ArticlePublisher preview available

Day similarity metric model for short-term load forecasting supported by PSO and artificial neural network

Authors:
Zoran Janković at University of Novi Sad
Zoran Janković
Aleksandar Selakov
Aleksandar Selakov
Aleksandar Selakov
  • This person is not on ResearchGate, or hasn't claimed this research yet.
Dusko Bekut at University of Novi Sad
Dusko Bekut
Marija Đorđević
Marija Đorđević
Marija Đorđević
  • This person is not on ResearchGate, or hasn't claimed this research yet.
Read publisher preview
Download citation
To read the full-text of this research, you can request a copy directly from the authors.

Abstract and Figures

This paper proposes a new model for optimal similar days selection and its use in short-term load forecasting based on artificial neural network. Proposed day similarity metric model is based on the multi-filtering process and introduces a few novelties: (1) introduction of pre-history of similar days in a selection process; (2) extension of forecasting factors: load inertia, daylight duration and load profiles; (3) open model with possibility to add additional contribution factors; (4) particle swarm optimization is applied for calculation of the impact of different contributing factors. This approach results in optimal similar days selection even in a case where it is not obvious in advance which factors are the most relevant. Finally, the artificial neural network is used as a basic procedure for the short-term load forecast. The proposed model has been tested in the transmission system utility in Serbia, and the results are presented.
Figures - available from: Electrical Engineering
This content is subject to copyright. Terms and conditions apply.
A preview of this full-text is provided by Springer Nature.
Learn more
Content available from Electrical Engineering
This content is subject to copyright. Terms and conditions apply.
Vol.:(0123456789)
1 3
Electrical Engineering (2021) 103:2973–2988
https://doi.org/10.1007/s00202-021-01286-6
ORIGINAL PAPER
Day similarity metric model forshort‑term load forecasting supported
byPSO andartificial neural network
ZoranJanković1· AleksandarSelakov1 · DuškoBekut1· MarijaĐorđević2
Received: 7 September 2019 / Accepted: 1 April 2021 / Published online: 16 April 2021
© The Author(s), under exclusive licence to Springer-Verlag GmbH Germany, part of Springer Nature 2021
Abstract
This paper proposes a new model for optimal similar days selection and its use in short-term load forecasting based on
artificial neural network. Proposed day similarity metric model is based on the multi-filtering process and introduces a few
novelties: (1) introduction of pre-history of similar days in a selection process; (2) extension of forecasting factors: load
inertia, daylight duration and load profiles; (3) open model with possibility to add additional contribution factors; (4) particle
swarm optimization is applied for calculation of the impact of different contributing factors. This approach results in opti-
mal similar days selection even in a case where it is not obvious in advance which factors are the most relevant. Finally, the
artificial neural network is used as a basic procedure for the short-term load forecast. The proposed model has been tested
in the transmission system utility in Serbia, and the results are presented.
Keywords Artificial neural network· Particle swarm optimization· Short-term load forecast· Similar days
Abbreviations
STLF Short-term load forecasting
DSMM Day similarity metric model
PSO Particle swarm optimization
ANN Artificial neural network
GA Genetic algorithm
Sc Similarity coefficient
pc Participation coefficient
wf Weighting coefficient
FA Forecasted array
HA Historical array
nfw Number of weather factors
dd Daylight duration
HA Historical array
FA Forecasted array
in Inertia
ddwf Day decreasing weighting coefficient
avgP Average load
d Day
σ Standard load deviation
Cσ Load deviation contribution
Lσ Selected border value
MAPE Mean absolute percentage error
Pf Forecasted load
Pr Realized load
1 Introduction
In modern electrical utilities, short-term load forecasting is
essential for reliable power system operations, and it signifi-
cantly affects power system economics. Forecasting electri-
cal load a day ahead, usually referred to as short-term load
forecasting (STLF), is one of the critical needs for electrical
utility and transmission system operators to make decisions
regarding purchasing and selling electric power/energy,
balancing of the system, maintenance, load switching and
maintenance planning.
