The precise predicting of air temperature has a significant influence in many sectors such as agriculture, industry, modeling environmental processes. In this work, to predict the mean daily time series air temperatur...
详细信息
The precise predicting of air temperature has a significant influence in many sectors such as agriculture, industry, modeling environmental processes. In this work, to predict the mean daily time series air temperature in Mu & gbreve;la city (AT(m)), Turkey, initially, five different layer structures of Long Short-Term Memory (LSTM) and Gated Recurrent Unit (GRU) deep learning-based neural network models through the seq2seq regression forecast module are developed. Then, based on performance evaluation metrics, an optimal DL-based layer network structure designed is chosen to hybridize with the wavelettransform (WT) algorithm (i.e., WT-DNN model) to enhance the estimation capability. In this direction, among potential meteorological variables considered, the average daily sunshine duration (SSD) (hours), total global solar radiation (TGSR) (kw. hour/m(2)), and total global insolation intensity (TGSI) (watt/m(2)) from Jan 2014 to Dec 2019 are picked as the most effective input variables through correlation analysis to predict AT(m). To thwart overfitting and underfitting problems, different algorithm tuning along with trial-and-error procedures through diverse types of hyper-parameters are performed. Consistent with the performance evaluation standards, comparison plots, and Total Learnable Parameters (TLP) value, the state-of-the-art and unique proposed hybrid WT-(LSTM x GRU) model (i.e., hybrid WT with the coupled version of LSTM and GRU models via Multiplication layer (x)) is confirmed as the best model developed. This hybrid model under the ideal hyper-parameters resulted in an R-2 = 0.94, an RMSE = 0.56 (degrees C), an MBE = -0.5 (degrees C), AICc = -382.01, and a running time of 376 (s) in 2000 iterations. Nonetheless, the standard single LSTM layer network model as benchmark model resulted in an R-2 = 0.63, an RMSE = 4.69 (degrees C), an MBE = -0.89 (degrees C), AICc = 1021.8, and a running time of 186 (s) in 2000 iterations.
In this paper, a new hybrid approach by combining the Support Vector Machine (SVM) with wavelettransform (WT) algorithm is developed to predict horizontal global solar radiation. The predictions are conducted on both...
详细信息
In this paper, a new hybrid approach by combining the Support Vector Machine (SVM) with wavelettransform (WT) algorithm is developed to predict horizontal global solar radiation. The predictions are conducted on both daily and monthly mean scales for an Iranian coastal city. The proposed SVM-WT method is compared against other existing techniques to demonstrate its efficiency and viability. Three different sets of parameters are served as inputs to establish three models. The results indicate that the model using relative sunshine duration, difference between air temperatures, relative humidity, average temperature and extraterrestrial solar radiation as inputs shows higher performance than other models. The statistical analysis demonstrates that SVM-WT approach enjoys very good performance and outperforms other approaches. For the best SVM-WT model, the obtained statistical indicators of mean absolute percentage error, mean absolute bias error, root mean square error, relative root mean square error and coefficient of determination for daily estimation are 6.9996%, 0.8405 MJ/m(2), 1.4245 MJ/m(2), 7.9467% and 0.9086, respectively. Also, for monthly mean estimation the values are 3.2601%, 0.5104 MJ/m(2), 0.6618 MJ/m(2), 3.6935% and 0.9742, respectively. Based upon relative percentage error, for the best SVM-WT model, 88.70% of daily predictions fall within the acceptable range of -10% to +10%. (C) 2014 Elsevier Ltd. All rights reserved.
Diffuse solar radiation is a fundamental parameter highly required in several solar energy applications. Despite its significance, diffuse solar radiation is not measured in many locations around the world due to tech...
