Confidence intervals for neural network based short-term loadforecasting

Using traditional statistical models, like ARMA and multilinear regression, confidence intervals can be computed for the short-term electric load forecasting, assuming that the forecast errors are independent and Gaussian distributed. In this paper, the 1 to 24 steps ahead load forecasts are obtained through multilayer perceptrons trained by the backpropagation algorithm. Three techniques for the computation of confidence intervals for this neural network based short-term load forecasting are presented: (1) error output; (2) resampling; and (3) multilinear regression adapted to neural networks. A comparison of the three techniques is performed through simulations of online forecasting     

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     

New models for integrated short-term forward electricity markets

In a typical short-term forward wholesale electricity market where products are auctioned sequentially, one often observes significant market inefficiency and price volatility-thus the growing impetus in developing integrated short-term forward markets where electric energy, reserves, and transmission capacity are auctioned simultaneously. Such markets need new computational methods and models for determining market clearing prices and physical (delivery/consumption) schedules. The purpose of this paper is to examine key aspects of current modeling and pricing methods in short-term forward wholesale electricity markets and to introduce new models suitable for clearing price-based markets of integrated trades of energy, reserve, and transmission. Specifically, an analysis of the impacts of various pricing rules and bidding requirements on market operations is presented, the selection of optimization objectives is discussed, and a new model of transmission congestion and multiproducts simultaneous auction is introduced. Examples are used where appropriate.     

考虑分时电价的居民用户短期用电量分类预测及修正方法

为了更好地利用智能电网中的用户用电信息,提高短期用电量预测精度,针对居民用户提出一种考虑分时电价的分类短期用电量预测及修正方法。首先,通过模糊聚类将用户按用电行为分类,将电价、用电量和加权气象日期影响因素作为预测模型输入量。然后,针对各类用户的用电特点,经仿真对比选择相适应的BP、Elman、LSTM神经网络算法构建预测模型。最后,运用修正算法对误差较大的峰谷值进行修正,将修正后的分类预测结果相加以获得整体预测值。以广东省云浮市某小区为例对该方法进行仿真分析,并与随机森林、CART等算法进行对比。实验结果证明所提方法具有更高的预测精度。     

Some analytical results pertaining to Cournot models for short-term electricity markets

This paper provides some theoretical results pertaining to the Cournot model applied to short-term electricity markets. Price, quantities and profits are first obtained, and then results related to sensitivities and limit values are derived and discussed. The cases of both several identical Cournot producers and one dominant Cournot producer are analyzed. A case example illustrates the results obtained.     

一种新型短期电力负荷组合预测方法

短期负荷预测的准确性对电力系统的安全运行和生产安排起着至关重要的作用。针对单一负荷预测方法较难准确预测电力负荷的情况,提出了一种新颖的组合预测方法。该方法采用混沌时间序列和基于RBF-ARX模型的预测方法对日负荷进行加权组合预测,并通过与单个预测方法进行比较,证明了该组合预测方法精度更高,相对误差更小。     

A proposed grey model for short-term electricity price forecasting in competitive power markets

In order to improve the performance of traditional grey models for short-term electricity price prediction in competitive power markets, a novel grey model, named PGM(1,2,a,b), is proposed in this paper. In the proposed model, the reference sequence is determined with correlation method. Furthermore, considering the limitation of the least square method (LSM), particle swarm optimization algorithm (PSO) is adopted to identify the parameters instead of LSM. To demonstrate the superiority of the proposed model, public available data coming from Nordpool, California, and Ontario power markets were used for training and testing. Simulation results show that the proposed model is capable of forecasting short-term electricity price efficiently and accurately.     

基于注意力机制的CNN-LSTM短期电价预测

短期电价预测结果的准确性对存在多元化竞争格局的电力市场具有重要意义。为提高在电价跳跃点和尖峰点的预测精度及预测效率,针对多因素融合影响的电价序列与其影响因素间隐含的非线性关系,提出了一种基于ATT-CNN-LSTM的短期电价预测方法。首先,采用灰色关联度分析法分析负荷因素与电价之间的关联程度,筛选出关联度较高的数据作为最优模型输入。其次,通过注意力机制(Attention, ATT)自适应分配输入数据的权重,以权重大小区分强弱特征数据。再利用卷积神经网络(Convolution Neural Networks, CNN)对数据集进行二次特征提取及降维处理,优化输入长短期记忆神经网络(LongShort-TermMemory, LSTM)中的数据,从而提升LSTM网络的预测精度与训练速度。对澳大利亚电力市场的实测数据进行算例分析,通过与其他主流算法对比,验证了所提方法具有更高的预测精度和计算效率。     

An electricity efficiency index for a regional economy

The purpose of this research is to develop a “macro” index which measures changes in electricity efficiency over time at the regional level for three major sectors: residential, commercial, and industrial. There is no attempt to link these to specific programs but rather to document all changes in efficiency over time. The results indicate that electricity efficiency in the Northwest increased 33 percent over the twenty-five-year period from 1990–2015.     

