Load Imbalance Research Articles

Control for Power Quality Improvement of Solar Photovoltaic-Distributed Static Synchronous Compensator Interfaced with Weak Grid Using Multi-Variable Filter Dual Second-Order Generalized Integrator Phase-Locked Loop

This study proposes a control method for transferring power generated by a solar photovoltaic system. The control algorithm, called MVF-DSOGI-PLL or Multi-Variable Filter Dual Second-Order Generalized Integrator Phase-Locked Loop), accurately estimates symmetrical components, phase angle, and positive sequence voltages of the Common Point of Interface (CPI) in distorted grid conditions. The approach creates a compensatory current to reduce PQ difficulties by extracting the size of the basic current load undergoing operation employing the SOGI method. These systems can operate in either modes of PV-DSTATCOM or DSTATCOM, depending on solar insolation’s availability. The simulation models of the grid-interfaced SPV systems are developed in Matlab/Simulink to evaluate the performances of the suggested control method. The results of simulation demonstrate that the suggested control systems efficiently transfers the power generated by PV arrays to a grid while performing tasks like load balancing, power factor correction (PFC), and harmonics reduction. Furthermore, the MVF-DSOGI-PLL-based control algorithm successfully mitigates PQ issues in distribution networks with weak grids, ensuring that PQ levels remain between a limited set by the IEEE-519 standard. A laboratory prototype of the proposed system is implemented to assess its steady states and dynamic efficient under fluctuating solar insolation, and load imbalance distortions.

DLB: A Dynamic Load Balance Strategy for Distributed Training of Deep Neural Networks

Synchronous strategies with data parallelism are widely utilized in distributed training of Deep Neural Networks (DNNs), largely owing to their easy implementation yet promising performance. In these strategies, the workers with different computational capabilities need to wait for each other because of the essential gradient or weight synchronization. This will inevitably cause the high-performance workers to waste time waiting for the weak computational workers, which in turn results in the inefficiency of the cluster. In this paper, we propose a Dynamic Load Balance (DLB) strategy for the distributed training of DNNs to tackle this issue. Specifically, the performance of each worker is evaluated first based on the performance demonstration during the previous training epochs, and then the batch size and dataset partition are adaptively adjusted in consideration of the current performance of the workers. As a result, the waiting cost among the workers will be eliminated, thereby the utilization of the clusters is highly improved. Furthermore, the essential theoretical analysis has also been provided to justify the convergence of the proposed algorithm. Extensive experiments have been conducted on the CIFAR10 and CIFAR100 benchmark datasets with four state-of-the-art DNN models. The experimental results indicate that the proposed algorithm can significantly improve the utilization of the distributed cluster. In addition, the proposed algorithm can also prevent the load imbalance of the distributed DNN training from being affected by the disturbance and can be employed flexibly in conjunction with the other synchronous distributed DNN training methods.

DetonationFoam: An open-source solver for simulation of gaseous detonation based on OpenFOAM

Detonation has promising applications in advanced propulsion systems, and numerical simulation is widely used to gain insights into the complex interaction between the hydrodynamic flow and chemical reactions involved in detonation. In this work, detonationFoam, an open-source solver for accurate and efficient simulation of compressible reactive flow and detonation are developed based on OpenFOAM. detonationFoam can simulate compressible, multi-component, reactive flow and it can accurately evaluate the detailed transport coefficients using the mixture-averaged transport model. Compared to rhoCentralFoam, the improved HLLC-P approximate Riemann solver is used in detonationFoam and it helps to accurately resolve shock waves appearing in detonation. Besides, the adaptive mesh refinement and dynamic load balancing algorithms are used in detonationFoam, which greatly improves the computational efficiency. Validation tests including homogenous ignition, unsteady diffusion, shock tube problem, premixed flame, planar detonation, double Mach reflection, detonation cellular structure and oblique detonation wave are conducted. These tests demonstrate that detonationFoam can be used to accurately and efficiently to simulate the compressible, multi-component reactive flow and detonation processes. Program summaryProgram Title: detonationFoamCPC Library link to program files:https://doi.org/10.17632/x45zh4nz28.1Developer's repository link:https://github.com/JieSun-pku/detonationFoamLicensing provisions: GPLv3Programming language: C++Nature of problem: Gaseous detonation involves different length scales and complicated chemistry. To accurately and efficiently simulate the gaseous detonation, adaptive mesh refinement needs to be conducted and detailed chemistry should be considered. Besides, the severe load imbalance caused by the chemical source term evaluation may greatly reduce the computation efficiency.Solution method: An open-source solver, detonationFoam is developed based on OpenFOAM. The species equations considering detailed chemistry are solved in detonationFoam and thereby detonation in a compressible, multi-component, reactive flow can be simulated. The adaptive mesh refinement technique and the dynamic load balancing algorithm are incorporated into detonationFoam. It is demonstrated that detonationFoam can accurately and efficiently simulate gaseous detonation.

