Daily electric load forecasting using artificial neural network

A multilayered-type neural network is attractive for daily electric load forecasting because the neural network can acquire a nonlinear relationship among the electric load data and their factors (weather, temperature, etc.) automatically. This paper discusses first some essential issues to be considered in neural network applications. One is difficulty of obtaining sufficient effective training data, another is the influence of abnormal learning data, and one more is the inevitable outerpolation. For these issues, the following three methods are developed in order to forecast more accurately: (1) a structure of the neural networks for insufficient training data; (2) detection and diminishing the influence of abnormal data; (3) employment of interpolation network and outerpolation network with additional data for outerpolation. Furthermore, to increase the sensitivity between electric loads and factors, (4) removal of base load is developed. Those methods work effectively to decrease the average absolute errors of peak-load forecasting and 24-hour load forecasting to 1.78 percent and 2.73 percent, respectively.     

基于SoPC的人工神经网络的硬件实现方法

提出了一种基于SoPC的神经网络的硬件实现方法,该方法以FPGA器件为硬件载体,NIOSⅡ软核处理器为CPU,Avalon片内总线为数据交换架构。研究了多层前馈神经网络在FPGA上的实现方法,描述了神经网络模块与Avalon片内总线的接口技术。整个系统在Altera的EP2C8Q208C8器件上实现,结果表明,该方法的应用不仅提高了人工神经网络的运算速度,还提高了整个系统的灵活性。     

An artificial neural network approach to transformer faultdiagnosis

This paper presents an artificial neural network (ANN) approach to the diagnosis and detection of faults in oil-filled power transformers based on dissolved gas-in-oil analysis. A two-step ANN method is used to detect faults with or without cellulose involved. Good diagnosis accuracy is obtained with the proposed approach     

An adaptive neural network approach to one-week ahead loadforecasting

A neural network approach is proposed for one-week ahead load forecasting. This approach uses a linear adaptive neuron or adaptive linear combiner called Adaline. An energy spectrum is used to analyze the periodic components in a load sequence. The load sequence mainly consists of three components: base load component, and low and high frequency load components. Each load component has a unique frequency range. A load decomposition is made for the load sequence using digital filters with different passband frequencies. After load decomposition, each load component can be forecasted by an Adaline. Each Adaline has an input sequence, an output sequence, and a desired response-signal sequence. It also has a set of adjustable parameters called the weight vector. In load forecasting, the weight vector is designed to make the output sequence, the forecasted load, follow the actual load sequence; it also has a minimized least mean square error. This approach is useful in forecasting unit scheduling commitments. Mean absolute percentage errors of less than 3.4% are presented from five months of utility data, thus demonstrating the high degree of accuracy that can be obtained without dependence on weather forecasts     

战略资产管理及其在电力企业中的应用

介绍了 UMS公司用以评估国电华东公司资产管理水平的战略资产管理 (SAM)流程框架 ,分析了SAM各组成部分间的相互关系及其所体现的先进管理思想。对电力企业采纳 SAM改进资产管理工作提出一些建议 ,指出电力企业应在行业改革的情况下 ,借鉴 SAM管理理念 ,结合企业自身的实际 ,适时应用 SAM流程框架 ,提高企业的核心竞争能力     

Cellular neural network for fast adaptive equalization

The use of an analogue neural network in the adaptive equalization of time-varying communication channels is proposed. the network is used to compute the coefficients of a linear transversal filter. the settling time decreases as the filter order increases and as the signal-to-noise ratio decreases. Owing to the real-time processing capabilities, the network can be useful when it is of interest to track fast variations, as in radio links. the special properties of the tap input correlation matrix result in a cellular network architecture which greatly simplifies the VLSI implementation. Simulation results are presented which point out very satisfactory performance.     

