Application of deep learning algorithms to confluent flow-rate forecast with multivariate decomposed variables

Tebong, Njogho Kenneth; Simo, Théophile; Takougang, Armand Nzeukou; Sandjon, Alain Tchakoutio; Herve, Ntanguen Patrick

doi:10.1016/j.ejrh.2023.101357

Cited by 3 publications

(4 citation statements)

References 32 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Loess is an algorithm for handling nonlinear correlations. STL is highly desirable for time series decomposition due to its ability to handle any type of seasonality, its ability to be controlled by the user in trend-cycle smoothening, its ability to be robust to outliers, and its ability to allow the seasonal component to change over time [ 25 , 26 , 28 ]. From Figs.…”

Section: Methodsmentioning

confidence: 99%

“…It is located in Bertoua in the East region of Cameroon. Due to the stochastic nature of the hydroelectricity production in the Southern interconnected grid of Cameroon, the reservoir was constructed to regulate the flow rate of water for hydroelectricity production [ 26 ]. The reservoir has the following coordinates 5 0 22′16.87″ N, 13 0 30′44.67″ E. Its minimum and maximum storage capacity are 100.000.000 m 3 and 6280 000 000 m 3 respectively.…”

Section: Study Areamentioning

confidence: 99%

“…STL is used to decompose the reservoir inflow and precipitation into random, seasonal, and trend components. STL was chosen over other decomposition methods such as classical decomposition, and SEATS (seasonal extraction in autoregressive integrated moving average (ARIMA) time series) due to its ability to handle different types of seasonality and user control of smoothness in trend-cycle interaction among others [ 25 , 26 ]. Deep learning models namely the dense model, LSTM model, and Conv1D have been implemented to forecast the STL decomposed components and the resultant reservoir inflow is the sum of the forecasted decomposed inflows of the reservoir.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

STL-decomposition ensemble deep learning models for daily reservoir inflow forecast for hydroelectricity production

Tebong

Simo

Takougang

et al. 2023

Heliyon

View full text Add to dashboard Cite

Section: Methodsmentioning

confidence: 99%

Section: Study Areamentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

STL-decomposition ensemble deep learning models for daily reservoir inflow forecast for hydroelectricity production

Tebong

Simo

Takougang

et al. 2023

Heliyon

View full text Add to dashboard Cite

“…As a versatile method, STL has an advantage over other decomposition methods on account of its ability to deal with any type of seasonality and to be robust to outliers. [28,29] It is based on locally weighted regression (LOESS), a sequence of smoothing procedures. [26] STL consists of two recursive processes: an inner loop nested an outer loop.…”

Section: Seasonal Trend Lossesmentioning

confidence: 99%

Dynamic changes of net anthropogenic nitrogen/phosphorus input and its riverine export in the Beijing's subcenter

Luo,

Peng,

Shang

et al. 2024

CLEAN Soil Air Water

View full text Add to dashboard Cite

Due to human activities, the overabundance of regional nitrogen/phosphorus (N/P) has a huge impact on ecological environment worldwide. It is necessary to comprehensively understand the human‐induced N/P input and spatiotemporal variations of nutrients in a river. As the adjustment to the spatial structure of Beijing, the Beijing's subcenter is an important strategic position in promoting the coordinated development of Beijing–Tianjin–Hebei. In this study, the dynamic changes of ammonia nitrogen (NH3‐N) and total phosphorus (TP) from 2012 to 2022 were analyzed, then the net anthropogenic N input (NANI) and net anthropogenic P input (NAPI) was estimated in the Beijing's subcenter. The results showed that both NH3‐N and TP had an overall downward trend with a three‐stage pattern. The NANI and NAPI declined from 5.68 × 104 and 0.95 × 104 to 0.78 × 104 and 0.15 × 104 kg km−2 per year, respectively. The chemical fertilizers input accounted for the largest part at 53–81% of NANI and 49–80% of NAPI. There were high values of N and P input in the southeast of Tongzhou, and the fluxes of N and P were positively correlated with the input of anthropogenic N and P in the Chaobai River Basin (Tongzhou District) and the Beiyun River Basin (Tongzhou District). This study would provide the foundation for setting priorities and enacting more targeted N/P management strategies in the Beijing's subcenter.

show abstract

RNN-Based Monthly Inflow Prediction for Dez Dam in Iran Considering the Effect of Wavelet Pre-Processing and Uncertainty Analysis

Adib,

Pourghasemzadeh,

Lotfirad

2024

Hydrology

View full text Add to dashboard Cite

In recent years, deep learning (DL) methods, such as recurrent neural networks (RNN). have been used for streamflow prediction. In this study, the monthly inflow into the Dez Dam reservoir from 1955 to 2018 in southwestern Iran was simulated using various types of RNNs, including long short-term memory (LSTM), bidirectional long short-term memory (Bi-LSTM), gated recurrent unit (GRU), and stacked long short-term memory (Stacked LSTM). It was observed that considering flow discharge, temperature, and precipitation as inputs to the models yields the best results. Additionally, wavelet transform was employed to enhance the accuracy of the RNNs. Among the RNNs, the GRU model exhibited the best performance in simulating monthly streamflow without using wavelet transform, with RMSE, MAE, NSE, and R2 values of 0.061 m3/s, 0.038 m3/s, 0.556, and 0.642, respectively. Moreover, in the case of using wavelet transform, the Bi-LSTM model with db5 mother wavelet and decomposition level 5 was able to simulate the monthly streamflow with high accuracy, yielding RMSE, MAE, NSE, and R2 values of 0.014 m3/s, 0.008 m3/s, 0.9983, and 0.9981, respectively. Uncertainty analysis was conducted for the two mentioned superior models. To quantify the uncertainty, the concept of the 95 percent prediction uncertainty (95PPU) and the p-factor and r-factor criteria were utilized. For the GRU, the p-factor and r-factor values were 82% and 1.28, respectively. For the Bi-LSTM model, the p-factor and r-factor values were 94% and 1.06, respectively. The obtained p-factor and r-factor values for both models are within the acceptable and reliable range.

show abstract

Application of deep learning algorithms to confluent flow-rate forecast with multivariate decomposed variables

Cited by 3 publications

References 32 publications

STL-decomposition ensemble deep learning models for daily reservoir inflow forecast for hydroelectricity production

STL-decomposition ensemble deep learning models for daily reservoir inflow forecast for hydroelectricity production

Dynamic changes of net anthropogenic nitrogen/phosphorus input and its riverine export in the Beijing's subcenter

RNN-Based Monthly Inflow Prediction for Dez Dam in Iran Considering the Effect of Wavelet Pre-Processing and Uncertainty Analysis

Contact Info

Product

Resources

About