A simple closed-loop interphase power controller for cascaded H-bridge converter-based DG units in islanded microgrids with imbalanced loads

A simple closed-loop interphase power controller for cascaded H-bridge converter-based DG units in islanded microgrids with imbalanced loads

A dynamic load balancing algorithm for CFD–DEM simulation with CPU–GPU heterogeneous computing

Computational fluid dynamics combined with discrete element method (CFD–DEM) is widely adopted to study particle–fluid multiphase systems. However, due to the global dynamical evolution and the naturally non-uniform distribution of particles, the computational load in parallel computing varies dynamically in each process, leading to significant load imbalance. Thus, a dynamic load balancing (DLB) algorithm is proposed for parallel CFD–DEM simulations in the CPU–GPU heterogeneous architecture. A type of computing load analyzing grid (AG) is developed which covers the entire simulation region. Then Kalman filtering and linear weighted method is used to solve the weight of fluid grids and particles. The domain decomposition scheme is updated according to the weights of AGs dynamically. This DLB method is tested for gas–solid fluidized beds and periodic gas–solid systems of different scales. Results show that the DLB algorithm can significantly speed up CFD–DEM simulation and the highest performance is 2.28 times faster with DLB.

Analisis Ketidakseimbangan Beban Pada Tranformator Distribusi Di Gardu Induk Budurn

Electricity is a basic need for humans today, it has been proven that humans are very dependent on electricity for their daily activities. If electricity is not available, then human daily activities will be disrupted. The occurrence of load imbalance that occurs in the R, S, and T phases due to the absence of concurrent when the load is turned on, resulting in a neutral current flow on the secondary side of the transformer. with the development of electric power distribution systems. The benefit of reducing losses on the neutral conductor of the distribution transformer is that it can produce savings in the supply of electricity, and other benefits obtained include the resulting system. The method used is that the data resulting from field measurements are converted into mathematical equations. From the research conducted, the results showed the average value of the loading current and load imbalance in the distribution transformer at the Buduran Substation, namely the average loading current of (14.4 A), Load Imbalance of (28.3 A), Neutral Current Losses of (2.34 W). And Copper Losses of (0.510945 kW).

Pengaruh Pembebanan Terhadap Efisiensi Transformator Distribusi Di PT. PLN (Persero) UP3 Garut

Transformer efficiency is defined as the ratio of the incoming power to the outgoing electric power. Even so the ideal transformer has 100% efficiency, this ideal transformer does not exist, because the energy that goes out is always less than the energy that goes in, meaning that there is energy loss. Load imbalance in an electric power distribution system always occurs and the imbalance is in single-phase loads in low-voltage network subscribers. Due to the load imbalance, a current appears in the neutral of the transformer. Perform analysis and compare efficiency. Analysis of used load capacity is not efficient. Analysis of the results of current and voltage measurements at the substation to obtain efficiency, and transformer losses in day and night conditions using LWBP and WBP methods. In this study, observations were made regarding the transformer efficiency analysis using the Outside Peak Load Time (LWBP) and Peak Load Time (WBP) methods. The LWBP method is carried out during the day for 20 hours, while the WBP method is carried out at night for 4 hours, from 18.00 to 22.00. Then measurements were made at PT. PLN (Persero) UP3 Garut. For the highest percentage of loading which is close to the standard of 80%, namely the afternoon loading at the PDDK substation, which is 50.93%, the load used is 50.93 kVA. The night load at the CITL substation is 77.49%, the load used is 77.49 kVA. For the lowest percentage of loading that deviates from the standard 80%, namely the daytime loading at the SJRK substation, which is 16.32%, the load used is 16.32 kVA. The night load at the SJRK substation is 21.52%, the load used is 21.52 kVA. Based on observations, it is not efficient because the unused load is very large, it is recommended to use a smaller transformer capacity. The highest efficiency in the LWBP method is at the CITK substation 97.87%, the lowest at the PDDK substation 94.55%. The highest efficiency in the WBP method is at SJRK substations 95.5%, the lowest at PDDK substations is 92.1%.

