基于改进RBF碟式太阳集热器出口气流温度预测

为了改善传统的数学建模预测集热器出口气流温度较为复杂的问题,采用RBF(径向基函数)神经网络建立温度预测模型。考虑到碟式太阳集热器出口气流温度影响因素较多,在大量历史实验数据基础上,分析其主要影响因素。为提高预测精度,提出一种自适应聚类算法对RBF神经网络进行改进,并利用Kennard-Stone选取方法(简称K-S法)的空间均匀分布原理提高训练样本质量。根据碟式太阳能光热系统的实测数据对所提的模型进行测试,结果表明,改进算法可进一步提高预测精度和效率,验证了改进算法的可行性和有效性。     

基于GRA-BPNN时变权重的光伏短期出力组合预测

针对并网情况下光伏出力预测精度低和稳定性差问题,提出了一种基于灰色关联分析(GRA)结合BP神经网络(GRA-BPNN)的变权重系数组合预测模型。首先,利用3种单一预测模型对光伏出力分别进行预测,然后,利用GRA-BPNN模型对3个单一模型不同时刻的权重系数进行预测,最后,根据权重系数计算出预测结果。文章利用武汉某并网光伏电站的实测数据对GRA-BPNN变权重组合预测模型预测结果的准确性进行检验。分析结果表明:GRA-BPNN变权重组合预测模型的相对均方根误差和相对平均误差均低于单一模型和等权重组合模型;根据各预测模型的残差直方图可知,GRA-BPNN变权重组合预测模型预测结果中出现较大残差的概率很小,有效地解决了单一模型预测结果不稳定的问题。     

梯度提升树算法在陕北风电场短期风电功率预测中的应用

为了对地形和气候条件复杂的陕北风电场短期风电功率进行准确预测,通过将（weather research and forecasting,WRF）模式输出结果和同期实测风电功率资料相结合,利用梯度提升树算法进行预报气象场和实测风电功率之间的统计关系分析,从而建立了一套陕北风电场短期风电功率预测模型。以陕北靖边某风电场为例,预测结果表明：所提模型年平均预测准确率伟15.7%;月平均归一化均方根误差在20%以下。模型对风电场风电功率预测精度较好。     

街谷空气温度预测模型的改进研究

介绍了针对街谷空气温度预测的建筑群热时间常数模型的改进研究,通过分析原模型的不足,从二维模型向三维模型扩展、地表温度准确计算模型、长波辐射温降计算方法改进、乔木下长波辐射计算和绿地对流换热修正因子计算方法5方面进行改进。利用改进模型对广州某居住小区室外空气温度预测进行计算,结果表明预测与实测有较高的一致性,相关度为0.99。该模型可用于小区微气候规划设计和城市不同区域建筑能耗准确预测,具有重要意义。     

秸秆类生物质灰软化温度的灰预测模型研究

根据秸秆类生物质灰成分和灰软化温度实测数据构成的数据样本,采用灰色关联度分析法研究灰成分对灰软化温度的影响,建立灰软化温度GM(0,8)预测模型。结果表明:该预测模型预测精度较高,适用于预测秸秆类生物质灰软化温度,具有工程应用价值。     

高海拔荒漠地区硅电池板输出电流预测模型研究

摘要： 高海拔荒漠地区具有环境温度低并且日照强度高的气候特点,现有光伏组件模型无法实现精确预测。在分析硅电池板等效电路的基础上,计算得到了输出电流的预测模型。提出了硅电池板开路电压、短路电流、最大电流和最大电压的修正公式,从而建立起适用高海拔荒漠环境下的硅电池板输出电流预测模型。利用I-V曲线测试仪,在格尔木地区采集了单块多晶硅的I-U曲线和P-U曲线,并记录了实时的环境温度和日照强度。采用所提出模型对输出电流和功率进行预测,并与实验I-U曲线和P-U曲线对比。模型仿真与实测输出电流值和功率值的相对误差均小于9.8%,能较好地预测高海拔荒漠环境下的硅电池板输出电流和输出功率。     

基于改进极限学习机ELM的光伏发电预测方法研究

提升光伏发电预测的准确性有助于电网的调度管理和经济运行,其关键在于建立高效适用的预测模型。首先采用皮尔逊（Person）相关系数法对影响光伏出力的主要因素进行筛选,确立预测模型的输入特征向量;提出了一种融合Person相关系数法与遗传算法（GA）优化的ELM混合预测模型,并对预测模型中随机生成的参数进行了优化;以某一光伏电站的历史数据为研究对象,采用GA-ELM预测模型对预处理后的数据进行训练和测试,基于模型开展了四个季节典型日的光伏发电功率预测。结果表明：混合预测模型比单一的ELM预测模型和Person相关系数与ELM混合的预测模型的预估偏差率分别降低了19.2%和4.3%,验证了本文模型具有更高的准确性和稳定性。     

