PDF(6396 KB)
PDF(6396 KB)
PDF(6396 KB)
基于若干数据分解技术的优化快速学习网地下水水位预测
Optimized Fast Learning Network for Groundwater Level Prediction Based on Several Data Decomposition Techniques
为提高地下水水位时间序列预测精度,同时探讨经验模态分解(EMD)、集合经验模态分解(EEMD)、完整集合经验模态分解(CEEMD)、改进完整集合经验模态分解(ICEEMD)、局部均值分解(LMD)、鲁棒局部均值分解(RLMD)、固有时间尺度分解(ITD)、极点对称模态分解(ESMD)、小波变换(WT)、小波包变换(WPT)、经验小波变换(EWT)、变分模态分解(VMD)、奇异谱分析(SSA)、时变滤波经验模态分解(TVF-EMD)、傅里叶分解(FDM)、辛几何模态分解(SGMD)、逐次变分模态分解(SVMD)17种分解技术在地下水水位时序数据分解中的应用效果,提出基于17种分解技术的爱情进化算法(LEA)-快速学习网(FLN)预测模型。首先,利用EMD等17种分解技术对地下水水位时序数据进行分解处理,得到若干分解分量;其次,基于各分解分量训练集构建适应度函数,利用LEA优化适应度函数获得最佳FLN输入层权值和隐藏层阈值,建立EMD-LEV-FLN等17种模型对各分解分量进行预测和重构;最后,通过云南省草坝地下水监测井2019—2023年逐日水位时间序列预测实例对各模型进行验证。结果表明: ①WPT-LEV-FLN、EWT-LEV-FLN、FDM-LEV-FLN、TVF-EMD-LEV-FLN模型预测精度最高,预测的平均绝对百分比误差MAPE、平均绝对误差MAE、均方根误差RMSE分别在0.000%~0.001%、0.002~0.020 m、0.002~0.032 m之间,决定系数R2均为1.000 0;SSA-LEV-FLN、WT-LEV-FLN、VMD-LEV-FLN模型次之,预测的MAPE、MAE、RMSE、R2分别在0.003%~0.007%、0.041~0.087 m、0.063~0.131 m、0.999 5~0.999 9之间,其他模型预测精度相对较差,预测的MAPE、MAE、RMSE、R2分别在0.017%~0.033%、0.221~0.417 m、0.385~0.705 m、0.985 3~0.995 6之间。②WPT、EWT、FDM、TVF-EMD分解效果最好,其中WPT不但分解效果好,而且分解分量少,最具优势;SSA、WT、VMD分解效果较好,增加分解分量数可进一步提升分解效果;其他分解效果相对较差,其中SGMD、SVMD分解分量最少,最具潜力。③WPT-LEV-FLN模型预测精度高、计算规模小,最具实用价值和实用意义。
[Objective] To improve the accuracy of groundwater level time series prediction and explore the application effects of 17 decomposition techniques—EMD, EEMD, CEEMD, ICEEMD, LMD, RLMD, ITD, ESMD, WT, WPT, EWT, VMD, SSA, TVF-EMD, FDM, SGMD, and SVMD—in the decomposition of groundwater level time series data, a love evolution algorithm (LEA) - fast learning network (FLN) prediction model based on these 17 decomposition techniques is proposed. [Methods] Firstly, 17 decomposition techniques including EMD were used to decompose the groundwater level time series data, and several decomposition components were obtained. Secondly, based on the training set of each decomposition component, a fitness function was constructed, and LEA was used to optimize the fitness function to obtain the optimal FLN input layer weight and hidden layer threshold for FLN. Seventeen models, including EMD-LEV-FLN, were established to predict and reconstruct each decomposition component. Finally, the daily water level time series prediction of the Caoba groundwater monitoring well in Yunnan Province from 2019 to 2023 was used as an example to verify each model. [Results] (1) WPT-LEV-FLN, EWT-LEV-FLN, FDM-LEV-FLN, TVF-EMD-LEV-FLN models achieved the highest prediction accuracy, with average absolute percentage error (MAPE), average absolute error (MAE), and root mean square error (RMSE) ranging 0.000%-0.001%, 0.002-0.020 m, and 0.002-0.032 m, respectively. The determination coefficients (R2) were all 1.000 0. The SSA-LEV-FLN, WT-LEV-FLN, and VMD-LEV-FLN models came in second place, with predicted MAPE, MAE, RMSE, and R2 