Explore deep reinforcement learning for efficient task processing based on federated optimization in big data

In recent years, along with the extensive application of consumer electronics, the task execution with cloud computing for big data has become one of the research focuses. Nevertheless, the traditional theories and algorithms are still employed by existing research work to explore the feasible solutions, which takes a beating from low generalization performance, system load imbalance, more response delay, etc. To solve the matter, a task execution method called DROP (Deep Reinforcement network aided Optimization method aiming at task Processing) has been put forward, which is capable of completing task request allocation through virtual network embedding. The prominence of this method is explained by its effect in reducing load balancing degree, minimizing bandwidth resource overhead, and preserving electric energy as well as meeting customer demands. It makes use of Deep Deterministic Policy Gradient (DDPG) instead of depending on tons of iterations for better path selection schemes in previous methods, through continuous environment interaction and trial-and-error evaluation to get better strategy selection for virtual link embedding. To realize the virtual node embedding in the federated optimization based system architecture, the intentional deep feature learning network is applied. Compared with the cutting edge approaches, the performance benefits of DROP can be verified by the experimental results in terms of bringing down the extra cost on resources and energy of the substrate network during the task execution for big data.

БАЛАНСИРОВКА ФАЗНЫХ НАГРУЗОК В НИЗКОВОЛЬТНЫХ ТРЕХФАЗНЫХ ЧЕТЫРЕХПРОВОДНЫХ СЕТЯХ

The imbalance of phase loads is a natural negative property of a low-voltage three-phase four-wire network due to uneven connection to the phases and the difference in individual consumer demand for electricity. The imbalance of phase loads in the network nodes leads to a decrease in the quality of electricity, overestimated technical losses in the network and to current overloads in one of the phases of power lines. In the paper, balancing of phase loads in the network is performed using a heuristic algorithm based on average hourly readings of loads and voltages readings of consumer smart meters. The effectiveness of the algorithm was evaluated by reducing losses, maximum currents in the phase wires and the neutral wire in a real feeder of the electrical network

Optimization Method of Energy Storage Configuration for Distribution Network with High Proportion of Photovoltaic Based on Source–Load Imbalance

After a high proportion of photovoltaic is connected to the distribution network, it will bring some problems, such as an unbalanced source and load and voltage exceeding the limit. In order to solve them, this paper proposes an optimization method of energy storage configuration for a high-proportion photovoltaic distribution network considering source–load imbalance clustering. Taking the minimum total voltage deviation, the minimum total power loss and the minimum total operating cost as the objective function, and considering various constraints such as power balance constraints and energy storage operation constraints, a mathematical model for energy storage configuration optimization is established. Firstly, the source–load imbalance of the distribution network with a high proportion of photovoltaic is defined. Therefore, according to the 24 h photovoltaic and load data, the 24 h source–load imbalance can be obtained, and the optimal k value can be determined by the elbow rule, so that 24 h a day can be clustered into k periods by the k-means algorithm. Then, the fuzzy comprehensive evaluation method is used to determine the weight factors of each objective function in each period, and three scenes are determined according to the different amount of energy storage. Then, the hybrid particle swarm optimization algorithm proposed in this paper is used to solve the model, and the minimum objective function value, optimal position and optimal capacity of each energy storage grid in each scene are obtained. Finally, it is applied to an example of IEEE33. In the results, the total voltage deviation is increased by more than 10%, the total power loss is increased by more than 8% and the total operating cost is increased by more than 12%, which verifies the effectiveness of the proposed model.