智能用电网络的设计与初步实现

发展智能用电是建设智能电网的重要环节之一。本文综合以往智能用电、智能家居与自动需求响应的认识,提出了"智能用电网络",给出其定义。智能用电网络是将用户侧的各种电器通过能量信息网关互连而形成的网络。它旨在建立基于物联网、互联网与智能电网技术的联系海量电力用户的大型用电网,不仅可以帮助用户节能,给电力用户提供智能友好的个性化、差异化服务,还可以引导电力用户灵活互动地参与协同需求响应,帮助电网实现优化运行。智能用电网络的硬件部分由智能插座、智能红外控制器、能量信息网关、云端服务器、移动终端等部分组成;软件部分基于Android平台进行开发。目前,该网络已在清华大学紫荆公寓投入试运行,运行结果表明本文方法的可行性与有效性。     

基于优化神经网络的电网谐波检测方法

为了检测电力系统中的谐波,本文提出了一种基于优化神经网络的电网谐波测量方法.该方法应用一个结构和训练算法都优化了的多层前馈神经网络(MLFNN)对电网中的谐波进行检测,即首先考虑到神经网络的权值记忆负担主要来自谐波幅值和相角的变化,因此先对相角进行确定;再用基于神经网络理论方法对幅值进行检测,并使每一个输出神经元都对应于自己的隐层;然后利用多层前馈神经网络对当前及上一时刻的采样值进行分析,实现了对电网谐波的检测.实验仿真结果证明了该方法的有效性.     

A novel neural network controller and its efficient DSP implementation for vector-controlled induction motor drives

An artificial neural network controller is experimentally implemented on the Texas Instruments TMS320C30 digital signal processor (DSP). The controller emulates indirect field-oriented control for an induction motor, generating direct and quadrature current command signals in the stationary frame. In this way, the neural network performs the critical functions of slip estimation and matrix rotation internally. There are five input signals to the neural network controller, namely, a shaft speed signal, the synchronous frame present and delayed values of the quadrature axis stator current, as well as two neural network output signals fed back after a delay of one sample period. The proposed three-layer neural network controller contains only 17 neurons in an attempt to minimize computational requirements of the digital signal processor. This allows DSP resources to be used for other control purposes and system functions. For experimental investigation, a sampling period of 1 ms is employed. Operating at 33.3 MHz (16.7 MIPS), the digital signal processor is able to perform all neural network calculations in a total time of only 280 /spl mu/s or only 4700 machine instructions. Torque pulsations are initially observed, but are reduced by iterative re-training of the neural network using experimental data. The resulting motor speed step response (for several forward and reverse step commands) quickly tracks the expected response, with negligible error under steady-state conditions.     

弹性自适应人工鱼群-BP神经网络模型及在短期电价预测中的应用

为了提高短期电价预测精度,分析了人工鱼群算法及其缺点,提出了一种弹性自适应人工鱼群算法(RAAFSA).应用RAAFSA算法训练BP神经网络,实现了BP神经网络参数优化,形成弹性自适应人工鱼群-BP神经网络混合算法(RAAFSA-BP),对贵州电网进行短期电价预测.仿真表明,弹性自适应人工鱼群优化的BP神经网络算法收敛速度快于BP神经网络算法和人工鱼群-BP神经网络算法,该混合算法克服了BP神经网络和人工鱼群算法易陷于局部极值、搜索质量差和精度不高的缺点,改善了BP神经网络的泛化能力,输出稳定性好,预报精度显著提高,各日预测电价的平均百分比误差可控制在2%以内,平均绝对误差最大值为1.762$/MWh.该混合算法可有效用于电力市场短期电价预测.     

Daily peak electric load forecasting using an artificial neural network and an improvement method for reducing the forecasting errors

This paper proposes a forecasting method for shortterm peak electric loads using a 3-layer neural network of locally active units. Each unit in the hidden layer of the neural network is activated only by input vectors in a bounded domain of vector space. This characteristic enables additional learning. Furthermore, it is supposed to provide the network structure with information that helps to improve forecasting accuracy. The neural network is applied to daily peak load forecasting simulations in summer. The results show that the proposed method is superior to a conventional neural network with the backpropagation algorithm. To make the best use of the neural network, an error-oriented method of parameter modification is also examined.     

