Energy load forecasting: Bayesian and exponential smoothing hybrid methodology

2019 ◽  
Vol ahead-of-print (ahead-of-print) ◽  
Author(s):  
Eman Khorsheed
Keyword(s):  
Hybrid Approach ◽  
Peak Load ◽  
Exponential Smoothing ◽  
Coefficient Of Determination ◽  
Percentage Error ◽  
Original Research ◽  
Model Parameters ◽  
Learning Tools ◽  
Content Type ◽  
Hybrid Methodology

Purpose The purpose of this study is to present a hybrid approach to model and predict long-term energy peak load using Bayesian and Holt–Winters (HW) exponential smoothing techniques. Design/methodology/approach Bayesian inference is administered by Markov chain Monte Carlo (MCMC) sampling techniques. Machine learning tools are used to calibrate the values of the HW model parameters. Hybridization is conducted to reduce modeling uncertainty. The technique is applied to real load data. Monthly peak load forecasts are calculated as weighted averages of HW and MCMC estimates. Mean absolute percentage error and the coefficient of determination (R2) indices are used to evaluate forecasts. Findings The developed hybrid methodology offers advantages over both individual combined techniques and reveals more accurate and impressive results with R2 above 0.97. The new technique can be used to assist energy networks in planning and implementing production projects that can ensure access to reliable and modern energy services to meet the sustainable development goal in this sector. Originality/value This is original research.

Predicting likelihood of cost overrun in educational projects

2017 ◽  
Vol 24 (1) ◽  
pp. 21-39 ◽  
Author(s):  
Richard Ohene Asiedu ◽  
Nana Kena Frempong ◽  
Hans Wilhelm Alfen
Keyword(s):  
Linear Regression Analysis ◽  
Model Development ◽  
Multiple Linear Regression Analysis ◽  
Coefficient Of Determination ◽  
Percentage Error ◽  
Cost Overrun ◽  
Cost Overruns ◽  
Content Type ◽  
Public Projects ◽  
Contract Signing

Purpose Being able to predict the likelihood of a project to overrun its cost before the contract signing phase is crucial in developing the required mitigating measures to avert it. Known parameters that permit the timely prediction of cost overrun provide the basis for such predictions. Therefore, the purpose of this paper is to develop a model for forecasting cost overruns. Design/methodology/approach Ten predictive variables known before the contract signing phase of a project are identified. Based on a survey approach, information on 321 educational projects completed are compiled. A multiple linear regression analysis is adopted for the model development. Findings Five variables – initial contract sum, gross floor area, number of storeys, source of funds and contractors’ financial classification are observed to influence cost overruns. The model, however, yields a fairly weak coefficient of determination with a mean absolute percentage error of 30.22 and 138 per cent, respectively. Research limitations/implications The model developed focussed on data only educational projects sampled from three out of the ten administration regions in Ghana based on a purposive sampling approach. Practical implications Policy makers and construction managers working on public projects stand to gain tremendous assistance in formulating and strengthening their own in-house cost forecasting at the precontract phase based on “what if” analysis to generate various alternative predictions of cost overruns. Originality/value Considering the innate nature of cost overruns within the Ghanaian construction industry often resulting to project abandonment, this research presents a unique dimension for tackling cost overruns based on a predictive approach.

An unbiased GM(1,1)-based new hybrid approach for time series forecasting

2016 ◽  
Vol 6 (3) ◽  
pp. 322-340 ◽  
Author(s):  
R.M. Kapila Tharanga Rathnayaka ◽  
D.M.K.N. Seneviratna ◽  
Wei Jianguo ◽  
Hasitha Indika Arumawadu
Keyword(s):  
Time Series ◽  
Hybrid Approach ◽  
Time Series Forecasting ◽  
Series Data ◽  
Percentage Error ◽  
Second Phase ◽  
Suitable Model ◽  
Content Type ◽  
Ex Post ◽  
Hybrid Forecasting