A typical STLF process relies on the forecasted weather
information and takes weather and load history as the inputs
to the modeling process. After the estimation of the param-
eters, the model and weather forecast are extrapolated to
generate the final forecast [1].
Among many STLF approaches, one of the critical steps
in the STLF process is similar days selection [2–26, 39, 40].
* Aleksandar Selakov
aselakov@uns.ac.rs
1 University ofNovi Sad, Faculty ofTechnical Sciences,
Department ofPower, Electronic andTelecommunication
Engineering, Trg Dositeja Obradovića 6, 21000NoviSad,
Serbia
2 Serbian Transmission System Company, 11 Kneza Miloša,
11000Belgrade, Serbia
Content courtesy of Springer Nature, terms of use apply. Rights reserved.
... The most similar day was then used as the forecast [14]. The concept of the SDM was used in [15,16] to develop a day similarity metric model for the STLF (DSMM-STLF tool) that performed well in practical applications. However, the methodology based on this method lacks good results when there are no similar days in the historical data or the forecasted day is a special day, like a national or religious holiday. ...
... Holidays and other predetermined special days were excluded from the experiment. The solution proposed in this paper is an improvement of the DSMM-STLF tool and methodologies presented in [15][16][17]. The process structure and flow of the proposed solution are illustrated in Figure 1. ...
... The proposed work is the culmination of several years of research, with multiple papers published throughout this time [13,[15][16][17][38][39][40][41]. Older models have been rigorously tested across diverse datasets, geographies, and power systems, and the current model represents the latest advancement in this ongoing development. ...
Autoencoder-Driven Training Data Selection Based on Hidden Features for Improved Accuracy of ANN Short-Term Load Forecasting in ADMS
Article
Full-text available
  • Oct 2024
This paper presents a novel methodology for short-term load forecasting in the context of significant shifts in the daily load curve due to the rapid and extensive adoption of Distributed Energy Resources (DERs). The proposed solution, built upon the Similar Days Method (SDM) and Artificial Neural Network (ANN), introduces several novelties: (1) selection of similar days based on hidden representations of day data using Autoencoder (AE); (2) enhancement of model generalization by utilizing a broader set of training examples; (3) incorporating the relative importance of training examples derived from the similarity measure during training; and (4) mitigation of the influence of outliers by applying an ensemble of ANN models trained with different data splits. The presented AE configuration and procedure for selecting similar days generated a higher-quality training dataset, which led to more robust predictions by the ANN model for days with unexpected deviations. Experiments were conducted on actual load data from a Serbian electrical power system, and the results were compared to predictions obtained by the field-proven STLF tool. The experiments demonstrated an improved performance of the presented solution on test days when the existing STLF tool had poor predictions over the past year.
View
... In this context, forecasting methods based on neural networks [11] have been increasingly used in recent years in multiple applications. In general, these approaches allow flexible implementations and yield more accurate results compared to traditional methods such as regression and exponential smoothing [12]. Current research related to power systems uses neural networks-based solutions to predict quantities such as short-term energy demand (in [12], authors present a model for short-term load forecasting that relies on particle swarm optimisation [13] and on artificial neural networks; in [14], authors propose a model based on a recurrent long short-term memory (LSTM) neural network; and in [15], the authors employ empirical wavelet transforms and random forests [16] to the developed short-term forecasting model), solar irradiation (in [17], the authors use wavelets combined with recurrent neural networks; and in [18], authors employ use LSTM networks), photo-voltaic generation (in [19], authors employ a super-resolution perception approach; and in [20], LSTM neural networks are used) and, also, power transformer temperature. ...
... In general, these approaches allow flexible implementations and yield more accurate results compared to traditional methods such as regression and exponential smoothing [12]. Current research related to power systems uses neural networks-based solutions to predict quantities such as short-term energy demand (in [12], authors present a model for short-term load forecasting that relies on particle swarm optimisation [13] and on artificial neural networks; in [14], authors propose a model based on a recurrent long short-term memory (LSTM) neural network; and in [15], the authors employ empirical wavelet transforms and random forests [16] to the developed short-term forecasting model), solar irradiation (in [17], the authors use wavelets combined with recurrent neural networks; and in [18], authors employ use LSTM networks), photo-voltaic generation (in [19], authors employ a super-resolution perception approach; and in [20], LSTM neural networks are used) and, also, power transformer temperature. ...