ranging 0.003%-0.007%, 0.041-0.087 m, 0.063-0.131 m, and 0.999 5-0.999 9, respectively. Other models had relatively poor prediction accuracy, with predicted MAPE, MAE, RMSE, and R2 ranging 0.017%-0.033%, 0.221-0.417 m, 0.385-0.705 m, and 0.985 3-0.995 6, respectively. Among them, WPT-LEV-FLN model had high prediction accuracy and small computational scale, demonstrating the greatest practical value and significance. (2) WPT, EWT, FDM, and TVF-EMD showed the best decomposition performance, among which WPT not only had good decomposition performance, but also produced fewer decomposition components, making it the most advantageous. SSA, WT, and VMD showed relatively good decomposition performance, and increasing the number of decomposition components could further improve the decomposition effectiveness. The other models performed relatively poorly, among which SGMD and SVMD had the least decomposition components and the greatest potential. [Conclusion] This study compares the application performance of 17 current mainstream time series decomposition techniques for processing groundwater level time series decomposition and proposes 17 prediction models, providing reference and guidance for the selection of time series decomposition methods and research on groundwater time series prediction.
地下水水位预测 / 分解技术 / 爱情进化算法 / 快速学习网 / 权阈值优化 / 时间序列
groundwater level prediction / decomposition techniques / love evolution algorithm / fast learning network / weight threshold optimization / time series
| [1] |
席东洁, 赵雪花, 张永波, 等. 基于经验模态分解与Elman神经网络的月径流预测[J]. 中国农村水利水电, 2017(7):112-115.
针对径流序列的弱相关，非线性特征，直接预测会导致精度低，建立基于经验模态分解（empirical mode decomposition，EMD）与Elman神经网络的模型进行月径流预测，选取黄河上游唐乃亥水文站1979-2009年的月径流资料为研究对象。首先利用EMD对月径流资料进行处理，得到5个固有模态函数（intrinsic mode function，IMF）和1个趋势项。为了避免分量太多导致误差过大，将分量进行重组之后得到三个子序列，然后对三个子序列进行Elman神经网络预测。结果表明，EMD与Elman神经网络组合预测精度更高，更加适合复杂的水文序列。该方法可用于径流的中长期预测中。
(
In view of the weak correlation, nonlinear characteristics of runoff series, the EMD and Elman neural network are adopted to forecast monthly runoff, and monthly runoff data from 1979 to 2009 at Tangnaihai hydrological station in the upper reaches of the Yellow River is selected. Firstly, EMD is used to decompose the data, and then five intrinsic mode function (IMF) components and one trend are obtained. In order to avoid errors from multi-components, the original component should be restructured to three sub-series, which are predicted using Elman neural network. Results show that the precision of EMD-Elman forecast model was higher，and more suitable for complicated hydrological sequence. This method can be used for long-term runoff forecasting.
|
| [2] |
王佳, 王旭, 王浩, 等. 基于EEMD与ANN混合方法的水库月径流预测[J]. 人民黄河, 2019, 41(5):43-46.
(
|
| [3] |
王文川, 杜玉瑾, 和吉, 等. 基于CEEMDAN-VMD-BP模型的月径流量预测研究[J]. 华北水利水电大学学报(自然科学版), 2023, 44(1): 32-40, 48.