Purpose The time series forecasting is an essential methodology which can be used for analysing time series data in order to extract meaningful statistics based on the information obtained from past and present. These modelling approaches are particularly complicated when the available resources are limited as well as anomalous. The purpose of this paper is to propose a new hybrid forecasting approach based on unbiased GM(1,1) and artificial neural network (UBGM_BPNN) to forecast time series patterns to predict future behaviours. The empirical investigation was conducted by using daily share prices in Colombo Stock Exchange, Sri Lanka. Design/methodology/approach The methodology of this study is running under three main phases as follows. In the first phase, traditional grey operational mechanisms, namely, GM(1,1), unbiased GM(1,1) and nonlinear grey Bernoulli model, are used. In the second phase, the new proposed hybrid approach, namely, UBGM_BPNN was implemented successfully for forecasting short-term predictions under high volatility. In the last stage, to pick out the most suitable model for forecasting with a limited number of observations, three model-accuracy standards were employed. They are mean absolute deviation, mean absolute percentage error and root-mean-square error. Findings The empirical results disclosed that the UNBG_BPNN model gives the minimum error accuracies in both training and testing stages. Furthermore, results indicated that UNBG_BPNN affords the best simulation result than other selected models. Practical implications The authors strongly believe that this study will provide significant contributions to domestic and international policy makers as well as government to open up a new direction to develop investments in the future. Originality/value The new proposed UBGM_BPNN hybrid forecasting methodology is better to handle incomplete, noisy, and uncertain data in both model building and ex post testing stages.

scholarly journals Estimation of life-cycle costs of buildings: regression vs artificial neural network

2016 ◽  
Vol 6 (1) ◽  
pp. 30-43 ◽  
Author(s):  
Ayedh Alqahtani ◽  
Andrew Whyte
Keyword(s):  
Life Cycle ◽  
Cost Estimation ◽  
Coefficient Of Determination ◽  
Percentage Error ◽  
Data Set ◽  
Building Projects ◽  
Content Type ◽  
Testing Stage ◽  
Running Cost ◽  
Artificial Neural

Purpose – The purpose of this paper is to compare the performance of regression and artificial-neural-networks (ANNs) methods to estimate the running cost of building projects towards improved accuracy. Design/methodology/approach – A data set of 20 building projects is used to test the performance of these two (ANNs/regression) models in estimating running cost. The concept of cost-significant-items is identified as important in assisting estimation. In addition, a stepwise technique is used to eliminate insignificant factors in regression modelling. A connection weight method is applied to determine the importance of cost factors in the performance of ANNs. Findings – The results illustrate that the value of the coefficient of determination=99.75 per cent for ANNs model(s), with a value of 98.1 per cent utilising multiple regression (MR) model(s); second, the mean percentage error (MPE) for ANNs at a testing stage is 0.179, which is less than that of the MPE gained through MR modelling of 1.28; and third, the average accuracy is 99 per cent for ANNs model(s) and 97 per cent for MR model(s). On the basis of these results, it is concluded that an ANNs model is superior to a MR model when predicting running cost of building projects. Research limitations/implications – A means for continuous improvement for the performance of the models accuracy has been established; this may be further enhanced by future extended sample. Originality/value – This work extends the knowledge base of life-cycle estimation where ANNs method has been found to reduce preparation time consumed and increasing accuracy improvement of the cost estimation.

Wear Modeling Revisited Using Electrical Analogy

2017 ◽  
Vol 139 (6) ◽  
Author(s):  
M. Hanief ◽  
M. F. Wani
Keyword(s):  
Electrical Circuit ◽  
Operating Conditions ◽  
Coefficient Of Determination ◽  
Percentage Error ◽  
Model Parameters ◽  
Wear Process ◽  
Wear Modeling ◽  
Proposed Model ◽  
Electrical Analogy ◽  
High Degree

Electrical analogy has been used extensively in modeling various mechanical systems such as thermal, hydraulic, and other dynamic systems. However, wear modeling of a tribosystem using electrical analogy has not been reported so far. In this paper, an equivalent electrical analogous system is proposed to represent the wear process. An analogous circuit is developed by mapping the wear process parameters to that of the electrical parameters. The circuit, thus, developed is solved by conventional electrical circuit theory. The material properties and operating conditions are taken into account by model parameters. Accordingly, a model equation in terms of model parameters is developed to represent the wear rate. It is also demonstrated how this methodology can be used to take various system parameters into account by incorporating the equivalent resistance of the parameters. The nonlinear model parameters are evaluated by Gauss–Newton (GN) algorithm. The proposed model is validated by using experimental data. A comparison of the proposed model with the experimental results, based on statistical methods: coefficient of determination (R2), mean-square-error (MSE) and mean absolute percentage error (MAPE), indicates that the model is competent to predict the wear with a high degree of accuracy.