An artificial neural network based digital differential protectionscheme for synchronous generator stator winding protection

This paper describes a new artificial neural network (ANN) based digital differential protection scheme for generator stator winding protection. The scheme includes two feedforward neural networks (FNNs). One ANN is used for fault detection and the other is used for internal fault classification. This design uses current samples from the line-side and the neutral-end in addition to samples from the field current. Fundamental and/or second harmonic present in the field current during a fault help the ANN, used for fault detection, to differentiate between generator states (normal, external fault and internal fault states). Results showing the performance of the protection scheme are presented and indicate that it is fast and reliable     

Detection and classification of power quality disturbances using S-transform and modular neural network

This paper presents an S-transform based modular neural network (NN) classifier for recognition of power quality disturbances. The excellent time—frequency resolution characteristics of the S-transform makes it an attractive candidate for the analysis of power quality (PQ) disturbances under noisy condition and has the ability to detect the disturbance correctly. On the other hand, the performance of wavelet transform (WT) degrades while detecting and localizing the disturbances in the presence of noise. Features extracted by using the S-transform are applied to a modular NN for automatic classification of the PQ disturbances that solves a relatively complex problem by decomposing it into simpler subtasks. Modularity of neural network provides better classification, model complexity reduction and better learning capability, etc. Eleven types of PQ disturbances are considered for the classification. The simulation results show that the combination of the S-transform and a modular NN can effectively detect and classify different power quality disturbances.     

一种基于自适应神经网络的传感器故障检测方法

传感器的故障检测技术是自动控制系统能否正确并运行的先提条件,针对系统存在未知输入的特点,提出了一种新的检测方法,预测出正常情况下传感器的状况,并进行了大量仿真实验研究。与传感器实际输出比较,实验结果证明该故障检测方法的有效性和实用性。     

基于动态面控制的间接自适应神经网络块控制

针对一类可转化为"标准块控制形"的多输入多输出的非线性系统,基于动态面控制技术,提出一种间接自适应神经网络控制器的设计方案.该方法通过引入1阶滤波器,消除了后推设计中由于反复对虚拟控制的求导而导致的复杂性问题,同时完全避免了反馈线性化方法中可能出现的控制器奇异性问题,且无需控制增益矩阵正定、可逆的条件.利用李亚普诺夫方法,证明了闭环系统是半全局一致终结有界,通过适当选取设计常数,跟踪误差可收敛到原点的一个小邻域内.仿真结果表明所提控制方法的有效性.     

永磁同步电动机自适应神经网络IP位置控制器

针对永磁同步电动机伺服系统（PMSM）存在结构复杂以及系统性能受不确定性影响严重等问题,在IP位置控制系统上,提出了神经网络补偿器的方法。该方法利用参考模型和系统输出的偏差在线训练神经网络的权值,有效克服了参数变化和负载扰动等不确定因素的影响,而且控制方法简单,动态快速性好。仿真结果表明,采用神经网络补偿器的方法能明显改变系统的品质,增强系统的鲁棒性。     

Evaluation of adaptive neural network models for freeway incident detection

Automated incident detection is an essential component of a modern freeway traffic monitoring system. A number of neural network (NN)-based incident detection models have been tested independently over the past decade. This paper evaluates the adaptability of three promising NN models for this problem: a multilayer feed-forward NN (MLFNN), a basic probabilistic NN (BPNN) and a constructive probabilistic NN (CPNN). These three models have been developed on an original freeway site in Singapore and then adapted to a new freeway site in California. In addition to their incident detection performance, their ability to adapt to new freeway sites, and network sizes have also been compared. A novel updating scheme has been used for adjustment of smoothing parameter of the BPNN. Results of this study show that the MLFNN model has the best incident detection performance at the development site while CPNN model has the best performance after model adaptation at the new site. In addition, the adaptation method for CPNN model is less laborious. The efficient network pruning procedure for the CPNN network resulted in a smaller network size, making it easier to implement it for real-time application. The results suggest that CPNN model has good potential for application in an operational automatic incident detection system for freeways.     

人工神经网络在电力系统负荷预测中的应用

论述了人工神经网络预测电力系统负荷的方法和步骤,并以BP神经网络在石嘴山地区短期负荷预测中的应用为例,探讨负荷预测的重要性。     

Resilience management practices for electric utilities and extreme weather

This paper describes management practices of a model resilient electric utility that can serve as a framework for advancing planning and preparation for extreme weather and climate hazards. The framework focuses on practices grouped into eight domains progressing through five levels of maturity. For each domain, a discussion of resilience management practices is provided along with examples. By assessing its maturity level and taking steps to increase it, a utility can realize increased resilience benefits.