A forecasting methodology based on second-generation wavelets and neural networks: application to power transformer oil temperature
Article
Full-text available
  • Nov 2023
  • ELECTR ENG
This paper presents a novel forecasting methodology based on neural networks and second-generation wavelets, and an example of application of such methodology for forecasting the oil temperature of an actual power transformer. The proposed method consists on selecting the variables with the highest correlation concerning the magnitude to be predicted (in situations featuring multiple variables), selecting the most appropriate second-generation wavelet from a set of candidate lifting wavelet transforms to smooth each variable’s data and using the values smoothed by the selected second-generation wavelet to train the considered neural network. The authors applied the proposed methodology to predict the oil temperature of an existing transformer for four days and for one month using field data collected during four years of equipment operation. The obtained results indicate that selecting the most appropriate second-generation wavelet considering different selection techniques and using them to smooth the input variables significantly reduce the forecasting errors. The application of the proposed approach to forecast the oil temperature from the field-collected data using a long short-term memory neural network considering different batch sizes, neuron numbers and second-generation wavelet selection techniques decreased the average forecasting errors presented by using just the considered long short-term memory network by up to eleven times. The innovations of this proposal about literature are the use of correlation-based variable selection techniques given improving neural network forecasting accuracy, which is unprecedented in transformer insulating oil temperature forecasting applications, and the employment of SGWs to smooth the network’s inputs to increase the accuracy of the model significantly. For the application case study presented in this research, the minor average forecasting errors in percentage to predict the four-day and one-month future temperature were, respectively, 0.012 and 0.326%. Also, for one month of forecasting, the average error in degrees Celsius was less than 0.2.
View
... However, they have limited capability in handling nonlinearity. Machine learning methods encompass techniques such as random forests [17], support vector machines [18], and artificial neural networks [19]. By incorporating machine learning algorithms, these methods address the nonlinear relationships among data effectively. ...
... When we introduce the TS2Vec layer, we achieve a MAPE of 6.01% on dataset area1. This results in a substantial reduction of19.11% compared to the baseline. ...
TS2ARCformer: A Multi-Dimensional Time Series Forecasting Framework for Short-Term Load Prediction
Article
Full-text available
  • Aug 2023
Accurately predicting power load is a pressing concern that requires immediate attention. Short-term load prediction plays a crucial role in ensuring the secure operation and analysis of power systems. However, existing research studies have limited capability in extracting the mutual relationships of multivariate features in multivariate time series data. To address these limitations, we propose a multi-dimensional time series forecasting framework called TS2ARCformer. The TS2ARCformer framework incorporates the TS2Vec layer for contextual encoding and utilizes the Transformer model for prediction. This combination effectively captures the multi-dimensional features of the data. Additionally, TS2ARCformer introduces a Cross-Dimensional-Self-Attention module, which leverages interactions across channels and temporal dimensions to enhance the extraction of multivariate features. Furthermore, TS2ARCformer leverage a traditional autoregressive component to overcome the issue of deep learning models being insensitive to input scale. This also enhances the model’s ability to extract linear features. Experimental results on two publicly available power load datasets demonstrate significant improvements in prediction accuracy compared to baseline models, with reductions of 43.2% and 37.8% in the aspect of mean absolute percentage error (MAPE) for dataset area1 and area2, respectively. These findings have important implications for the accurate prediction of power load and the optimization of power system operation and analysis.
View
... The paper [21] shows a long short-term memory neural network mode for predicting the load for a day-ahead in advance and solar insolation of residential consumers who are equipped with rooftop PV based on weather classification. In [22], a model for selecting similar days that goes through several levels and short-term load forecasting based on artificial neural network is presented. In [23], the application of an artificial neural network (ANN) enabled the prediction of electricity consumption in the USA from 2021 to 2030. ...