RBF神经网络在特高拱坝混凝土浇筑仓最高温度预测中的应用

准确预测特高拱坝混凝土浇筑仓最高温度对其温控防裂具有重要意义。基于RBF神经网络,利用Matlab神经网络工具箱建立了特高拱坝混凝土浇筑仓最高温度预测模型,以浇筑温度、水管布置、通水水温、通水流量及现场环境气温作为输入层,以浇筑仓最高温度为输出层,并通过高温季节、低温季节工程实测数据分别对模型进行验证。结果表明,该模型简单有效、精度高,可满足预测要求。     

风电场风功率实时预测效果综合评价方法

超短期风电功率预测对含大规模风电的电力系统安全经济运行有着重要意义。但目前对预测结果的评价均停留在常规统计学指标上,缺乏合理的评价体系来评价某特定风电场所选取预测模型的优劣。简述了目前风电功率预测结果评价指标的不足,提出一种基于预测误差评价和预报考核等指标的风电场输出功率实时预测效果评估方法,为不同地区风电场根据其风电输出功率变化的特点,选择预测模型以及风电场输出功率预测效果的工程检验提供依据。最后,利用吉林省某风电场实测数据,采用该评估方法对不同预测模型的实时预测结果进行分析评价,实现了该风电场不同预测模型间的择优,验证了该评价方法的指导价值。     

选择预报因子的新方法及其对预测的影响分析

针对选择适合预报因子预测负荷时采用相关系数法仅能确定两个随机变量之间线性关系的问题,采用相关概率法选择预报因子构建预测模型,分别考虑了单因素与双因素,并利用最小二乘支持向量机法预测电力负荷。结果表明,该方法精度较高,预测结果较好。     

Estimation for inner surface geometry of furnace wall using inverse process combined with grey prediction model

In this work the inner surface geometry of a cylindrical furnace wall is estimated using inverse process method combined with grey prediction model. In estimating process a virtual area extended from the inner surface of furnace wall is used for analysis. The heat conduction equation and the boundary condition are first discretized by finite difference method to form a linear matrix equation; the inverse model is then optimized by linear least-squares error method and the temperatures of virtual boundary are obtained from a few of measured temperatures in furnace wall using the linear inverse model; and finally the temperature distribution of system is got by direct process and the inner surface geometry of furnace wall can be estimated accordingly. The result shows that using inverse process combined with grey prediction model the geometry can be exactly estimated from relatively small number of measured temperatures. Moreover, the effects of measurement error, location, and number of measured points on the estimation for inner surface geometry of furnace wall are discussed in detail.     

基于CEEMDAN-KELM的锅炉对流受热面状态预测研究

为了解决表征锅炉受热面表面健康状态的清洁因子在未来时间段内预测时呈现非平稳问题，以省煤器受热面为例，提出一种结合核极限学习机(Kernel Extreme Learning Machine,KELM)和自适应噪声完备集成经验模态分解(Complete Ensemble Empirical Mode Decomposition with Adaptive Noise,CEEMDAN)的清洁因子预测方法。首先，通过CEEMDAN分解算法对省煤器表面清洁因子序列进行分解和降低复杂程度，获得各固有模态函数(Intrinsic Mode Function,IMF)；其次，利用皮尔逊相关性分析确定主蒸汽流量、进出口烟温等9个参数为输入，建立核极限学习机模型对清洁因子的各IMF进行预测；最后，将各IMF预测结果相加获得最终预测结果。结果表明：与基本核极限学习机、支持向量机等预测模型相比，本文模型具有较高的预测精度和较优预测时间，可为基于受热面状态开展的锅炉智慧吹灰应用提供参考。     

基于系统辨识的地表太阳辐射Box-Jenkins模型的建立

利用天文辐射作为输入数据,采用系统辨识的方法得到地表太阳辐射的BJ(Box-Jenkins)模型,并通过残差分析和零极点检验。该方法可用于预测5～15min时间间隔的地表太阳辐射,为太阳能电站的功率输出预测提供太阳能辐射数据。     

ANN modeling of water consumption in the lead-acid batteries

Due to importance of the quantity of water loss in the life cycle of lead-acid batteries, water consumption tests were performed on 72 lead-acid batteries with low antimony grid alloy at different charge voltages and temperatures. Weight loss of batteries was measured during a period of 10 days. The behavior of batteries in different charge voltages and temperatures were modeled by artificial neural networks (ANNs) using MATLAB 7 media. Four temperatures were used in the training set, out of which three were used in prediction set and one in validation set. The network was trained by training and prediction data sets, and then was used for predicting water consumption in all three temperatures of prediction set. Finally, the network obtained was verified while being used in predicting water loss in defined temperatures of validation set. To achieve a better evaluation of the model ability, three models with different validation temperatures were used (model 1 = 50 °C, model 2 = 60 °C and model 3 = 70 °C). There was a good agreement between predicted and experimental results at prediction and validation sets for all the models.     