(
|
| [4] |
毛钦男, 刘招, 李杰, 等. 联合ICEEMDAN与多种智能优化算法的径流预测及对比研究[J]. 水电能源科学, 2023, 41(10): 23-26, 31.
(
|
| [5] |
井露茜, 文传博. 基于LMD-SECNN的风机变桨系统故障检测[J]. 上海电机学院学报, 2024, 27(1): 13-19.
(
|
| [6] |
陈志刚, 赵志川, 钟新荣, 等. 基于鲁棒局部均值分解与二阶瞬态提取变换的滚动轴承故障诊断[J]. 科学技术与工程, 2022, 22(1): 157-165.
(
|
| [7] |
关焦月, 田晶, 赵金明, 等. 基于固有时间尺度分解与多尺度形态滤波的滚动轴承故障特征提取方法[J]. 科学技术与工程, 2019, 19(14):178-182.
(
|
| [8] |
刘尚东, 孙东永, 许晶晶, 等. 基于ESMD-NNBR耦合的月径流预测模型[J]. 水文, 2023, 43(1): 52-56.
(
|
| [9] |
徐成贵, 崔东文. 基于WT-HBBO-RBF模型的年径流时间序列预测[J]. 水力发电, 2023, 49(4): 17-22, 95.
(
|
| [10] |
李菊, 崔东文. 基于数据分解与斑马算法优化的混合核极限学习机月径流预测[J]. 长江科学院院报, 2024, 41(6): 42-50.
为提高月径流预测精度,改进混合核极限学习机(HKELM)预测性能,提出小波包分解(WPT)－斑马优化算法(ZOA)－HKELM组合模型。利用WPT处理月径流时序数据,构建局部高斯径向基核函数和全局多项式核函数相混合的HKELM;通过ZOA优化HKELM超参数(正则化参数、核参数、权重系数),建立WPT-ZOA-HKELM组合模型,并构建WPT-遗传算法(GA)-HKELM、WPT-灰狼优化(GWO)算法-HKELM、WPT-鲸鱼优化算法(WOA)-HKELM、WPT-ZOA-极限学习机(ELM)、WPT-ZOA-最小二乘支持向量机(LSSVM)、ZOA-HKELM作对比模型,通过黑河流域莺落峡、讨赖河水文站月径流时间序列预测实例对各模型进行检验。结果表明:①莺落峡、讨赖河水文站月径流时间序列WPT-ZOA-HKELM模型预测的平均绝对百分比误差分别为1.054%、0.761%,决定系数均达0.999 9,优于其他对比模型,具有更高的预测精度,预测效果更好。②利用ZOA优化HKELM超参数,可提高HKELM预测性能,优化效果优于GWO、WOA、GA。③预测模型能充分发挥WPT、ZOA和HKELM优势,提高月径流预测精度;在相同分解和优化情形下,HKELM的预测性能优于ELM、LSSVM。
(
|
| [11] |
莫崇勋, 邓云, 阮俞理, 等. 基于EWT-PSO-Elman耦合模型在径流预测中的应用[J]. 科学技术与工程, 2022, 22(22): 9775-9780.
(
|
| [12] |
张上要, 罗军刚, 石国栋, 等. 基于VMD-TCN模型的渭河流域月径流量预测研究[J]. 人民黄河, 2023, 45(10):25-29.
(
|
| [13] |
杨琼波, 崔东文. WPD-RSO-ESN和SSA-RSO-ESN模型在径流时间序列预测中应用比较[J]. 中国农村水利水电, 2022(2):61-67,75.