详细信息
Diffuse solar radiation is a fundamental parameter highly required in several solar energy applications. Despite its significance, diffuse solar radiation is not measured in many locations around the world due to technical and fiscal limitations. On this account, determining the amount of diffuse radiation alternatively based upon precise and reliable estimating methods is indeed essential. In this paper, a coupled model is developed for estimating daily horizontal diffuse solar radiation by integrating the support vector machine (SVM) with wavelettransform (WT) algorithm. To test the validity of the coupled SVM-WT method, daily measured global and diffuse solar radiation data sets for city of Kerman situated in a sunny part of Iran are utilized. For the developed SVM-WT model, diffuse fraction (cloudiness index) is correlated with clearness index as the only input parameter. The suitability of SVM-WT is evaluated against radial basis function SVM (SVM-RBF), artificial neural network (ANN) and a 3rd degree empirical model established for this study. It is found that the estimated diffuse solar radiation values by the SVM-WT model are in favourable agreements with measured data. According to the conducted statistical analysis, the obtained mean absolute bias error, root mean square error and correlation coefficient are 0.5757 MJ/m2, 0.6940 MJ/m(2) and 0.9631, respectively. While for the SVM-RBF ranked next the attained values are 1.0877 MJ/m(2),12583 MJ/ m(2) and 0.8599, respectively. In fact, the study results indicate that SVM-WT is an efficient method which enjoys much higher precision than other models, especially the 3rd degree empirical model. (C) 2015 Elsevier Ltd. All rights reserved.
Technical design of sewer systems requires highly accurate prediction of sediment transport. In this study, the capability of the combined support vector machine-wavelettransform (SVM-wavelet) model for the predictio...
详细信息
Technical design of sewer systems requires highly accurate prediction of sediment transport. In this study, the capability of the combined support vector machine-wavelettransform (SVM-wavelet) model for the prediction of the densimetric Froude number (Fr) was compared to the single SVM and different existing sediment transport equations at the limit of deposition. The performance evaluation was performed using the R-square (R-2), three relative indexes (MRE, MARE, MSRE) and three absolute indexes (ME, MAE, RMSE). The factors affecting the Fr were initially determined. After categorizing them into different dimensionless groups, six different models were found to predict the Fr. Comparisons between the obtained results showed that both the SVM and SVM-wavelet can predict the Fr with high accuracy. However, it was found that the SVM-wavelet (R-2=0.995, MRE=0.002, MARE=0.021, MSRE=0.001, ME=0.007, MAE=0.086 and RMSE=0.114) offers higher performance than the SVM and the existing equations. (C) 2015 Elsevier Ltd. All rights reserved.
Fuel cell/ultracapacitor/battery hybrid power systems represent a promising architecture to satisfy the energy requirements for road vehicles. When the primary power source is a fuel cell, the objective of an energy m...
详细信息
Fuel cell/ultracapacitor/battery hybrid power systems represent a promising architecture to satisfy the energy requirements for road vehicles. When the primary power source is a fuel cell, the objective of an energy management strategy (EMS) is to optimise the hydrogen consumption and fuel cell efficiency without compromising the performance of the overall system. In this study, a fuzzy logic controller together with a wavelet-transformalgorithm (WTA) is proposed as EMS, which is a reasonable option to achieve an adequate solution of the problem since this supervisory control strategy can deal with a considerable amount of variables. Besides, a WTA is capable of separating the high and low frequency components of the load profile and help allocate them among the on-board energy sources. A simulation environment has been developed to test the EMS;it includes models for the power sources and energy storage devices, the power electronics and the vehicular power demand.
Aiming at the low illumination image, we proposed a method to make images clearer which is based on our forward fog-removed method[7]. This method combines Multi-Scale Retinex(MSR) algorithm with wavelettransform alg...
详细信息
ISBN:
(纸本)9781467395878
Aiming at the low illumination image, we proposed a method to make images clearer which is based on our forward fog-removed method[7]. This method combines Multi-Scale Retinex(MSR) algorithm with wavelet transform algorithm. The improved method firstly use MSR algorithm to enhance the whole image, then the wavelet transform algorithm is used to strengthen the details information of image, finally a clearly image which has better overall contrast and more clear details information can be obtained. Through analyzing subjective and objective evaluation of the experimental results, the performance indicators of image which are processed by the proposed method in this paper are higher than those processed by other algorithms.
暂无评论