基于GRU-RF模型的太阳辐照度短时预测

针对现有太阳辐照度短期预测方法的建模复杂、准确度低等问题,提出一种基于深度学习的GRU-RF动态权值组合预测方法。大气因素与太阳辐照度数据融合,将运算速度较快且模型复杂度较低的随机森林（RF）模型与带有时序记忆的门控循环单元（GRU）神经网络进行动态权值的加权集成,分别将地表接收到的太阳辐照度、近地层气温、相对湿度、近地层风速和相对气压等变化特征进行预测研究。通过几种模型对比分析,结果表明使用GRU-RF模型预测短时（9 h）太阳辐照度结果较好,运行速度较快,在不同时间间隔（5、10以及15 min）下能够很好地预测太阳辐照度数据。     

A validated methodology for the prediction of heating and cooling energy demand for buildings within the Urban Heat Island: Case-study of London

This paper describes a method for predicting air temperatures within the Urban Heat Island at discreet locations based on input data from one meteorological station for the time the prediction is required and historic measured air temperatures within the city. It uses London as a case-study to describe the method and its applications. The prediction model is based on Artificial Neural Network (ANN) modelling and it is termed the London Site Specific Air Temperature (LSSAT) predictor. The temporal and spatial validity of the model was tested using data measured 8 years later from the original dataset; it was found that site specific hourly air temperature prediction provides acceptable accuracy and improves considerably for average monthly values. It thus is a very reliable tool for use as part of the process of predicting heating and cooling loads for urban buildings. This is illustrated by the computation of Heating Degree Days (HDD) and Cooling Degree Hours (CDH) for a West–East Transect within London. The described method could be used for any city for which historic hourly air temperatures are available for a number of locations; for example air pollution measuring sites, common in many cities, typically measure air temperature on an hourly basis.     

生物质成型机成型模孔内表面摩擦力分形预测

选择2种模孔材料,利用粗糙仪测出其表面形貌,提取轮廓数据,采用尺码法计算出表面分形参数D和G,再基于经典接触公式和M-B分形接触模型建立成型模孔内表面滑动摩擦力分形预测模型并进行模拟,最后利用摩擦磨损试验台进行验证。结果表明：当环模模孔内表面的粗糙度R_a值3.428 μm时（45钢）,D值为1.373,G值为1.55×10_-6 m;而当其R_a值为4.002 μm时（40Cr）,D值为1.359,G值为0.82×10_-6 m;摩擦力随真实接触面积的增加呈增大趋势,单位摩擦力随真实接触面积的增加呈减小趋势;在A_rrc时,摩擦力增大速度较快,单位摩擦力减小速度较快;在A_r>A_rc时,摩擦力增幅较平缓,单位摩擦力降幅较平缓。经实验验证,摩擦力模拟结果比试验结果略高,且随真实接触面积的增大,摩擦力的模拟值正逐渐接近实验值,模型预测较准确。     

Investigation of radiative contribution in a high temperature fluidized-bed using the alternate-slab model

The modified alternate-slab model of Gabor is examined for the prediction of radiative contribution to the total heat transfer from a high temperature fluidized-bed system of air-sand to an immersed surface. The results are compared with the predictions of other models and experimental data on average heat transfer coefficient, and percentage radiative contribution as a function of various influencing parameters. The heat transfer coefficients are overestimated by the model within reasonable limits and approach the experimental data for high values of heat transfer surface temperature. The percentage radiative contribution is substantial for large values of particle diameter, surface and bed temperatures. The model is found reliable and simple to handle over a wide temperature range. Results are also presented for air-ash and air-dolomite systems in view of their practical significance.     

Optimization of fermentative hydrogen production process using genetic algorithm based on neural network and response surface methodology

A central composite design was carried out to investigate the effect of temperature, initial pH and glucose concentration on fermentative hydrogen production by mixed cultures in batch test. The modeling abilities of the response surface methodology model and neural network model, as well as the optimizing abilities of response surface methodology and the genetic algorithm based on a neural network model were compared. The results showed that the root mean square error and the standard error of prediction for the neural network model were much smaller than those for the response surface methodology model, indicting that the neural network model had a much higher modeling ability than the response surface methodology model. The maximum hydrogen yield of 289.8 mL/g glucose identified by response surface methodology was a little lower than that of 360.5 mL/g glucose identified by the genetic algorithm based on a neural network model, indicating that the genetic algorithm based on a neural network model had a much higher optimizing ability than the response surface methodology. Thus, the genetic algorithm based on a neural network model is a better optimization method than response surface methodology and is recommended to be used during the optimization of fermentative hydrogen production process.     

基于深度学习的翼型气动弹片流场参数预测

弹片是解决翼型流动分离的重要技术手段,合理的弹片参数对翼型表面压力分布尤为重要。基于数据驱动的深度学习方法与计算流体力学(Computational Fluid Dynamics, CFD)相结合,可快速有效地完成对复杂流场特征的识别与提取。本文提出一种基于卷积神经网络(Convolutional Neural Network, CNN)的翼型表面压力分布预测方法,通过提取流场的尾流速度、压力等流动特征构建翼型表面压力分布的预测模型。首先,通过数值模拟计算了8种不同抬起角度的NACA 0012弹片翼型的流场;其次,采用提取的流场数据建立CNN预测模型;最后,将预测值和CFD计算值进行对比。结果表明：基于CNN的预测模型对翼型表面压力系数分布有较高的预测精度,其中尾流速度模型在弹片抬起角度为15°时的预测均方根误差仅为0.1,说明尾流速度中包含丰富的流场信息。