为提高径流时间序列预测精度，提出小波包分解（WPD）与奇异谱分解（SSA）-鼠群优化（RSO）算法-回声状态网络（ESN）相混合的径流时间序列预测方法。分别利用WPD和SSA将非平稳径流时间序列分解为若干子序列，有效降低径流时间序列的复杂性；介绍RSO算法原理，在不同维度条件下选取6个典型函数对RSO算法进行仿真测试；利用RSO算法对ESN储备池规模、稀疏度等超参数进行优化，建立WPD-RSO-ESN、SSA-RSO-ESN模型，并分别构建WPD-RSO-SVM、WPD-ESN、WPD-SVM和SSA-RSO-SVM、SSA-ESN、SSA-SVM作对比分析模型；利用云南省江边街水文站1957-2014年逐月径流时间序列数据对8种模型进行检验及对比分析。结果表明：RSO算法在不同维度条件下均具有较好的寻优精度和全局搜索能力。WPD-RSO-ESN、SSA-RSO-ESN模型对实例后10年120个月月径流时间序列预测的平均绝对百分比误差分别为2.73%、3.90%，预测精度优于同一分解条件下的其他模型。RSO算法能有效优化ESN网络超参数，提高ESN网络的预测性能。WPD对径流时间序列数据的分解效果优于SSA方法。
(
To improve the prediction accuracy of runoff time series this paper proposes a runoff time series prediction method that combines wavelet packet decomposition （WPD） with singular spectrum decomposition （SSA）-rat swarm optimization （RSO） algorithm-echo state network （ESN）. WPD and SSA is used to decompose the non-stationary runoff time series into several sub-sequences to effectively reduce the complexity of the runoff time series； the principle of the RSO algorithm is introduced， 6 typical functions are selected under different dimensional conditions to simulate the RSO algorithm； RSO algorithm is used to optimize hyperparameters such as ESN reserve pool size and sparsity， WPD-RSO-ESN， SSA-RSO-ESN models are established， and WPD-RSO-SVM， WPD-ESN， WPD-SVM and SSA-RSO- SVM， SSA-ESN are constructed， and SSA-SVM are used as comparative analysis models； the monthly runoff time series data from 1957 to 2014 at Jiangbian Street Hydrological Station in Yunnan Province are used to test and compare 8 models. The results show that the RSO algorithm has better optimization accuracy and global search ability under different dimensional conditions. The WPD-RSO-ESN and SSA-RSO-ESN models have predicted average absolute percentage errors of monthly runoff time series for 10 years and 120 months after the example. The average absolute percentage errors are 2.73% and 3.90%， respectively. The prediction accuracy is better than other models under the same decomposition conditions. The RSO algorithm can effectively optimize the hyperparameters of the ESN network and improve the prediction performance of the ESN network. The decomposition effect of WPD on runoff time series data is better than SSA method. |
| [14] |
王秀杰, 张帅, 田福昌, 等. 基于TVF-EMD-PSO-GRU的月径流模型与应用研究[J]. 天津大学学报(自然科学与工程技术版), 2022, 55(8):802-810.
(
|
| [15] |
张辉, 张超, 辛阔, 等. 基于FDM和TEO的滚动轴承故障诊断研究[J]. 机电工程, 2021, 38(7): 850-856.
(
|
| [16] |
李显泽, 龙海洋, 郑直, 等. 基于SGMD线性峭度和log-SAM的滚动轴承故障诊断方法[J]. 噪声与振动控制, 2020, 40(6):121-127.
(
|
| [17] |
任宏宇, 余瑶怡, 杜雄, 等. 基于优化长短期记忆神经网络的IGBT寿命预测模型[J]. 电工技术学报, 2024, 39(4):1074-1086.
(
|
| [18] |
侯金霄, 黄林显, 胡晓农, 等. 基于EMD-LSTM耦合模型的趵突泉岩溶地下水水位预测应用[J]. 水资源与水工程学报, 2023, 34(4): 92-98.
(
|
| [19] |
张建锋, 刘见宝, 崔树军, 等. 小波-神经网络混合模型预测地下水水位[J]. 长江科学院院报, 2016, 33(8):18-21.
