
基于LMD-QPSO-CRJ模型汾河上游月径流预测方法研究
邢馨予, 赵雪花
基于LMD-QPSO-CRJ模型汾河上游月径流预测方法研究
Monthly Runoff Prediction on the Upper Fenhe River Based on LMD-QPSO-CRJ Model
针对径流序列具有较强的非平稳性和非线性特征,为提高预测精度,提出一种月径流组合预测模型LMD-QPSO-CRJ。选用局部均值分解法(Local Mean Decomposition,LMD)对径流数据分解降噪,并对第一个高频分量用变模态分解法(Variational Mode Decomposition,VMD)进行二次分解,采用量子粒子群优化算法(QPSO)对确定性循环跳跃网络(Cycle Reservoir with Regular Jumps,CRJ)进行参数优化,最终建立LMD-QPSO-CRJ模型。将该模型应用于汾河上游汾河水库站和上静游站的月径流预测,并与单一QPSO-CRJ模型及CEEMD-QPSO-CRJ模型进行对比分析。结果表明,验证期LMD-QPSO-CRJ模型的MAE值和RMSE值与单一QPSO-CRJ模型相比分别减少32%~40%和23%~31%,与CEEMD-QPSO-CRJ模型相比分别减少11%~26%和11%~18%,LMD-QPSO-CRJ模型的NSE值最接近于1。因此,LMD-QPSO-CRJ模型具有较好的预测精度,可以用于指导实际的生产建设。
In order to improve the accuracy of runoff prediction, a runoff forecasting hybrid model named LMD-QPSO-CRJ is developed for monthly forecasting featured with strong non-stationarity and non-linearity. Firstly, the runoff data is decomposed with Local Mean Decomposition to reduce noise, and Variational Mode Decomposition (VMD) is employed to further decompose the first high-frequency component. Secondly, Quantum Particle Swarm Optimization (QPSO) is utilized to optimize the parameters of the Cycle Reservoir with Regular Jumps (CRJ). Finally, the LMD-QPSO-CRJ model is established. The model is applied to the monthly runoff prediction of Fenhe Reservoir Station and Shangjingyou Station in the upper reaches of Fenhe River. LMD-QPSO-CRJ model is compared with other models such as the single QPSO-CRJ model and CEEMD-QPSO-CRJ model for a comparative analysis. The results show that the MAE value and RMSE value of the LMD-QPSO-CRJ model has reduced by 32%~40% and 23%~31% compared with the single QPSO-CRJ model, by 11%~26% and 11%~18% compared with CEEMD-QPSO-CRJ model during the validation period. The NSE value of the LMD-QPSO-CRJ model is close to 1. The results highlight the LMD-QPSO-CRJ model with great prediction accuracy, the model can be used to guide practical production and construction.
汾河上游 / LMD-QPSO-CRJ模型 / CEEMD-QPSO-CRJ模型 / 月径流预测 {{custom_keyword}} /
upper reaches of the Fenhe River / LMD-QPSO-CRJ model / CEEMD-QPSO-CRJ model / monthly runoff prediction {{custom_keyword}} /
表1 2个水文站各模型预测误差对比Tab.1 Comparison of prediction errors of each model of two hydrologic stations |
站名 | 模型 | 训练期 | 验证期 | ||||
---|---|---|---|---|---|---|---|
MAE/万m3 | RMSE/万m3 | NSE | MAE/万m3 | RMSE/万m3 | NSE | ||
汾河水库站 | QPSO-CRJ | 702.85 | 1 684.85 | 0.79 | 425.69 | 532.61 | 0.83 |
CEEMD-QPSO-CRJ | 539.64 | 1 473.32 | 0.84 | 388.16 | 492.02 | 0.85 | |
LMD-QPS0-CRJ | 418.70 | 1 158.32 | 0.90 | 287.40 | 406.18 | 0.90 | |
上静游站 | QPSO-CRJ | 97.68 | 300.35 | 0.78 | 59.46 | 67.84 | 0.70 |
CEEMD-QPSO-CRJ | 82.59 | 238.09 | 0.86 | 43.14 | 53.08 | 0.81 | |
LMD-QPSO-CRJ | 63.77 | 201.67 | 0.90 | 35.62 | 47.01 | 0.85 |
1 |
吕晗芳,赵雪花,桑宇婷,等. 基于VMD-LSSVM的月径流预测方法研究[J]. 中国农村水利水电,2020(8):166-170,176.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
2 |
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.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
3 |
{{custom_citation.content}}
{{custom_citation.annotation}}
|
4 |
程军圣,张 亢,杨 宇. 局部均值分解方法在调制信号处理中的应用[J]. 振动、测试与诊断,2010,30(4):362-366,470.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
5 |
李天成,钟平安,吴业楠,等. 基于产流机制的中长期径流预测RBF神经网络模型研究[J]. 水力发电,2017,43(3):13-17.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
6 |
李代华. 基于改进布谷鸟搜索算法优化的SVM月径流预测[J]. 中国农村水利水电,2020(8):171-176.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
7 |
张 梦,赵靓芳,全 星. PSO-ESN在城市内涝点降雨积水预测中的应用[J]. 中国农村水利水电,2019(6):56-59,65.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
8 |
{{custom_citation.content}}
{{custom_citation.annotation}}
|
9 |
{{custom_citation.content}}
{{custom_citation.annotation}}
|
10 |
{{custom_citation.content}}
{{custom_citation.annotation}}
|
11 |
秋兴国,王瑞知,张卫国,等. 基于PCA-CRHJ模型的矿井突水水源判别[J]. 工矿自动化,2020,46(11):65-71.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
12 |
韩 锐,董增川,王雪薇,等. 加权平均集成模型在径流预测中的应用[J]. 人民黄河,2017,39(6):16-20,23.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
13 |
吴东升,杨 青,张继云,等. 基于VMD-LMD-CNN的集合型故障诊断方法[J]. 轴承,2020(10):57-63.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
14 |
韩 敏,姜 涛,冯守渤. 基于VMD循环随机跳跃状态网络的时间序列长期预测[J]. 控制与决策,2020,35(9):2 175-2 181.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
15 |
{{custom_citation.content}}
{{custom_citation.annotation}}
|
16 |
{{custom_citation.content}}
{{custom_citation.annotation}}
|
17 |
孟飞,兰巨龙,胡宇翔. 基于改进的量子粒子群优化小波神经网络的网络流量预测[J]. 计算机应用研究,2015,32(5):1 450-1 453.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
{{custom_ref.label}} |
{{custom_citation.content}}
{{custom_citation.annotation}}
|
/
〈 |
|
〉 |