Application of artificial neural network to power consumption forecasting for the Sarajevo region
Article
Full-text available
  • Sep 2024
  • ELECTR ENG
This paper presents an innovative method for forecasting power consumption in the power system using an artificial neural network (ANN). The method was validated in the case of predicting power consumption for the Sarajevo region in Bosnia and Herzegovina. Power consumption is planned daily for the day-ahead with hourly resolution. Measured data on air temperature, wind speed, and insolation for 2017 to 2020 were utilized as input variables in the proposed power consumption forecasting method. The influence of these input variables on power consumption was analyzed using the Pearson correlation coefficient. The neural network underwent training with data on input variables and power consumption from 2017 to 2020 and was subsequently applied to forecast day-ahead power consumption for 2021. Due to the implementation of a neural network with a greater number of input variables, a smaller error in the power consumption forecast for 2021 was achieved compared to the forecast performed by the Electric Power Company. Therefore, the proposed method can be used as a more reliable tool for day-ahead power consumption forecasting. Additionally, the continual increase in the historical data on power consumption and influencing variables over time is expected to further enhance the reliability of power consumption forecasting using ANN.
View
... Fig. 2 shows a population matrix with a population size value of m and number n of genes in each chromosome [9]. Variables to be optimised were represented by genes [10]. Each cell in the population matrix represents one gene. ...
View
View
View
View
View
Research on Power Load Forecasting Using Deep Neural Network and Wavelet Transform
Article
Full-text available
  • Jan 2023
In today's rapid economic development, industrial and civil electricity consumption is growing year by year, and how to guarantee stability of power system operation has become the focus of attention of the power sector in each country. Power load forecasting has been closely associated with the modernization of power system management and is a vital guarantee for the safe and stable operation and economic efficiency of the power system. In this article, the authors propose a recurrent neural network (RNN) decision fusion forecasting framework based on the wavelet transform to address the power load forecasting problem. The framework firstly performs the wavelet transform on the power load data and uses Daubechies wavelets to extract the high-frequency and low-frequency parts of the data; then the data with different frequencies are combined with the original data and fed into the RNN model separately, and the decision fusion is performed in the output layer; finally, the prediction results are obtained by superposition of two RNN networks. The results showed that the error of the predicted data in the last nine years decreased by 50%, compared with the traditional method of feeding the data into the RNN model for training, which provides a new idea for future power load forecasting.
View
Two-Stage Short-Term Load Forecasting for Power Transformers Under Different Substation Operating Conditions
Article
Full-text available
  • Nov 2019
The load of transformers shows higher volatility and uncertainty than do the system-level and substation-level loads. This paper proposes a two-stage short-term load forecasting (STLF) model for power transformers. 1) Three state-of-the-art technologies are applied to predict the aggregated substation-level load by taking the historical load, weather, and calendar data as inputs. In this stage, no specific STLF model needs to be developed, which allows the forecasters to select the most accurate prediction results for transformer-level load forecasting. 2) The load distribution factor (LDF) is defined as the ratio of the transformer load to the substation load. The relationship between LDF and substation load is captured by nonlinear regression functions under different substation operating conditions, and the load of each parallel transformer is predicted using these nonlinear regression functions. Each nonlinear function can be accurately established even if the historical load data are scarce under some irregular operating conditions. Three application examples show the effectiveness and rationality of the proposed method. The third example demonstrates that STLF of transformers is necessary because it provides important information for optimizing substation operating schemes and equipment maintenance plans.
