A FUZZY NEURAL NETWORK DECISION MODEL ON THE OPERATION PROCESS OF ELECTRIC FURNACE FOR CLEANING SLAG AND ITS APPLICATION

ＡＦＵＺＺＹＮＥＵＲＡＬＮＥＴＷＯＲＫＤＥＣＩＳＩＯＮＭＯＤＥＬＯＮＴＨＥＯＰＥＲＡＴＩＯＮＰＲＯＣＥＳＳＯＦＥＬＥＣＴＲＩＣＦＵＲＮＡＣＥＦＯＲＣＬＥＡＮＩＮＧＳＬＡＧＡＮＤＩＴＳＡＰＰＬＩＣＡＴＩＯＮ￥Ｐｅｎｇ，Ｘｉａｏｑｉ；Ｍｅｉ，Ｃｈｉ；Ｚｈ...

Analysis and Prediction of Regional Electricity Consumption Based on BP Neural Network

Electricity consumption forecasting is one of the most important tasks for power system workers,and plays an important role in regional power systems.Due to the difference in the trend of power load and the past in the new normal,the influencing factors are more diversified,which makes it more difficult to predict the current electricity consumption.In this paper,the grey system theory and BP neural network are combined to predict the annual electricity consumption in Jiangsu.According to the historical data of annual electricity consumption and the six factors affecting electricity consumption,the gray correlation analysis method is used to screen the important factors,and three factors with large correlation degree are selected as the input parameters of BP neural network.The power forecasting model uses nearly 18 years of data to train and validate the model.The results show that the gray correlation analysis and BP neural network method have higher accuracy in power consumption prediction,and the calculation is more convenient than traditional methods.

A chaos genetic algorithm for optimizing an artificial neural network of predicting silicon content in hot metal

A genetic algorithm based on the nested intervals chaos search (NICGA) hasbeen given. Because the nested intervals chaos search is introduced into the NICGA to initialize thepopulation and to lead the evolution of the population, the NICGA has the advantages of decreasingthe population size, enhancing the local search ability, and improving the computational efficiencyand optimization precision. In a multi4ayer feed forward neural network model for predicting thesilicon content in hot metal, the NICGA was used to optimize the connection weights and thresholdvalues of the neural network to improve the prediction precision. The application results show thatthe precision of predicting the silicon content has been increased.

Application of variable-filtrating technique on fuzzy-reasoning neural network system predicting BOF end-point carbon content

Artificial intelligence techniques have been used to predict basic oxygen furnace(BOF) end-points. However,the main challenge is to effectively reduce the input nodes as too many input nodes in neural network increase complexity,decrease accuracy and slow down the training speed of the network.Simply picking-up variables as input usually influence validity of model.It is quite necessary to develop an effective method to reduce the number of input nodes whereby to simplify the network and improve model performance.In this study,a variable-filtrating technique combining both metallurgical mechanism model and partial least-squares(PLS ) regression method has been proposed by taking the advantages of both of them,i.e.qualitive and quantative relationships between variables respectively.Accordingly,a fuzzy-reasoning neural network(FNN) prediction model for basic oxygen furnace(BOF) end-point carbon content based on this technique has been developed.The prediction results showed that this model can effectively improve the hit rate of end-point carbon content and increase network training speed.The successful hit rate of the model can reach up to 94.12%with about 0.02% error range.

A Prediction Method of Charging Station Planning Based on BP Neural Network

The construction of charging service facilities is a very important factor in the popularization of electric vehicles. Therefore, the planning problems of electric vehicle charging station are urgent to be solved. Considering the standard of natural environment, society, traffic, power grid and economy, an evaluation system is created for electric vehicle charging station project through 15 sub-standards. Planning model of charging station is constructed based on BP neural network adopted in the analysis. It is used for location and capacity prediction of charging station planning. By analyzing the model with data samples, a stable network structure is established and the feasibility of the model is verified in the charging station planning.