An individual welfare maximization algorithm for electricity markets

An algorithm that allows a market participant to maximize its individual welfare in electricity spot markets is presented. The use of the algorithm in determining market equilibrium points, called Nash equilibria, is demonstrated. The start of the algorithm is a spot market model that uses the optimal power flow (OPF), with a full representation of the transmission system and inclusion of consumer bidding. The algorithm utilizes price and dispatch sensitivities, available from the Hessian matrix and gradient of the OPF, to help determine an optimal change in an individual's bid. The algorithm is shown to be successful in determining local welfare maxima, and the prospects for scaling the algorithm up to realistically sized systems are very good. Nash equilibria are investigated assuming all participants attempt to maximize their individual welfare. This is done by iteratively solving the individual welfare maximization algorithm until all individuals stop modifying their bids.     

An automated laser system for eye surgery

The use of lasers in ophthalmology is discussed, and the procedure and requirements of ophthalmic laser surgery are described. An overview of a proposed laser system for eye surgery is given, and its operation is described. Progress and research to date are reviewed.     

An automated learning system for Java programming

The increase in the numbers enrolling in university computer courses makes huge demands on resources, and maintaining standards of teaching and tutorial support is almost impossible without a massive increase in staff. To overcome this situation, the authors have developed RoboProf, an automated learning environment which, as well as generating and assessing programming exercises, provides ongoing assistance and feedback to students without extra demands on lecturer and tutors' time. This system also contains a technique for detecting plagiarism, an increasing problem in computing courses worldwide. For this research, RoboProf was used to teach Java programming to a class containing nearly 300 students in the first year of a computing degree. Use of the system by students was monitored and recorded on log files in order to investigate the extent to which usage patterns influence achieved programming skill. An analysis shows that students who complete the set of RoboProf exercises perform significantly better than those who do not. The timeliness in which these exercises are completed relative to other students is significant: early solvers get higher marks, and students solving the problems with fewer attempts get higher marks. Not surprisingly, plagiarists achieve a lower score than those who do their own work. Other factors that were found to influence programming performance included entry standards and gender. Entry qualifications impacted positively on performance, and males performed significantly better than females. There was a significant positive correlation between the score achieved in the RoboProf course and the performance in a subsequent computing course administered in the traditional manner.     

An automated microwave measurements experience for undergraduates

Experiments in microwave measurements for an undergraduate laboratory course are described that bridge the gap between point-by-point manual measurements and automatic network analyzer measurements. In the sequence of 10 laboratory experiments, the basic parameters of frequency versus wavelength in a waveguide are explored, first in point-by-point measurements versus frequency. Swept frequency measurements are taken up next in the form of an open-ended design experiment investigating a simple two-pole band pass filter. Following this, scalar transmission and reflection measurements are made using a PC as a controller and data-acquisition unit. Lastly, the Hewlett-Packard 8510B Automatic Network Analyzer is used by the students to perform the measurements on the same device as a verification of the earlier results. This paper concentrates on the design of the computer-aided measurements experiment, describing the requirements for the experimental content as well as the peripheral content, such as the controller language (HP BASIC) and the interface (HPIB-HP Interface Bus)     

Coordination between medium-term generation planning and short-term operation in electricity markets

This paper analyzes the coordination between medium-term generation planning and short-term operation in electricity markets. This coordination is particularly important from a practical point of view in order to guarantee that certain aspects of the operation that arise in the medium-term level are explicitly taken into account: limited-energy resources and obligatory-use resources. Three different approaches are proposed in order to guarantee that short-term decisions made by a generation company are consistent with its operation objectives formulated from a medium-term perspective. These approaches make use of technical and economic signals to coordinate both time scopes: primal information, dual information, and resource-valuation functions. This paper presents the main advantages and drawbacks of the three approaches and applies them to a case study that uses a conjectural-variation-based representation of the market.     