由于过量开采地下水，华北平原的许多城市出现地下水水位持续下降趋势，由此导致了许多严重的环境问题，如地下水枯竭、地面沉降和海水入侵等。为了准确预测城市地下水水位变化，利用小波变换的多尺度分析特征，建立了小波-神经网络混合模型(以下简称“混合模型”)，并研究了其在地下水水位预测中的精度。利用北京市平谷区地下水水位观测资料，分别用BP网络和混合模型对该区地下水水位进行了预测。采用均方根误差(RMSE)、平均绝对误差(MAE)和线性相关系数(R)对模型预测的精度进行度量。预测结果表明混合模型第1至第3个月的地下水水位平均绝对误差分别是0.535，0.598和0.634 m;而BP模型的平均绝对误差分别为0.566，0.824和0.940 m。混合模型的预测误差分别为BP模型的95%，73%和67%。使用混合模型能明显提高预测的精度，显著增加有效预测时段长度。
(
Due to over-exploitation of groundwater in many cities of North China Plain, there is a tendency of lasting decrease in groundwater level, which results in serious problems, such as groundwater exhaustion, land subsidence and seawater intrusion. In order to accurately predict changes of urban groundwater level, based on artificial neural network (ANN) and analysis of multi-scale of wavelet transform (WT), we established a wavelet-ANN conjugate model and test its accuracy to predict groundwater level. Measured data of groundwater level at Pinggu district of Beijing were taken as research objects. We predicted groundwater levels at the district by back propagation (BP) model and hybrid model. Then, we calculated the prediction accuracy by using statistical parameters including root mean square error (RMSE), mean absolute error (MAE) and correlation coefficient (R). Results showed that the MAE of the hybrid model from the first month to the third month was 0.535, 0.598 and 0.634 m, respectively, whereas 0.566, 0.824 and 0.940 m for BP model. The MAE of hybrid model from the first month to the third month was 95%, 73% and 67% of that of BP model, respectively. Comparison of results reveals that the hybrid model has advantages of better prediction accuracy and longer effective prediction duration.
|
| [20] |
付智勇, 陈文强, 唐伟雄, 等. 基于CEEMD-RF模型的渣土边坡地下水埋深预测[J]. 人民长江, 2020, 51(1):141-148.
(
|
| [21] |
|
| [22] |
胡坚, 刘超. CPSO-FLN模型在汽轮机热耗率预测中的应用研究[J]. 中国电力, 2019, 52(8): 179-184.
(
|
| [23] |
汪婵婵. 基于改进狮群算法的汽轮机热耗率模型预测[J]. 计量学报, 2021, 42(7):853-860.
(
|
| [24] |
李霞, 牛培峰, 马云鹏, 等. 基于改良的鸡群优化算法优化锅炉NOx排放质量浓度[J]. 动力工程学报, 2017, 37(4): 293-300.
为达到锅炉NO<sub>x</sub>排放质量浓度最优的目标，针对鸡群优化算法收敛速度慢、容易早熟等缺点，对算法中母鸡和小鸡的觅食行为分别进行了修改，得到改良的鸡群优化 （A-CSO） 算法，并通过测试函数验证了A-CSO算法性能优于粒子群算法、引力搜索算法、磷虾群算法和原鸡群优化算法.利用A-CSO算法以及快速学习网建立锅炉NO<sub>x</sub>排放质量浓度预测模型，对锅炉运行时的可调参数进行优化，获得锅炉燃烧优化调整方式.结果表明：优化后所有工况的NO<sub>x</sub>排放质量浓度明显下降，相对下降率优于文献[12]中的结果；由于锅炉燃烧中飞灰含碳量的影响，可适当调整优化后的氧量和一次风量，以达到锅炉高效低污染燃烧.