AN INTELLIGENT DECISION SUPPORT SYSTEM （IDSS） IN THE OPERATION PROCESS OF ELECTRIC FURNACE FOR CLEANING SLAG

ＡＮＩＮＴＥＬＬＩＧＥＮＴＤＥＣＩＳＩＯＮＳＵＰＰＯＲＴＳＹＳＴＥＭ（ＩＤＳＳ）ＩＮＴＨＥＯＰＥＲＡＴＩＯＮＰＲＯＣＥＳＳＯＦＥＬＥＣＴＲＩＣＦＵＲＮＡＣＥＦＯＲＣＬＥＡＮＩＮＧＳＬＡＧＰｅｎｇＸｉａｏｑｉＭｅｉＣｈｉＺｈｏｕＪｉｅｍｉｎ（Ｄｅｐａｒ...

Designing Evolutionary Wavelet Neural Network for Estimating Foaming Slag Quality in Electric Arc Furnace Using Power Quality Indices

In the present study,a novel approach based on an evolutionary wavelet neural network(EWNN)is proposed to estimate the slag quality in an electric arc furnace(EAF)employing power quality indices.In the EWNN,an evolutionary method is applied to train the parameters for a combination of neural networks and wavelets.I For this purpose,all of the electrical parameters for six melting processes are measured with a power quality analyzer,attached to the secondary component of an EAF transformer at a Saba steel complex,to estimate the foaming slag quality.Experimental results on various combinations of measured electrical parameters,applying the designed EWNN estimator,demonstrate that utilizing five leading indicators leads to the highest precision.The obtained 99%accuracy for estimating the foaming slag quality by EWNN compared to the other methods illustrates the proposed method's efficiency.

Intelligent predictive model of ventilating capacity of imperial smelt furnace

In order to know the ventilating capacity of imperial smelt furnace(ISF), and increase the output of plumbum, an intelligent modeling method based on gray theory and artificial neural networks(ANN) is proposed, in which the weight values in the integrated model can be adjusted automatically. An intelligent predictive model of the ventilating capacity of the ISF is established and analyzed by the method. The simulation results and industrial applications demonstrate that the predictive model is close to the real plant, the relative predictive error is 0.72%, which is 50% less than the single model, leading to a notable increase of the output of plumbum.

Genetic algorithm-determined deep feedforward neural network architecture for predicting electricity consumption in real buildings

A genetic algorithm-determined deep feedforward neural network architecture(GA-DFNN)is proposed for both day-ahead hourly and week-ahead daily electricity consumption of a real-world campus building in the United Kingdom.Due to the comprehensive relationship between affecting factors and real-world building electricity consumption,the adoption of multiple hidden layers in the deep neural network(DFNN)algorithm would improve its prediction accuracy.The architecture of a DFNN model mainly refers to its quantity of hidden layers,quantity of neurons in the hidden layers,activation function in each layer and learning process to obtain the connecting weights.The optimal architecture of DFNN model was generally determined through a trial-and-error process,which is an exponential combinatorial problem and a tedious task.To address this problem,genetic algorithm(GA)is adopted to automatically design an optimal architecture with improved generalization ability.One year and six months of measurement data from a campus building is used for training and testing the proposed GA-DFNN model,respectively.To demonstrate the effectiveness of the proposed GA-DFNN prediction model,its prediction performance,including mean absolute percentage error,coefficient of determination,root mean square error and mean absolute error,was compared to the reference feedforward neural network models with single hidden layer,DFNN models with other architecture,random search determined DFNN model,long-short-term-memory model and temporal convolutional network model.The comparison results show that the proposed GA-DFNN predictive model has superior performance than all the reference prediction models,demonstrating the optimization effectiveness of GA and the prediction effectiveness of DFNN model with multiple hidden layers and optimal architecture.