scholarly journals Influence of Operating Parameters on Running-in Wear of EN-31 Steel

2016 ◽  
Vol 13 (1) ◽  
pp. 1-2
Author(s):  
M. Hanief ◽  
M. F. Wani
Keyword(s):  
Surface Roughness ◽  
Wear Rate ◽  
Least Square Method ◽  
Least Square ◽  
Coefficient Of Determination ◽  
Percentage Error ◽  
Model Parameters ◽  
Operating Parameters ◽  
Proposed Model ◽  
En 31

Abstract In this paper, effect of operating parameters (temperature, surface roughness and load) was investigated to determine the influence of each parameter on the wear rate. A mathematical model was developed to establish a functional relationship between the running-in wear rate and the operating parameters. The proposed model being non-linear, it was linearized by logarithmic transformation and the optimal values of model parameters were obtained by least square method. It was found that the surface roughness has significant effect on wear rate followed by load and temperature. The adequacy of the model was estimated by statistical methods (coefficient of determination (R2) and mean absolute percentage error (MAPE)) .

Identification of particular hydrodynamic parameters for a modular type 4 DOF underwater vehicle by means of CFD method

2022 ◽  
Vol ahead-of-print (ahead-of-print) ◽  
Author(s):  
Serhat Yilmaz ◽  
Gülten Altıokka Yılmaz
Keyword(s):  
Turbulent Flow ◽  
Stokes Equations ◽  
Flow Analysis ◽  
Underwater Vehicle ◽  
Hydrodynamic Forces ◽  
Percentage Error ◽  
Model Parameters ◽  
Cad Model ◽  
Content Type ◽  
Hydrodynamic Parameters

Purpose The development of robust control algorithms for the position, velocity and trajectory control of unmanned underwater vehicles (UUVs) depends on the accuracy of their mathematical models. Accuracy of the model is determined by precise estimation of the UUV hydrodynamic parameters. The purpose of this study is to determine the hydrodynamic forces and moments acting on an underwater vehicle with complex body geometry and moving at low speeds and to achieve the accurate coefficients associated with them. Design/methodology/approach A three-dimensional (3D) computer-aided design (CAD) model of UUV is designed with one-to-one dimensions. 3D fluid flow simulations are conducted using computational fluid dynamics (CFD) software programme in the solution of Navier Stokes equations for laminar and turbulent flow analysis. The coefficients depending on the hydrodynamic forces and moments are determined by the external flow analysis using the CFD programme. The Flow Simulation k-ε turbulence model is used for the transition from laminar flow to turbulent flow. Hydrodynamic properties such as lift and drag coefficients and roll and yaw moment coefficients are calculated. The parameters are compared with the coefficient values found by experimental methods. Findings Although the modular type UUV has a complex body geometry, the comparative results of the experiments and simulations confirm that the defined model parameters are accurate and close to the actual experimental values. In the proposed k-ε method, the percentage error in the estimation of drag and lifting coefficients is decreased to 4.2% and 8.39%, respectively. Practical implications The model coefficients determined in this study can be used in high-level control simulations which leads to the development of robust real-time controllers for complex-shaped modular UUVs. Originality/value The Lucky Fin UUV with 4 degrees of freedom is a specific design and its CAD model is first extracted. Verification of simulation results by experiments is generally less referenced in studies. However, it provides more precise parameter identification of the model. Proposed study offers a simple and low-cost experimental measurement method for verification of the hydrodynamic parameters. The extracted model and coefficients are worthwhile references for the analysis of modular type UUVs.

Modeling of a simplified hybrid algorithm for short-term load forecasting in a power system network

2021 ◽  
Vol ahead-of-print (ahead-of-print) ◽  
Author(s):  
Kathiresh Mayilsamy ◽  
Maideen Abdhulkader Jeylani A, ◽  
Mahaboob Subahani Akbarali ◽  
Haripranesh Sathiyanarayanan
Keyword(s):  
Time Series ◽  
Hybrid Algorithm ◽  
Hybrid Approach ◽  
Load Forecasting ◽  
Series Data ◽  
Percentage Error ◽  
Short Term ◽  
Content Type ◽  
Highly Nonlinear ◽  
Short Term Load Forecasting

