Short-term Power Demand Forecasting using the Differential Polynomial Neural Network

Abstract

Power demand forecasting is important for economically efficient operation and effective control of power systems and enables to plan the load of generating unit. The purpose of the short-term electricity demand forecasting is to forecast in advance the system load, represented by the sum of all consumers load at the same time. A precise load forecasting is required to avoid high generation cost and the spinning reserve capacity. Under-prediction of the demands leads to an insufficient reserve capacity preparation and can threaten the system stability, on the other hand, over-prediction leads to an unnecessarily large reserve that leads to a high cost preparations. Differential polynomial neural network is a new neural network type, which forms and resolves an unknown general partial differential equation of an approximation of a searched function, described by data observations. It generates convergent sum series of relative polynomial derivative terms which can substitute for the ordinary differential equation, describing 1-parametric function time-series. A new method of the short-term power demand forecasting, based on similarity relations of several subsequent day progress cycles at the same time points is presented and tested on 2 datasets. Comparisons were done with the artificial neural network using the same prediction method.Keywords: power demand prediction, week and day load cycle, differential polynomial neural network, sum relative derivative term, ordinary differential equation composition * Typeset names in 10 pt Times Roman, uppercase. Use the footnote to indicate the present or permanent address of the author.

Bayesian Optimized Echo State Network Applied to Short-Term Load Forecasting

“…Besides, accurate load forecasting is critical for power system planning and operational decision making. Nevertheless, recent approaches to load forecasting consume a huge amount of available load time series data to produce machine learning models for load forecasting such as deep learning neural networks [1][2][3][4][5][6], ensemble of artificial neural networks [7][8][9][10][11][12], single artificial neural networks [13][14][15][16], Gaussian process [17], long short-term memory networks (LSTM) [18][19][20][21][22], deep belief networks [23,24], heterogeneous ensemble methods [25][26][27][28], k-nearest neighbor [29], echo state network [30,31], deep echo state network [32,33], ensemble of echo state networks [34], extreme learning machines [35,36], ensemble learning of regression trees [37], support vector machines tuned with particle swarm optimization (PSO) algorithm [38], and optimized artificial neural networks [39][40][41][42][43]. A review may be found in …”

Determination and mapping the spatial distribution of radioactivity of natural spring water in the Eastern Black Sea Region by using artificial neural network method

“…Such structures are processing structures that calculate the weight of connections on the system. In recent years, artificial neural network technique is used in many different areas, and quite successful results are obtained even in complex problems (Amanatiadis et al 2014;Bagheri et al 2014;Das et al 2015;Falamaki 2013;Farah et al 2011;Negarestani et al 2002;Otağ et al 2015;Raza et al 2012;Vega-Carrillo et al 2007;Zjavka 2014). The purpose of ANN method is making proper predictions for samples never seen before by using known samples collected before or previously measured results of the region.…”