Machine Learning Empowered Electricity Consumption Prediction

Electricity,being the most efficient secondary energy,contributes for a larger proportion of overall energy usage.Due to a lack of storage for energy resources,over supply will result in energy dissipation and substantial investment waste.Accurate electricity consumption prediction is vital because it allows for the preparation of potential power generation systems to satisfy the growing demands for electrical energy as well as:smart distributed grids,assessing the degree of socioeconomic growth,distributed system design,tariff plans,demand-side management,power generation planning,and providing electricity supply stability by balancing the amount of electricity produced and consumed.This paper proposes amedium-termprediction model that can predict electricity consumption for a given location in Saudi Arabia.Hence,this study implemented a standalone ArtificialNeuralNetwork(ANN)model and bagging ensemble for predicting total monthly electricity consumption in 18 locations across Saudi Arabia.The dataset used in this research is gathered exclusively from the Saudi Electric Company.The pre-processing phase included normalizing the data using min-max method and mapping the cyclical attribute to its sine and cosine facets.The number of neurons and learning rate of the standalone model were optimized using hyperparameter tuning.Finally,the standalone model was tested against the bagging ensemble using the optimized ANN.The bagging ensemble with an optimized ANN as the chosen classifier outperformed the standalone ANN model.The results for the proposed model produced 0.9116 Correlation Coefficient(CC),0.2836 Mean Absolute Percentage Error(MAPE),0.4578,Root Mean Squared Percentage Error(RMSPE),0.0298 MAE,and 0.069 Root Mean Squared Error(RMSE),respectively.

融合工况预测的燃料电池汽车里程自适应等效氢耗最小控制策略

为有效地提高插电式燃料电池汽车的经济性,实现燃料电池和动力电池的功率最优分配,考虑到行驶工况、电池荷电状态(State of charge,SOC)、等效因子与氢气消耗之间的密切联系,制定融合工况预测的里程自适应等效氢耗最小策略.通过基于误差反向传播的神经网络来实现未来短期车速的预测,分析未来车辆需求功率变化,同时借助全球定位系统规划一条通往目的地的路径,智能交通系统便可获取整个行程的交通流量信息,利用行驶里程和SOC实时动态修正等效消耗最小策略中的等效因子,实现能量管理策略的自适应性.基于MATLAB/Simulink软件,搭建整车仿真模型与传统的能量管理策略进行仿真对比验证.仿真结果表明,采用基于神经网络的工况预测算法能够较好地预测未来短期工况,其预测精度相较于马尔可夫方法提高12.5%,所提出的能量管理策略在城市道路循环工况(UDDS)下的氢气消耗比电量消耗维持(CD/CS)策略下降55.6%.硬件在环试验表明,在市郊循环工况(EUDC)下的氢气消耗比CD/CS策略下降26.8%,仿真验证结果表明了所提出的策略相比于CD/CS策略在氢气消耗方面的优越性能,并通过硬件在环实验验证了所提策略的有效性.

基于动态自适应图神经网络的电动汽车充电负荷预测

电动汽车充电站负荷波动的不确定性与长时间预测任务给提升充电负荷预测精度带来巨大的挑战。文中提出一种基于动态自适应图神经网络的电动汽车充电负荷预测算法。首先,构建了一个充电负荷信息时空关联特征提取层,将多头注意力机制与自适应相关图结合生成具有时空关联性的综合特征表达式,以捕获充电站负荷的波动性;然后,将提取的特征输入到时空卷积层,捕获时间和空间之间的耦合关系;最后,通过切比雪夫多项式图卷积与多尺度时间卷积提升模型耦合长时间序列之间的能力。以Palo Alto数据集为例,与现有方法相比,所提算法在4种波动情况下的平均预测误差大幅降低。

Electrode Wear Prediction in Milling Electrical Discharge Machining Based on Radial Basis Function Neural Network

Milling electrical discharge machining(EDM) enables the machining of complex cavities using cylindrical or tubular electrodes.To ensure acceptable machining accuracy the process requires some methods of compensating for electrode wear.Due to the complexity and random nature of the process,existing methods of compensating for such wear usually involve off-line prediction.This paper discusses an innovative model of electrode wear prediction for milling EDM based upon a radial basis function(RBF) network.Data gained from an orthogonal experiment were used to provide training samples for the RBF network.The model established was used to forecast the electrode wear,making it possible to calculate the real-time tool wear in the milling EDM process and,to lay the foundations for dynamic compensation of the electrode wear on-line.This paper demonstrates that by using this model prediction errors can be controlled within 8%.