Purpose The purpose of this paper is to develop a hybrid algorithm, which is a blend of auto-regressive integral moving average (ARIMA) and multilayer perceptron (MLP) for addressing the non-linearity of the load time series. Design/methodology/approach Short-term load forecasting is a complex process as the nature of the load-time series data is highly nonlinear. So, only ARIMA-based load forecasting will not provide accurate results. Hence, ARIMA is combined with MLP, a deep learning approach that models the resultant data from ARIMA and processes them further for Modelling the non-linearity. Findings The proposed hybrid approach detects the residuals of the ARIMA, a linear statistical technique and models these residuals with MLP neural network. As the non-linearity of the load time series is approximated in this error modeling process, the proposed approach produces accurate forecasting results of the hourly loads. Originality/value The effectiveness of the proposed approach is tested in the laboratory with the real load data of a metropolitan city from South India. The performance of the proposed hybrid approach is compared with the conventional methods based on the metrics such as mean absolute percentage error and root mean square error. The comparative results show that the proposed prediction strategy outperforms the other hybrid methods in terms of accuracy.

scholarly journals A Wavelet Based Solar Radiation Prediction in Nigeria Using Adaptive Neuro-Fuzzy Approach

2018 ◽  
Vol 12 (3) ◽  
pp. 907 ◽  
Author(s):  
Sani Salisu ◽  
Mohd Wazir Mustafa ◽  
Mamunu Mustapha
Keyword(s):  
Solar Radiation ◽  
Meteorological Data ◽  
Hybrid Approach ◽  
Model Performance ◽  
Coefficient Of Determination ◽  
Maximum Temperature ◽  
Percentage Error ◽  
Anfis Model ◽  
Inference System ◽  
Neuro Fuzzy

<p><span>In this study, a hybrid approach combining an Adaptive Neuro-Fuzzy Inference System (ANFIS) and Wavelet Transform (WT) is examined for solar radiation prediction in Nigeria. Meteorological data obtained from NIMET Nigeria comprising of </span><span lang="EN-MY">monthly mean minimum temperature, maximum temperature, relative humidity and sunshine hours were used as inputs to the model and monthly mean solar radiation was used as the model output. The data used was divided into two for training and testing, with 70% used during the training phase and 30% during the testing phase. The hybrid model performance is assessed using three statistical evaluators, Mean Absolute Percentage Error (MAPE), Root Mean Square Error (RMSE) and Coefficient of determination </span><span lang="EN-SG">(R<sup>2</sup>). According to the results obtained, a very accurate prediction was achieved by the WT- ANFIS model by improving the value of (R<sup>2</sup>) by at least 14% and RMSE by at least 78% when compared with other existing models. And a MAPE of 2% is recorded using the proposed approach. The obtained results prove the developed WT-ANFIS model as an efficient tool for solar radiation prediction.</span></p>

PENGARUH SARANA BELAJAR DAN MOTIVASI BELAJAR TERHADAP KEMANDIRIAN BELAJAR MAHASISWA (Studi Empirical Pada Mahasiswa Beasiswa Bidikmisi UPBJJ-UT Ternate)

2018 ◽  
Vol 19 (2) ◽  
pp. 68
Author(s):  
Raden Sudarwo ◽  
Yusuf Yusuf ◽  
Anfas Anfas
Keyword(s):  
Student Learning ◽  
Significant Influence ◽  
Sense Of Self ◽  
Regression Equation ◽  
Learning Motivation ◽  
Coefficient Of Determination ◽  
Learning Tools ◽  
Constant Number ◽  
Self Sufficiency ◽  
Student Independence