View
Multi-objective optimal power flow considering wind power cost functions using enhanced PSO with chaotic mutation and stochastic weights
Article
Full-text available
  • Sep 2019
  • ELECTR ENG
A multi-objective optimal power flow (OPF) solution using an enhanced NSPSO, incorporating chaotic mutation and stochastic weight trade-off features, is proposed here. The objective functions considered embody the different aspects of the power system, viz. financial, reliability and operational efficiency. The proposed OPF algorithm balances the exploration of global and the utilization of local bests with the stochastic weight and dynamic coefficient trade-off techniques, thus enhancing the searching capability. Also, for countering the premature convergence issue and to improve diversity, the feature of chaotic mutation is incorporated. The pareto-optimal front is provided by the combination of crowding distance approach and non-dominated sorting principle. The best solution can be obtained from a dual-stage process, by selecting from the collective of local best compromises using a fuzzy function. To assess the performance, the proposed method is tested on a standard IEEE 30-bus test system and is compared with popular multi-objective algorithms. IEEE 118-bus system is also used for testing applicability to large systems. Based on the obtained results, the proposed method gives a more reliable operating solutions, better optimal front and hence an improved solution providing a better trade-off.
View
View
A Hybrid clustering and classification technique for forecasting short‐term energy consumption
Article
Full-text available
This paper presents a hybrid approach to predict the electric energy usage of weather‐sensitive loads. The presented method utilizes the clustering paradigm along with ANN and SVM approaches for accurate short‐term prediction of electric energy usage, using weather data. Since the methodology being invoked in this research is based on CRISP data mining, data preparation has received a great deal of attention in this research. Once data pre‐processing was done, the underlying pattern of electric energy consumption was extracted by the means of machine learning methods to precisely forecast short‐term energy consumption. The proposed approach (CBA‐ANN‐SVM) was applied to real load data and resulting higher accuracy comparing to the existing models. © 2018 American Institute of Chemical Engineers Environ Prog, 38: 66–76, 2019
View
A Hybrid BA-ELM Model Based on Factor Analysis and Similar-Day Approach for Short-Term Load Forecasting
Article
Full-text available
  • May 2018
Accurate power-load forecasting for the safe and stable operation of a power system is of great significance. However, the random non-stationary electric-load time series which is affected by many factors hinders the improvement of prediction accuracy. In light of this, this paper innovatively combines factor analysis and similar-day thinking into a prediction model for short-term load forecasting. After factor analysis, the latent factors that affect load essentially are extracted from an original 22 influence factors. Then, considering the contribution rate of history load data, partial auto correlation function (PACF) is employed to further analyse the impact effect. In addition, ant colony clustering (ACC) is adopted to excavate the similar days that have common factors with the forecast day. Finally, an extreme learning machine (ELM), whose input weights and bias threshold are optimized by a bat algorithm (BA), hereafter referred as BA-ELM, is established to predict the electric load. A simulation experience using data deriving from Yangquan City shows its effectiveness and applicability, and the result demonstrates that the hybrid model can meet the needs of short-term electric load prediction.
View
Short-Term Load Forecasting Using EMD-LSTM Neural Networks with a Xgboost Algorithm for Feature Importance Evaluation
Article
Full-text available
  • Aug 2017
Accurate load forecasting is an important issue for the reliable and efficient operation of a power system. This study presents a hybrid algorithm that combines similar days (SD) selection, empirical mode decomposition (EMD), and long short-term memory (LSTM) neural networks to construct a prediction model (i.e., SD-EMD-LSTM) for short-term load forecasting. The extreme gradient boosting-based weighted k-means algorithm is used to evaluate the similarity between the forecasting and historical days. The EMD method is employed to decompose the SD load to several intrinsic mode functions (IMFs) and residual. Separated LSTM neural networks were also employed to forecast each IMF and residual. Lastly, the forecasting values from each LSTM model were reconstructed. Numerical testing demonstrates that the SD-EMD-LSTM method can accurately forecast the electric load.