Prediction of office building electricity demand using artificial neural network by splitting the time horizon for different occupancy rates

Due to the impact of occupants’activities in buildings,the relationship between electricity demand and ambient temperature will show different trends in the long-term and short-term,which show seasonal variation and hourly variation,respectively.This makes it difficult for conventional data fitting methods to accurately predict the long-term and short-term power demand of buildings at the same time.In order to solve this problem,this paper proposes two approaches for fitting and predicting the electricity demand of office buildings.The first proposed approach splits the electricity demand data into fixed time periods,containing working hours and non-working hours,to reduce the impact of occupants’activities.After finding the most sensitive weather variable to non-working hour electricity demand,the building baseload and occupant activities can be predicted separately.The second proposed approach uses the artificial neural network(ANN)and fuzzy logic techniques to fit the building baseload,peak load,and occupancy rate with multi-variables of weather variables.In this approach,the power demand data is split into a narrower time range as no-occupancy hours,full-occupancy hours,and fuzzy hours between them,in which the occupancy rate is varying depending on the time and weather variables.The proposed approaches are verified by the real data from the University of Glasgow as a case study.The simulation results show that,compared with the traditional ANN method,both proposed approaches have less root-mean-square-error(RMSE)in predicting electricity demand.In addition,the proposed working and non-working hour based regression approach reduces the average RMSE by 35%,while the ANN with fuzzy hours based approach reduces the average RMSE by 42%,comparing with the traditional power demand prediction method.In addition,the second proposed approach can provide more information for building energy management,including the predicted baseload,peak load,and occupancy rate,without requiring additional building parameters.

基于全影响因素的轧钢加热炉板坯单耗预测

板坯实际生产过程中单耗计算受原料和燃料条件、操作工艺、钢种等因素影响,且各因素与板坯单耗之间的映射关系较为复杂。文章采用BP神经网络建立板坯单耗预测模型,以板坯加热炉实际生产数据为研究对象,加热过程中涉及的全部影响因素共17项作为输入变量,建立板坯单耗计算预测模型。结合试错法确定合理的BP神经网络结构为:输入层节点数为17,隐藏层节点数为10,输出层节点数为1。预测结果显示单耗预测值与实际值趋势一致,预测均方根误差仅为0.181 GJ/t,模型整体精度可达92.06%。

基于改进LSTM神经网络的电动汽车充电负荷预测

当前对电动汽车(EV)充电负荷预测的研究缺少真实的数据支撑,并且模型考虑场景过于简单,影响因素考虑不到位,预测结果缺乏说服力。基于此,提出一种考虑多种电动汽车充电负荷影响因素的电动汽车充电负荷预测方法。首先,考虑天气、季节、温度、工作日、节假日等因素对电动汽车充电负荷的影响,采用三标度层次分析法分析各影响因素权重;其次,建立LSTM神经网络预测模型,通过真实数据训练得到用于预测的LSTM神经网络模型,结合影响因素权重分析结果对预测模型进行修正,得到最终的改进LSTM神经网络负荷预测模型;最后,采用常州某小区的真实数据对所提预测方法进行试验验证。结果表明,所提方法可以实现电动汽车充电负荷的精确预测,且负荷预测结果可为有序充电策略研究提供参考。

基于需求功率预测的电动拖拉机能量管理策略

针对电动拖拉机在犁耕工况下电机需求电流波动比较大的特点,为了改善动力电池的输出电流过高或过低及电动拖拉机犁耕持续作业时间短的现象,利用超级电容高功率密度的特点,设计了一种锂电池+超级电容结构的双电源电动拖拉机,并建立了Amesim/Simulink联合仿真模型。以模型预测控制作为双电源系统的能量管理方法,基于长短期记忆神经网络建立电动拖拉机犁耕工况下的需求功率预测模型,使用动态规划算法求解最佳的锂电池输出电流。仿真结果表明:相比于模糊控制策略,基于模型预测控制策略有效降低了锂电池大电流放电的频率且峰值电流降低了40%,有效提高了锂电池的使用寿命;超级电容的SOC保持在比较高的范围内,且电动拖拉机在犁耕工况下的单位里程能量消耗降低了2.17%,实现了双电源电流分配最优,提高了电动拖拉机的动力性和经济性。