(
To reduce the NO<sub><i>x</i></sub> emission of a boiler, an ameliorated chicken swarm optimization (A-CSO) algorithm was proposed by respectively modifying the foraging behaviors of hens and chicks in the original chicken swarm optimization algorithm to overcome the shortcomings of lower convergent speed and premature convergence, of which the performance was proved to be better than the particle swarm optimization algorithm, gravitational search algorithm, krill swarm algorithm and the original chicken swarm optimization algorithm through test function verification. Meanwhile, a prediction model was established for boiler NO<sub><i>x</i></sub> emission on the basis of A-CSO algorithm and fast-learning network, so as to optimize the adjustable parameters in boiler operation and obtain the way of boiler combustion optimization. Results indicate that the NO<sub><i>x</i></sub> emission in all cases is significantly lowered after optimization, and the relative decline rates are superior to that in the literature. Considering the influence of unburned carbon in the fly ash from boiler combustion, it is proposed to appropriately adjust the optimized oxygen content and primary air flow to achieve high efficiency and low NO<sub><i>x</i></sub> emission of the boiler.
|
| [25] |
牛培峰, 王丘亚, 马云鹏, 等. 基于量子自适应鸟群算法的锅炉NOx排放特性研究[J]. 计量学报, 2017, 38(6):770-775.
(
|
| [26] |
牛培峰, 吴志良, 马云鹏, 等. 基于鲸鱼优化算法的汽轮机热耗率模型预测[J]. 化工学报, 2017, 68(3):1049-1057.
为了准确地建立汽轮机热耗率预测模型，提出了一种基于反向学习自适应的鲸鱼优化算法（AWOA）和快速学习网（FLN）综合建模的方法。首先将改进后的鲸鱼算法与经典改进的粒子群、差分进化算法和基本鲸鱼算法进行比较，结果证明其具有更高的收敛精度和更快的收敛速度；然后采用某热电厂600 MW超临界汽轮机组现场收集的运行数据建立汽轮机热耗率预测模型，并将改进后的鲸鱼算法优化的快速学习网模型的预测结果与基本快速学习网及经典改进的粒子群、差分进化算法和基本鲸鱼算法优化的快速学习网模型预测结果相比较。结果表明，AWOA-FLN预测模型具有更高的预测精度和更强的泛化能力，更能准确地预测汽轮机的热耗率。
(
|
| [27] |
牛培峰, 陈科, 马云鹏, 等. 基于磷虾群算法的汽轮机热耗率建模应用[J]. 动力工程学报, 2016, 36(10):781-787.
为了准确建立汽轮机热耗率预测模型，以某热电厂600 MW超临界汽轮机组为研究对象，采用基于反向学习自适应的磷虾群算法（OAKH）和快速学习网（FLN）进行综合建模，并将该模型的预测结果与基本快速学习网、粒子群算法、生物地理学优化算法和磷虾群算法优化的快速学习网模型的预测结果进行比较.结果表明：OAKH算法能够更好地优化FLN模型参数，使所建立的FLN汽轮机热耗率预测模型具有更高的预测精度和更强的泛化能力，能够准确、有效地预测热电厂的汽轮机热耗率.
(
To accurately predict the heat rate of steam turbine, a model was established with the sample data of a 600 MW supercritical steam turbine unit in a thermal power plant using opposition adaptive krill herd algorithm (OAKH) and fast learning network (FLN), of which the prediction results were compared with that of basic FLN model and those FLN models whose parameters were optimized by particle swarm optimization, biogeography-based optimization and krill herd algorithm. Results show that compared with other algorithms and models, the model of turbine heat rate based on OAKH algorithm has a higher accuracy in prediction and stronger capability in parameter optimization and generation, which may help to accurately and effectively predict the heat rate of steam turbines.
|
| [28] |
周勇. 基于MOSFLA与快速学习网的荷电状态预测[J]. 计算机应用与软件, 2020, 37(4): 329-333.
(
|
| [29] |
李霞, 刘建平, 纪佳琪. 基于樽海鞘群算法与快速学习网的旅游客流量预测研究[J]. 赤峰学院学报(自然科学版), 2021, 37(6):33-37.
(
|
/
| 〈 |
|
〉 |
