为模拟分钟尺度的太阳辐射波动,根据江苏省常州市2018—2021年逐分钟辐射数据,采用Garson权重算法优化模型输入特征,并引入前10分钟的清晰度指数kt时序数据作为附加特征,建立基于时序数据与MLP神经网络的分钟尺度新分离模型。在此基础上,对Engerer2模型、Starke模型和Yang模型3个最新提出的分钟尺度分离模型进行参数本地优化,并设计测试实验验证。验证结果表明：采用时序数据与MLP神经网络的新模型可有效提取短时间内的太阳辐射波动信息,新模型的归一化均方根误差（enRMSE）为10.690%,新模型精度较Yang模型提高了17.08%。
Abstract
In order to simulate the solar radiation fluctuation on the minute scale, based on the 2018—2021 minute-scale radiation data of Changzhou city, Jiangsu province, the Garson weight algerithm is used to optimize the input characteristics of the model, and the time series data of the first 10-minute clearness index kt are introduced as additional features to establish a new minute-scale separation model based on time series data and MLP neural network. On this basis, the parameters of three newly proposed minute-scale separation models, which are Engerer2 model, Starke model and Yang model, respectively, are locally optimized and verified by test experiments. The verification results show that the new model using time series data and MLP neural network can effectively extract solar radiation fluctuation information in a short time. The normalized root mean square error （enRMSE） of the new model is 10.690%. The accuracy of the new model is 17.08% higher than the Yang model.
关键词
太阳辐射 /
机器学习 /
神经网络 /
分离建模 /
波动性
Key words
solar radiation /
machine learning /
neural network /
separation modeling /
volatility
{{custom_sec.title}}
{{custom_sec.title}}
{{custom_sec.content}}
参考文献
[1] 冯巍, 董宏, 杨柳, 等. 太阳辐射直散分离模型比较研究: 以北京地区为例[J]. 土木建筑与环境工程, 2015, 37(1): 12-17.
FENG W, DONG H, YANG L, et al.Comparison of the fraction model of direct and diffuse solar radiation: taking Beijing area as an example[J]. Journal of civil and environmental engineering, 2015, 37(1): 12-17.
[2] LIU B Y H, JORDAN R C. The interrelationship and characteristic distribution of direct, diffuse and total solar radiation[J]. Solar energy, 1960, 4(3): 1-19.
[3] GUEYMARD C A, RUIZ-ARIAS J A. Extensive worldwide validation and climate sensitivity analysis of direct irradiance predictions from 1-min global irradiance[J]. Solar energy, 2016, 128: 1-30.
[4] ENGERER N A.Minute resolution estimates of the diffuse fraction of global irradiance for southeastern Australia[J]. Solar energy, 2015, 116: 215-237.
[5] INMAN R H, CHU Y H, COIMBRA C F M. Cloud enhancement of global horizontal irradiance in California and Hawaii[J]. Solar energy, 2016, 130: 128-138.
[6] 蒋俊霞, 高晓清, 吕清泉, 等. 基于地基云图的云跟踪与太阳辐照度超短期预报方法研究[J]. 太阳能学报, 2020, 41(5): 351-358.
JIANG J X, GAO X Q, LYU Q Q, et al.Study on cloud tracking and solar irradiance ultra-short-term forecasting based on TSI images[J]. Acta energiae solaris sinica, 2020, 41(5): 351-358.
[7] CHEN X Y, DU Y, LIM E, et al.Sensor network based PV power nowcasting with spatio-temporal preselection for grid-friendly control[J]. Applied energy, 2019, 255:113760.
[8] STARKE A R, LEMOS L F L, BOLAND J, et al. Resolution of the cloud enhancement problem for one-minute diffuse radiation prediction[J]. Renewable energy, 2018, 125: 472-484.
[9] YANG D Z, BOLAND J.Satellite-augmented diffuse solar radiation separation models[J]. Journal of renewable and sustainable energy, 2019, 11(2): 023705.
[10] YANG D Z.Temporal-resolution cascade model for separation of 1-min beam and diffuse irradiance[J]. Journal of renewable and sustainable energy, 2021, 13(5): 056101.
[11] BOLAND J, HUANG J, RIDLEY B.Decomposing global solar radiation into its direct and diffuse components[J]. Renewable and sustainable energy reviews, 2013, 28: 749-756.
[12] TAPAKIS R, MICHAELIDES S, CHARALAMBIDES A G.Computations of diffuse fraction of global irradiance: part 2:neural networks[J]. Solar energy, 2016, 139: 723-732.
[13] SANTOS C M D, ESCOBEDO J F, TERAMOTO É T, et al. Assessment of ANN and SVM models for estimating normal direct irradiation (Hb)[J]. Energy conversion and management, 2016, 126: 826-836.
[14] 李芬, 刘迪, 胡超, 等. 基于PCA-LMBP神经网络的北京地区直散分离预测[J]. 水电能源科学, 2017, 35(4):208-212.
LI F, LIU D, HU C, et al.Prediction of beam-diffuse radiation separation in Beijing based on PCA-LMBP neural network[J]. Water resources and power, 2017, 35(4): 208-212.
[15] BRIGHT J M, ENGERER N A.Engerer2: Global re-parameterisation, update, and validation of an irradiance separation model at different temporal resolutions[J]. Journal of renewable and sustainable energy, 2019, 11(3): 033701.
[16] STARKE A R, LEMOS L F L, BARNI C M, et al. Assessing one-minute diffuse fraction models based on worldwide climate features[J]. Renewable energy, 2021, 177: 700-714.
[17] IQBAL M.An introduction to solar radiation[M]. Vancouver: Academic Press, 1983: 1-28.
[18] GUEYMARD C A.A reevaluation of the solar constant based on a 42-year total solar irradiance time series and a reconciliation of spaceborne observations[J]. Solar energy, 2018, 168: 2-9.
[19] THRELKELD J L, JORDAN R C.Direct solar radiation available on clear days[J]. Heat piping air cond, 1957, 29(12): 135-145.
[20] FORSTINGER A, WILBERT S, JENSEN A R, et al.Expert quality control of solar radiation ground data sets[C]//Proceedings of the ISES Solar World Congress 2021, Virtual Conference: International Solar Energy Society, 2021: 1-12.
[21] GARSON G D.Interpreting neural-network connection weights[J]. AI expert, 1991, 6(4): 46-51.
[22] LAGARIAS J C, REEDS J A, WRIGHT M H, et al.Convergence properties of the nelder-mead simplex method in low dimensions[J]. SIAM journal on optimization, 1998, 9(1): 112-147.
基金
新能源与储能运行控制国家重点实验室（中国电力科学院有限公司）开放基金（NYB51202101990）
PDF(9148 KB)