An annual framework for clustering-based pricing for an electricity retailer

In the competitive environment, it is necessary for a retailer to increase his/her profit as much as possible. There are few researches focused on the subjects related to the retailer and the retail market. In addition, those researches have mostly focused on the participation of the retailer in the wholesale market. In order to determine the optimal selling price, the knowledge of how and when consumers use electricity is essential to the retailer. This type of information can be found in load profiles of customers. In this paper, an annual framework for optimal price offering by a retailer is proposed which is based on clustering technique. For this purpose, load profiles of customers are used as their consumption patterns. Also, a profit function is defined as the objective of optimization problem based on the load profile considering conditional value at risk (CVaR) for risk modeling. Also, a new acceptance function is proposed to overcome drawbacks of the traditional ones. The objective function is a mixed-integer nonlinear problem which is solved by GAMS software.     

An automated feedback system for computer organization projects

This paper describes a system, built and refined over the past five years, that automatically analyzes student programs assigned in a computer organization course. The system tests a student's program, then e-mails immediate feedback to the student to assist and encourage the student to continue testing, debugging, and optimizing his or her program. The automated feedback system improves the students' learning experience by allowing and encouraging them to improve their program iteratively until it is correct. The system has also made it possible to add challenging parts to each project, such as optimization and testing, and it has enabled students to meet these challenges. Finally, the system has reduced the grading load of University of Michigan's large classes significantly and helped the instructors handle the rapidly increasing enrollments of the 1990s. Initial experience with the feedback system showed that students depended too heavily on the feedback system as a substitute for their own testing. This problem was addressed by requiring students to submit a comprehensive test suite along with their program and by applying automated feedback techniques to help students learn how to write good test suites. Quantitative iterative feedback has proven to be extremely helpful in teaching students specific concepts about computer organization and general concepts on computer programming and testing.     

基于一种NW-FLNN神经网络的短期电价预测

针对传统神经网络收敛速度慢、容易陷入局部极值的问题,文中提出一种改进型小波神经网络以实现网络全局最优化。首先,将小波神经网络与随机矢量函数连接型网络相融合构建一种新型小波链神经网络(NW-FLNN);其次,以小波基函数作为NW-FLNN的隐含层的传递函数,并利用梯度修正法训练该模型各参数;最后,选用澳大利亚新南威尔士州电价数据作为实验数据集,分别对NW-FLNN神经网络、逆传播BP神经网络与小波神经网络进行预测性能比较。实验结果表明:该新型网络预测模型较BP神经网络与小波神经网络性能更优,可明显减少网络迭代次数与隐层神经元数目,且平均百分比误差最大降低至0. 0317,满足实时性要求。     

An approach for applying blockchain technology in centralized electricity markets

The electricity sector has recently begun to undergo a series of changes to improve the generation, transmission, and energy distribution. New technologies have been testing focused on reducing fossil fuels and implementing new tools to create a more efficient energy sector. Blockchain has emerged as one of the most promising tools due to its wide range of utilities, primarily supporting renewable energy, electric mobility, purchase, and energy sale. This technology has been revolutionary, contributing to the significant renewal of the electricity market model. In electricity markets that are not very mature, you can find some centralized activities and competition—creating a commercial monopoly in centralized activities in favor of a specific group. This article aims to analyze the shortcomings of previous research on blockchain in the electricity sector and, through a case study, propose an approach for applying blockchain in centralized electricity markets. Blockchain technology has valuable points, including the excellent security in the transactions carried out in the purchase and sale of energy through a peer-to-peer network stimulating the creation of a decentralized energy trading system.     

基于BP神经网络的短期负荷预测建模仿真

通过BP神经网络与Matlab相结合,建立起三层四功能单元的BP神经网络短期负荷预测模型,并采用某条线路1年的历史负荷波动数据对模型进行“学习”训练。预测一日24小时负荷数据的Matlab仿真及误差分析结果表明,所构筑的BP神经网络模型具有较高的可靠性和准确性,误差率可以有效地控制在2％以内。BP神经网络模型大大提高了短期负荷预测数据的处理效率与可信性,是研究电力系统经济调度的一种新的非线性建模仿真模型。     

考虑负荷周期性和变化率的短期电价预测

为了提高电价预测精确度以提高其实用价值,在电价预测模型中引入负荷周期性和变化率因素.根据负荷对电价的影响建立基于系统负荷的短期电价预测模型,使用小波分解对负荷和电价数据进行分析处理,采用神经网络的预测方法对短期市场清算电价进行预测.考虑负荷和电价的周期特性,在预测模型输入侧增加了负荷的周期性因素.考虑负荷剧变引起的电价变化,定义综合负荷变化率影响因素并加入模型输入侧来提高预测精确度.预测实例采用实际负荷值为输入,其结果表明引入负荷周期特性和综合负荷变化率因素后预测相对预测误差和单点最大预测误差分别降低35%和28%,有效地提高了模型的预测精确度.