture of CUDA modifying the serial algorithm. Finally, the efficiency of this parallel version of topo sort has been investigated on various structures of graph modeled from radial distribution networks and has been reported.

An enhanced parallel version of RSA public key crypto based algorithm using openMP

Due to increased data movement and information exchange over internet and web, preserving data confidentiality and security has emerged as a prime concern for the end users. From bank transactions to document verification portals, from government official websites to social media, all these platforms share data remotely over web which contains certain confidential information and thus uses security mechanisms based on cryptographic algorithm to preserve the confidentiality of data preventing any information leakage or breach. RSA being the most popular public key cryptography based algorithm uses data encryption and decryption technique. It uses the mathematical concept of factorization of prime integers. The efficiency of the algorithm in terms of enhancing the security aspect to guess the factors of prime numbers increase when the numbers have high powers. Computing prime factors of these large numbers is compute intensive task where the serial programming makes RSA algorithm to slow down. The paper here presents an OpenMP based algorithmic modification to the code of RSA algorithm to improve the execution time using the parallel processing power of modern day multi-core architecture based machines. Experimental results under the lights of graphical representation shows a considerable speed up gained over the traditional implementation. Further the paper also compares the efficiency with other parallel implementations of RSA algorithm where the reported algorithm is found to perform almost equivalent to the GPGPU based implementation without the need of extra hardware in the form of graphics card and with reduced power consumption of the machine. The paper sums up with a conclusion that GPU based acceleration can be clubbed with proposed method to generate factors severely large prime numbers making it highly difficult to guess the factors.

Parallelizing GA based heuristic approach for TSP over CUDA and OPENMP

Travelling Salesperson Problem being a classic combinatorial optimization problem is an interesting but a challenging problem to be solved. It falls under the class of NP-hard problem and becomes non-solvable for large data set by traditional methods like integer linear programming and branch and bound method, being the earlier popular approaches. Genetic Algorithm based solutions emerged as the most popular tool to solve this which is a heuristic mechanism to find the closest approximate solution to the problem. In this paper, we approach Travelling Salesman Problem using GA approach, a natural selection technique which repeatedly modifies a population of individual solutions, with the added power of modern computing systems. Here we come up with a parallel version of GA for both multicore and many core architectures over OpenMP and CUDA in order to make some of the challenging problems like Vehicle Routing Problem for google maps, DNA sequencing and many more optimization problems involving a good amount of complex computation and data handling. The paper presents a comparative analysis of the results for both CUDA and OpenMP for various degree and structure of graphs depicting close approximation to the accuracy in terms of most optimal path for traversal with highly reduced execution time.

Parallelization of Load Flow Analysis

This paper has been proposed to present a simple approach for load flow analysis of a radial distribution network using parallel programming in Computationally Unified Device Architecture (CUDA). The proposed approach applies Breadth First Search to evaluate the nodes in the network and Kirchhoff’s current law (KCL) as well as Kirchhoff’s Voltage Law (KVL) for evaluating the current and voltages at each of the network nodes. The procedure is repeated till the convergence criterion is achieved. The paper demonstrates the working of Breadth First Search using CUDA. The efficiency of load flow algorithm has been enhanced by utilizing parallel computational power of Graphics Processing Unit (GPU). This approach has been tested for 33-nodes as well as for 69-nodes radial distribution systems and comparison has been done between the performances of sequential approach over CPU and parallel approach on GPU. The results show that introducing CUDA to load flow analysis speeds up the performance of the system by faster executions and gives accurate desired results as compared to sequential approach.

VANET: Security Attacks, Solution and Simulation

Vehicular ad hoc network is introduced through MANET in the year of 2000. They are intelligent system which is implemented in the vehicles to provide the luxury, comfort, and security to the peoples. At present, the use of vehicles is growing per year as per user demand. Various researches have been done in VANET through different perspectives, and security issues in VANET are studied and resolved. This paper analyzes the various security issues and their proposed solutions, and the comparative study of this solution is done. Also, the performance of these security solutions is analyzed.

An OpenMP-Based Algorithmic Optimization for Congestion Control of Network Traffic

The last decade being a web revolution in the field of electronic media, data and information exchange in various forms has significantly increased. With the advancement in the technological aspects of the communication mechanism, the textual form of data has taken the shape of audiovisual format, and more and more content over internet is being shared in this form. Data sharing in this form calls for the need of high bandwidth consumption which may slow down the network resulting in performance degradation of content delivery networks due to congestion. Several attempts have been made by the researchers to propose various techniques and algorithms to achieve optimal performance of the network resources under high-usage circumstances. But due to high-dense network architectures, the performance implementations of suggested algorithms for congestion may not be able to produce the desired results in real time. In this paper, we have presented an optimized multi-core architecture-based parallel version of two congestion control algorithms—leaky bucket and choke packet. The experimental results over a dense network show that optimized parallel implementation using OpenMP programming specification gets the network rebalancing in a very short span of time as compared to its serial counterpart. The proposed approach runs 60% faster than the serial implementation. The graphical map for the speed up continues to increase with the size of the network and routers. The paper throws the light on the implementation aspects as well as result analysis in detail along with some existing algorithms for the problem.

Sudoku Game Solving Approach Through Parallel Processing

Sudoku is one of the most popular puzzle games of all time. Today, Sudoku is based on simple rules of placing the numbers from 1 to 9 in the empty cells of the Sudoku board. After solving the Sudoku, each row, column, and mini grid should have only one occurrence of each digit between 1 and 9. There are various approaches introduced to solve the Sudoku game till now. The solutions in the state of art are computationally complex as the algorithmic complexity for the solution falls under the class of NP-complete. The paper discusses various solutions for the problem along with their pros and cons. The paper also discusses the performance of a serial version solution with respect to execution time taken to solve Sudoku. With certain modifications and assumptions to the serial version code, we have evaluated the possible solution and bottleneck in the parallel approach. The results have been evaluated for the parallelized algorithm on a normal user machine first, and then on a supercomputing solution box PARAMSHAVAK. Finally, the paper is concluded with the analysis of all the three solution aspects proposed.

Parallel Computing of Load Flow Analysis

This paper presents a simple and an efficient approach for load flow analysis of a radial distribution network by using parallel computation. By introducing CUDA (Computationally Unified Device Architecture), a platform is provided for performing the parallel implementation of the analysis.The analysis starts by determining the order of the nodes present in the distribution network. To generate the order of the nodes, Breadth First Search is used in parallel. Further, the current and the voltages present at each node of the network are evaluated. This procedure is repeated till the convergence criterion is achieved. CUDA provides an efficient utilization of the computational power of GPU, by which the performance and throughput of sequential load flow algorithm, has been enhanced dramatically. With the immense support of many GPUs, CUDA helps load flow analysis to efficiently manage and manipulate a large blocks of data of the distribution networks. Moreover there has been a lot of reduction in the execution steps as compared to the serial computing.