This study aims to determine the influence of learning facilities and student learning motivation towards the independence of student learning. The result of the research shows that there is positive and significant influence of learning tool (X1) on learning independence (Y). It is obtained by tvalue (2,159) with p = 0,034 <0,05 and ttable at 5% significant level with df = 78 equal to 1,991. There is a positive and significant influence of learning motivation (X2) on learning independence (Y). It is obtained tvalue (7,858) with p = 0,000 <0,05 and ttable at 5% significant level with df = 78 equal to 1,991. There is a positive and significant influence of learning facilities (X1) and learning motivation (X2) simultaneously to the independence of learning (Y). This shows the coefficient of double correlation RY (1,2) = 0,746 and R² = 0,557 and price Fvalue equal to 48,980 with p = 0,000 <0,05 and Ftable = 3,11 at 5% significant level. Coefficient value X1 = 0,186 and X2 = 0,647, constant number equal to 8,650 so that can be made regression equation Y = 8,650 + 0,186X1 + 0,647X2. The higher the learning means (X1) and the learning motivation (X2), the higher the learning independence (Y). Coefficient of Determination is R² of 0,557. Means 55,7% learning independence is explained by learning tools and learning motivation. Meanwhile, 44,3% is explained by other factors not discussed in this study. The study concludes that partially, learning facilities and student learning motivation has a positive and significant effect on student independence (self-sufficiency) in learning.  In addition, both learning facility and motivation have a positive and significant effect on student learning independence or sense of self-sufficiency. Penelitian ini bertujuan untuk mengetahui pengaruh fasilitas belajar dan motivasi belajar siswa terhadap kemandirian belajar siswa. Hasil penelitian menunjukkan bahwa ada pengaruh yang positif dan signifikan sanara belajar (X1) terhadap kemandirian belajar (Y). Hal ini diperoleh dengan nilai thitung (2,159) dengan p = 0,034 <0,05 dan ttabel pada 5% tingkat signifikan dengan df = 78 sama dengan 1,991. Ada pengaruh positif dan signifikan motivasi belajar (X2) pada kemandirian belajar (Y). Diperoleh nilai thitung (7,858) dengan p = 0,000 <0,05 dan ttabel pada taraf signifikan 5% dengan df = 78 sebesar 1,991. Ada pengaruh yang positif dan signifikan dari fasilitas belajar (X1) dan motivasi belajar (X2) secara bersamaan terhadap kemandirian belajar (Y). Hal ini menunjukkan koefisien korelasi ganda RY (1,2) = 0,746 dan R² = 0,557 dan harga Fhitung sebesar 48,980 dengan p = 0,000 <0,05 dan Ftabel = 3,11 pada taraf signifikan 5%. Nilai koefisien X1 = 0,186 dan X2 = 0,647, bilangan konstan sebesar 8,650 sehingga dapat dibuat persamaan regresi Y = 8,650 + 0,186X1 + 0,647X2. Semakin tinggi nilai sarana belajar (X1) dan motivasi belajar (X2), semakin tinggi kemandirian belajar (Y). Koefisien Determinasi adalah R² 0,557. Berarti 55,7% kemandirian belajar dijelaskan oleh alat belajar dan motivasi belajar. Sementara itu, 44,3% dijelaskan oleh faktor-faktor lain yang tidak dibahas dalam penelitian ini. Penelitian ini menyimpulkan bahwa secara parsial, baik ketersediaan sarana prasaran belajar dan motivasi berpengaruh positif dan signifikan pada kemandirian mahasiswa, dari dari kedua variable tersebut motivasi mempunyai pengaruh lebih besar. Secara simultan ketersediaan sarana prasarana dalam belajar dan pembelajaran, serta motivasi berpengaruh positif terhadap kemandirian belajar.

scholarly journals Predicting Ink Transfer Rate of 3D Additive Printing Using EGBO Optimized Least Squares Support Vector Machine Model

2020 ◽  
Vol 2020 ◽  
pp. 1-12
Author(s):  
Shengpu Li ◽  
Yize Sun
Keyword(s):  
Support Vector Machine ◽  
Prediction Model ◽  
Least Squares ◽  
Transfer Rate ◽  
Coefficient Of Determination ◽  
Percentage Error ◽  
Support Vector ◽  
Experimental Sample ◽  
Machine Model ◽  
Ink Transfer

Ink transfer rate (ITR) is a reference index to measure the quality of 3D additive printing. In this study, an ink transfer rate prediction model is proposed by applying the least squares support vector machine (LSSVM). In addition, enhanced garden balsam optimization (EGBO) is used for selection and optimization of hyperparameters that are embedded in the LSSVM model. 102 sets of experimental sample data have been collected from the production line to train and test the hybrid prediction model. Experimental results show that the coefficient of determination (R2) for the introduced model is equal to 0.8476, the root-mean-square error (RMSE) is 6.6 × 10 (−3), and the mean absolute percentage error (MAPE) is 1.6502 × 10 (−3) for the ink transfer rate of 3D additive printing.

Sign in / Sign up

Export Citation Format

Share Document