Profiling and optimization of Python-based social sciences applications on HPC systems by means of task and data parallelism

The article presents optimization techniques for two Python-based large-scale social sciences applications: SN (Social Network) Simulator and KPM (Kernel Polynomial Method). These applications use MPI technology to transfer data between computing processes, which in the regular implementation leads to load imbalance and performance degradation. To avoid this effect, we propose a 2-stage optimization. In the first step, the order of tasks is changed, and in the second step, the tasks are divided into smaller ones for easier allocation. In addition, we focus on mitigating performance and memory bottlenecks using modern ccNUMA systems with multiple NUMA domains. As part of the performance analysis, the limitations of communication in data traffic between and within the processor were revealed and resolved through appropriate data allocation. Benchmarking was carried out, examining various environments, including vendors of traditional x86-64 and ARM-based processors for HPC.

The Strategy of Strengthening Efficiency and Environmental Performance of Product Changeover in the Multiproduct Production System

Product changeovers are especially critical for multiproduct environments where key production system requirements are flexibility, time, and quality. Massive waste of production time and environmental pollution increase with changeovers significantly. It is noted that the green supply chain is gradually emerging, and the environmental policies in countries are also increasing pressure on manufacturers globally. However, how to improve the changeovers’ environmental performance in manufacturing enterprise are not entirely focused. The present study aims to raise the changeovers’ time efficiency and reduce environmental pollution in the multiproduct production system. This paper first analyzes the characteristics of multiproduct production systems and the causes of inefficient work and pollution and then extracts the problems that need to be optimized. They are the frequent changeover work, complex operation programs, and load imbalance. Multiproduct Production Fast Changeover (MPFC) is developed based on these problems, which integrates the Analytic Hierarchy Process, Entropy Weight method, Divisive Analysis, and Firefly algorithm. In addition, Divisive Analysis’s distance calculation is improved for flexible clustering targets. Firefly algorithm’s exploration, exploitation, and population coordination mechanisms have also been enhanced. The effectiveness of MPFC is proved in a real multiple flow-lines production case: time efficiency was increased, while the multiple industrial pollutions and key resource consumption were also reduced.

Modified Adaptive Filter Based UPQC for Battery Supported Hydro Driven PMSG System

The standalone three phase system for power quality improvement of both the load side voltages as well as the currents of the machine terminal is presented in this paper. This standalone system consists of a three-phase permanent magnet synchronous generator (PMSG) driven by governor-less hydro turbine as a three-phase source feeding the sensitive/non-linear/linear loads, connected via a unified power quality conditioner (UPQC). The UPQC consists of a back-to-back connected voltage source converters (VSCs) sharing a common DC bus. The series compensator of the UPQC is used to eliminate voltage power quality (PQ) and to maintain the sinusoidal, balanced, and constant amplitude voltages at load terminals irrespective of any variation in the system. Moreover, the current quality such as reactive power compensation, harmonics elimination, etc. are provided by the shunt compensator of the UPQC, simultaneously while improving voltage quality issues. In addition, the efficiency of the overall system is significantly improved by running the generator below its rated voltage and reduces the losses by operating at unity pf. The system frequency and the generated voltage magnitudes are maintained during any load fluctuation or imbalance loading. The model of the UPQC based hydro turbine driven PMSG system is developed in the Simulink-MATLAB environment, and its results are presented for analysis. The validation of the system is also done on the hardware prototype, and it performs effectively for the voltage and current PQ enhancement simultaneously. It significantly improves the PQs at the generator terminals and ensures to adhere the IEC-61727 and the IEEE-519 standards even with either balanced or unbalanced loads.

A Novel MIMO Control for Interleaved Buck Converters in EV DC Fast Charging Applications

This brief proposes a new multiple input multiple output (MIMO) control for off-board electric vehicle (EV) dc fast chargers. The proposed feedback matrix design avoids multiple tuning of controllers in multiple and interconnected loops while improving the performance of interleaved dc buck converters over classical PI/PID controls. The innovative features of the presented strategy are the reference current monotonic tracking from any initial state of charge with an arbitrarily fast settling time and the fast compensation of both load variations and imbalances among the legs. Numerical results validate the performance improvements of the proposed discrete-time MIMO algorithm for interleaved buck converters over classical PI/PID controls. Full-scale hardware-in-the-loop (HIL) and scaled-down prototype experimental results prove the feasibility and effectiveness of the proposal.