View
Improved Deep Belief Network for Short-Term Load Forecasting Considering Demand-Side Management
Article
  • Oct 2019
Demand-side management (DSM) increases the complexity of the forecasting environment, which makes traditional forecasting methods difficult to meet the firm's need for predictive accuracy. Since deep learning can comprehensively consider various factors to improve the prediction results, this paper improves the deep belief network from three aspects of input data, model and performance, and uses it to solve the short-term load forecasting problem in DSM. In the data optimization stage, the Hankel matrix is constructed to increase the input weight of DSM data, and the gray relational analysis is used to select strongly correlated data from other data. In the model optimization stage, the Gauss-Bernoulli restricted Boltzmann machine is used as the first restricted Boltzmann machine of the deep network to transform the continuity features of the input data into binomial distribution features. In the performance optimization stage, a pre-training method combining error constraint and unsupervised learning is proposed to provide good initial parameters, and the global fine-tuning of network parameters is realized based on genetic algorithm. Based on the actual data of Tianjin Power Grid in China, the experimental results show that the proposed method is superior to other methods.
View
Multivariate Quantile Regression for Short-Term Probabilistic Load Forecasting
Article
  • Jun 2019
Short-term probabilistic load forecasting is essential to power systems management and optimization of power flows across transmission networks. Developing forecasting tools capable of providing accurate predictions must comply with their practical implementation in short-term operations, mainly in terms of fast computing and high efficiency. Many data-driven forecasting solutions are often unnecessarily verbose, thus making their practical value limited. This problem occurs more frequently when multiple loads have to be predicted simultaneously, as in power transmission system analysis and optimization. In this paper, we propose a new cooperative forecasting system that refines probabilistic forecasts of individual loads online. The refining procedure is based on a multivariate quantile regression, which is dynamically applied to the individual forecasts as new observations become available. The proposal is validated on the load data published by ISO New England for eight regions, covering six States of the United States. The quality of probabilistic forecasts is assessed in terms of reliability and sharpness, comparing the results to three benchmarks. The proposed method outperforms the best benchmark by up to 6% w.r.t. the reduction in pinball loss.
View
Strategic Placements of PMUs for Power Network Observability Considering Redundancy Measurement
Article
  • Nov 2018
  • MEASUREMENT
This paper proposes a mixed integer nonlinear programming (MINLP) formulation with binary-valued variables for the optimal installation of the phasor measurement units (PMUs) in a power network. The MINLP framework has been solved by a two-phase branch-and-bound algorithm (BBA) for the optimal PMU placement (OPP) problem. The present work deals with the solution of the OPP problem without considering the radial buses including in the optimal solution. PMU is pre-assigned to each bus connected to a radial bus such that the radial buses are excluded from the list of potential PMU locations. The programming technique is used to determine the optimum number of PMUs and their locations to make the interconnected power network completely observable. The proposed BBA may provide multiple solutions at the lowest objective value, each time the optimizer routine restarts. Therefore, the obtained optimal solutions are being ranked regarding the measurement redundancy index. The two-phase branch-and-bound algorithm is applied to IEEE standard test systems as well as to New England 39- bus and a northern regional power grid 246- bus system. The proposed algorithm has been compared with the existing programming techniques in the recent literature. The simulation results obtained by the proposed BBA indicate the effectiveness and the efficiency to detect the desired target.
View
Priority Index Considering Temperature and Date Proximity for Selection of Similar Days in Knowledge-based Short Term Load Forecasting Method
Article
  • Dec 2017
  • ENERGY
Short term load forecasting (STLF) is one of the important issues in the energy management of power systems. Increasing the accuracy of STLF results leads to improving the energy system scheduling and decreasing the operating costs. Different methods have been proposed and applied in the STLF problem such as neural network, fuzzy system, regression-based and neuro-fuzzy methods. This paper investigates the knowledge-based method that has less computation time and memory compared with other methods. The accuracy of knowledge-based STLF method is improved by proposing a novel priority index for selection of similar days. In this index, temperature similarity and date proximity are simultaneously considered. In order to consider the effect of temperature in STLF more efficiently, the system is partitioned into the smaller regions and the STLF of the whole system is calculated by gathering the STLF of all regions. The proposed method is implemented on a sample real data, Iran's national power network, to show the advantages of the proposed method compared with Bayesian neural network and locally linear neuro-fuzzy methods in aspects of accuracy and computation time. It is shown that the proposed method decreases yearly mean absolute percentage error (MAPE), and generates more reliable load forecasting.
View