自回归神经网络的预测值反馈再训练策略及应用

为提高非线性自回归神经网络(NARX-NN)的多步预测性能,提出了一种预测值反馈再训练(FR)策略.首先采用常规训练策略对NARX-NN进行训练,然后利用模型的单步预测结果替换实测值,得到重构训练集,并指导网络再次训练.为验证FR的有效性,将其应用于3种典型的NARX-NN模型:非线性自回归深度神经网络(NARX-DNN)、基于长短期记忆网络的编码器-解码器(LSTMED)和深度自回归网络(DeepAR),以预测燃煤锅炉NO_(x)质量浓度或综合能源系统电负荷.与常规训练策略和计划采样的对比结果表明,采用FR的NARX-NN具有最高的多步预测精度,其中,LSTMED对NO_(x)质量浓度前向15步预测的平均绝对百分比误差(MAPE)为4.01%;DeepAR对电负荷前向24步预测的平均MAPE为4.34%.配对样本T检验结果表明,FR对NARX-NN的多步预测性能提升具有显著性.通过保持训练阶段和预测阶段输入的一致性,FR有效提升了NARX-NN模型的多步预测精度.

基于LSTM⁃MIV神经网络的SF_(6)断路器触头电寿命预测

文中基于LW30⁃40.5型SF_(6)断路器全寿命烧蚀试验数据,采用归一化异常指数F反映断路器触头剩余电寿命,0表示电寿命完备,1表示电寿命耗尽。根据断路器分闸动态电阻—行程曲线,提出了将动态电阻—行程曲线上行程大于8 mm且电阻值大于500μΩ的第一个峰值点作为断路器主触头接触阶段与弧触头接触阶段分界点。据此得到反映断路器触头异常指数F的5个特征参数,采用皮尔逊相关系数法对特征参数进行独立性分析。通过LSTM神经网络建立了基于特征参数与异常指数F之间的回归预测模型,F回归预测均方差和绝对平均误差分别为0.014921和0.013053,采用MIV方法确定了各特征参数对异常指数F影响权重。

基于LSTM神经网络的牵引站电气设备耦联体系地震响应预测

铁路牵引变电站中,软导线-电气设备耦联体系具有较强的几何非线性。为提升系统分析效率,提出一种改进的软导线-电气设备耦联体系地震响应递归预测方法。基于长短期记忆(long short-term memory,LSTM)神经网络与Dropout防止过拟合技术搭建了LSTM神经网络预测模型。建立了充分考虑软导线对相邻设备的耦联作用的软导线-电气设备耦联体系理论分析模型。为验证预测模型的泛化能力,筛选出了41条在峰值、频谱和持续时间上具有较大差异的地震波。并按照递归方案,将选取的地震波以及软导线-电气设备耦联体系理论分析模型计算所得的位移响应,进行滑动切片处理,建立模型输入特征与输出响应标签的映射关系。在此基础上,利用该LSTM神经网络预测模型开展了软导线-电气设备耦联体系设备的地震位移响应预测,并采用多个评价指标进行较为全面的模型性能评估。研究结果表明:LSTM递归预测模型具有良好的地震响应预测性能,搭配Dropout技术能够有效防止模型训练过拟合,提高模型适应能力。对于差异较大的地震波数据,均能够快速预测出误差较小、相关度较高的地震响应,具有较好的准确性、高效性与泛化能力。所提方法能够较高效准确地预测任意时刻的软导线-电气设备耦联体系地震响应,为铁路牵引变电站抗震设计提供新的研究思路。