Phase Identification in Smart Grids – Case Study

. A widespread problem of the electricity distribution system is determining to which phase a consumer is connected. Generally, utility companies did not record the distribution of the phases over time, which is essential to address the problem of unbalanced voltage distribution networks. In addition, manual phase identification is not feasible due to its high cost. The massive deployment of smart meters allowed periodic readings of energy consumption, voltage, current, etc. Out of this large amount of data, several techniques based on machine learning have emerged, addressing the problem of phase identification automatically while maintaining the existing infrastructure. Hence, phase identification is essential for developing smart-grid solutions. In the present work, we applied an unsupervised machine-learning technique that allows the classification of the time-voltage series recorded by the smart meters of a low-voltage three-phase radial distribution network located in the Balearic Islands (Spain). The results show that there is a correlation between the time series of the feeder voltage and the consumer meters. The proposed method reached a 100% success rate in the case study. In addition, the results obtained open the way to deploy a new grid configuration to minimize the load imbalance.

PENGARUH KETIDAKSEIMBANGAN BEBAN TERHADAP PROFIL JARINGAN TEGANGAN RENDAH

Load imbalance occurs due to unequal distribution of load in each phase, this can occur in low voltage network distribution lines, one of which occurs in the DS0588 transformer, Merdeka feeder, Denpasar. Based on the data obtained, the percentage of DS0588 transformer load unbalance is 14%. The DS0588 transformer is one of the transformers at the Merdeka Denpasar Feeder with a capacity of 160 kVA which is located on Jalan Badak Agung, Renon Denpasar. Load unbalance data obtained at PLN UP3 South Bali. The research was carried out by the method of redistribution of the load on each phase under unbalanced load conditions, so that the total load of the three phases is the same or close. The results obtained were that the percentage of unbalanced loads in unbalanced conditions was 14% and when balanced conditions decreased to 2.2%. The losses in the distribution network when the unbalanced condition is 25 kW and the balanced condition drops to 23.3 kW, this shows that the load imbalance has an effect on the low voltage network losses. The higher the load imbalance, the higher the losses incurred.

A hybrid fast inference approach with distributed neural networks for edge computing enabled UAV swarm

Nowadays, unmanned aerial vehicle (UAV) swarm supported by mobile edge computing is attracting more and more attention, such as smart agriculture, smart transportation, smart security surveillance, and smart environmental monitoring. Nevertheless, small UAV devices cannot deploy advanced best-performing DNN models with insufficient computing and storage resources. Aerial computing centers serving as edges can collect tasks from UAVs, reduce UAV load, and provide high-accuracy inference. Limited radio resources in offloading and load imbalance in computing centers could lead to significant latency in task inference. Considering image processing in UAV applications, we propose a hybrid inference framework combining early exit and task offloading based on distributed neural network. The bottleneck-designed DNN provides a smaller intermediate data size than the original input. By deploying shallow neural networks at the drone terminals and advanced DNNs at the aerial computing servers, latency-constrained tasks can be exited after local inference, offloaded to the edge after local computation, or directly offloaded to the edge. Compared to binary or partial offloading, the early-exit-based framework provides fast inference as the workflow stops when a confident result is obtained. Combining task offloading into the inference process offers a more flexible option to release the UAV computing load. The proposed approach efficiently utilizes wireless transmission and balances the computation load to maximize the system inference accuracy within the latency constraints. Practical task inference simulations demonstrate that our approach obtains higher system accuracy under different radio resources, task generation rates, and terminal computing capacities conditions.

A Robust Control Strategy for the Automatic Load Commutation Device Considering Uncertainties of Source and Load

The long-term three-phase imbalance in distribution systems can lead to increased energy losses during transmission, reduced efficiency in energy utilization, and have serious implications for power supply security, power quality, and economic operations of the power system. The automatic load commutation device is an automatic device designed to address the issue of three-phase load imbalances in low-voltage distribution networks. By adjusting the phase of users without interrupting the power supply, this automatic device evenly distributes the load across all phases and effectively resolves the three-phase imbalance problem. Therefore, aiming at the above issue, a robust optimization method is adopted to address the control problem of automatic load commutation devices for a low-voltage distribution network. First, using historical data from photovoltaic and wind power generation systems as well as user load, a forecast analysis is conducted and uncertainty models for renewable energy and load demand are established. Then, a robust control strategy for the automatic load commutation device is proposed, which considers uncertainties of both source and load using the robust optimization method. The proposed model is then linearized using second-order cone technology and strong dual theory, and the column-and-constrained generation (C&CG) algorithm is employed to solve the problem iteratively. Finally, a modified IEEE 33-bus system is taken to verify the effectiveness of the proposed strategy. The simulation results show that the proposed load commutation device robust control model can enhance the ability of the distribution network to respond to load demand and renewable energy fluctuations while ensuring the economic operation of the distribution network. In addition, by adjusting the deviation amount and uncertainty parameters of load demand and renewable energy, a good balance between the robustness and economy of the proposed model can be achieved.

Performance Implication of Tensor Irregularity and Optimization for Distributed Tensor Decomposition

Tensors are used by a wide variety of applications to represent multi-dimensional data; tensor decompositions are a class of methods for latent data analytics, data compression, and so on. Many of these applications generate large tensors with irregular dimension sizes and nonzero distribution. CANDECOMP/PARAFAC decomposition ( Cpd ) is a popular low-rank tensor decomposition for discovering latent features. The increasing overhead on memory and execution time of Cpd for large tensors requires distributed memory implementations as the only feasible solution. The sparsity and irregularity of tensors hinder the improvement of performance and scalability of distributed memory implementations. While previous works have been proved successful in Cpd for tensors with relatively regular dimension sizes and nonzero distribution, they either deliver unsatisfactory performance and scalability for irregular tensors or require significant time overhead in preprocessing. In this work, we focus on medium-grained tensor distribution to address their limitation for irregular tensors. We first thoroughly investigate through theoretical and experimental analysis. We disclose that the main cause of poor Cpd performance and scalability is the imbalance of multiple types of computations and communications and their tradeoffs; and sparsity and irregularity make it challenging to achieve their balances and tradeoffs. Irregularity of a sparse tensor is categorized based on two aspects: very different dimension sizes and a non-uniform nonzero distribution. Typically, focusing on optimizing one type of load imbalance causes other ones more severe for irregular tensors. To address such challenges, we propose irregularity-aware distributed Cpd that leverages the sparsity and irregularity information to identify the best tradeoff between different imbalances with low time overhead. We materialize the idea with two optimization methods: the prediction-based grid configuration and matrix-oriented distribution policy, where the former forms the global balance among computations and communications, and the latter further adjusts the balances among computations. The experimental results show that our proposed irregularity-aware distributed Cpd is more scalable and outperforms the medium- and fine-grained distributed implementations by up to 4.4 × and 11.4 × on 1,536 processors, respectively. Our optimizations support different sparse tensor formats, such as compressed sparse fiber (CSF), coordinate (COO), and Hierarchical Coordinate (HiCOO), and gain good scalability for all of them.

FMap: A fuzzy map for scheduling elephant flows through jumping traveling salesman problem variant toward software‐defined networking‐based data center networks

SummaryNowadays data center networks (DCNs) should handle the ever‐growing load generated by diverse applications. It particularly occurs under concurrent flow requests and so‐called mice flows (MFs): elephant flows (EFs) ratio. Concentrating on a binary vision over flow size classification (EF or MF) results in subsequent unpredicted load imbalance due to neglecting EF's distinctions including a wide range of sizes. As a result, some EFs might utilize a path owing qualities beyond the given EF's demands, while another EF with higher requirements is attending to use an over‐utilized path. This article proposes FMap, a fuzzy map for scheduling EFs through our proposed variant of traveling salesperson problem (TSP) toward DCNs. FMap represents a novel EF scheduling scheme that integrates flow prioritization and routing decisions under the event pf parallel incoming flows besides the cooperation of the controller and OpenFlow switches in software‐defined networking (SDN) paradigm. FMap adopts fuzzy inference process to overcome the vagueness over EF's resource allocation. Mainly, FMap proposes a new variant of TSP (optimized by genetic algorithm) which enables EF's group forwarding with a minimum cost. FMap reduces the total hop count of EFs through considering a single optimal path for delivering groups of EFs that contain a same tag (priority). The outstanding results represent a major improvement as compared with equal cost multiple path, Hedera, Sonoum, and Size‐KP‐PSO. Particularly, the results illustrate an outperforming by 3.76×, 0.21×, 0.15×, 0.03×, and 0.03× in terms of total hop count, EFs FCT, packet loss, goodput, and received packets as compared with Size‐KP‐PSO, respectively.