Hourly Load Curve Research Articles

An adaptive hybrid fractal model for short-term load forecasting in power systems

This paper develops a hybrid short-term load forecasting (STLF) modeling with adaptive parametrization, based on composite linear fractal interpolation function (CLFIF), iterative learning (IL) and chimp optimization algorithm (ChOA). More precisely, after selecting similar days in power load data, an amendatory CLFIF model is constructed for an-hour-ahead prediction in terms of hourly load curves. Then, iterative learning based on ChOA optimizes the parameters of the amendatory CLFIF model with higher accuracy. Moreover, to confirm effectiveness of the CLFIF-IL-ChOA model, numerical examples are tested on the historical power load data from PJM and ENTSOE. The numerical results show that the proposed method can obtain higher accuracy, compared to some common methods about time series analysis.

Weekly hydropower scheduling of cascaded reservoirs with hourly power and capacity balances

A medium/long-term hydropower scheduling routinely assumes the energy produced in each week will be fully utilized in peak-shaving the power demands during every day in that week, ignoring the fluctuation of the hourly power loads. This work improves the traditional weekly hydropower scheduling by integrating the hourly power and capacity balances (HPCB), which has rarely been investigated, mainly due to great challenges imposed in solving the model. The HPCB are formulated into mixed integer linear constraints, involving the spare, maintenance, disabled, reserve and working capacities, as well as the order and levels of hydroplants in peak-shaving the hourly power load curve, all to be optimized. The formulation is then improved with successive strategies in updating water heads and peak-shaving hours, aiming to boost the solution efficiency by excluding more binary variables. The case studies involving 6 cascaded hydropower reservoirs in Lancang River reveal that the traditional weekly hydropower scheduling overestimates the benefits significantly, strongly suggesting the necessity to include the HPBC into a long/mid-term hydropower scheduling. The experiments also recommend an improved model formulation that achieves very consistent results on the top three prioritized objectives, while taking less than 0.1 s to solve the problem involving all six cascaded reservoirs, demonstrating a great prospect in solving problems in large scale.

Deep Learning for Short-Term Load Forecasting—Industrial Consumer Case Study

In the current trend of consumption, electricity consumption will become a very high cost for the end-users. Consumers acquire energy from suppliers who use short, medium, and long-term forecasts to place bids in the power market. This study offers a detailed analysis of relevant literature and proposes a deep learning methodology for forecasting industrial electric usage for the next 24 h. The hourly load curves forecasted are from a large furniture factory. The hourly data for one year is split into training (80%) and testing (20%). The algorithms use the previous two weeks of hourly consumption and exogenous variables as input in the deep neural networks. The best results prove that deep recurrent neural networks can retain long-term dependencies in high volatility time series. Gated recurrent units (GRU) obtained the lowest mean absolute percentage error of 4.82% for the testing period. The GRU improves the forecast by 6.23% compared to the second-best algorithm implemented, a combination of GRU and Long short-term memory (LSTM). From a practical perspective, deep learning methods can automate the forecasting processes and optimize the operation of power systems.

Optimal Dispatch Strategy of Virtual Power Plant for Day-Ahead Market Framework

Renewable energy sources prevail as a clean energy source and their penetration in the power sector is increasing day by day due to the growing concern for climate action. However, the intermittent nature of the renewable energy based-power generation questions the grid security, especially when the utilized source is solar radiation or wind flow. The intermittency of the renewable generation can be met by the integration of distributed energy resources. The virtual power plant (VPP) is a new concept which aggregates the capacities of various distributed energy resources, handles controllable and uncontrollable loads, integrates storage devices and empowers participation as an individual power plant in the electricity market. The VPP as an energy management system (EMS) should optimally dispatch the power to its consumers. This research work is proposed to analyze the optimal scheduling of generation in VPP for the day-ahead market framework using the beetle antenna search (BAS) algorithm under various scenarios. A case study is considered for this analysis in which the constituting energy resources include a photovoltaic solar panel (PV), micro-turbine (MT), wind turbine (WT), fuel cell (FC), battery energy storage system (BESS) and controllable loads. The real-time hourly load curves are considered in this work. Three different scenarios are considered for the optimal dispatch of generation in the VPP to analyze the performance of the proposed technique. The uncertainties of the solar irradiation and the wind speed are modeled using the beta distribution method and Weibull distribution method, respectively. The performance of the proposed method is compared with other evolutionary algorithms such as particle swarm optimization (PSO) and the genetic algorithm (GA). Among these above-mentioned algorithms, the proposed BAS algorithm shows the best scheduling with the minimum operating cost of generation.

Belarus Power Engineering System Modeling Taking into Account the Nuclear Power Plant Commissioning

The article presents the analysis of the specific features of modeling the operation of energy systems with a large share of nuclear power plants (NPP). The study of operating conditions and characteristics of different power units showed that a power engineering system with a large share of NPP and CHPP requires more detailed modeling of operating modes of generating equipment. Besides, with an increase in the share of installations using renewable energy sources, these requirements are becoming tougher. A review of the literature revealed that most often the curve of the load duration and its distribution between blocks are used for modeling energy systems. However, since this method does not reflect a chronological sequence, it can only be used if there are no difficulties with ensuring power balance. Along with this, when the share of CHP and nuclear power plants is high, to maintain a balance of power one must know the parameters and a set of powered equipment not only currently but, also, in the previous period. But this is impossible if a curve of load duration is used. For modeling, it is necessary to use an hourly load curve and to calculate the state of the energy system for each subsequent hour in chronological order. In the course of a comparative analysis of available computer programs, it was not possible to identify a suitable model among the existing ones. The article presents a mathematical model developed by the authors, which makes us possible to simulate the operation of a power engineering system with a large share of NPP and CHPP while maintaining the power balance for each hour of the forecast period. Verification of the proposed model showed good accuracy of the methods used.

Estimation of load curves for large-scale district heating networks

Decarbonisation and a transition towards sustainable energy systems in cities are key elements of the United Nations sustainability goal. Large-scale district heating networks sourced by excess heat or renewable energy allow to effectively transform building-related energy systems. This study proposes two different approaches for modelling load curves in large-scale district heating networks: 1) physics-based static energy balance model 2) data-driven regression model trained and adjusted on measured load curves. The load curves generated by application of these two approaches are compared with the actual load of an urban district heating network in Geneva, Switzerland. Both models allow to recreate the actual load curve of the district heating network, however with lower accuracy for higher time resolution in the case of the physics-based model. The physics-based static model can be used to simulate the demand and generate load curves of sufficient quality at monthly and daily resolution. For an hourly load curve, it is recommended to use the data-driven regression model if consumption data of the network is available.

SynCity: Using open data to create a synthetic city of hourly building energy estimates by integrating data-driven and physics-based methods

Cities officials are increasingly interested in understanding spatial and temporal energy patterns of the built environment to facilitate their city’s transition to a low-carbon future. In this paper, a new Augmented-Urban Building Energy Model (A-UBEM) is proposed that combines data-driven and physics-based simulation methods to produce synthetic hourly load curve estimates for every building within a city—similar to data an hourly smart meter would measure. By using only publicly available data, a generalizable two-step process is implemented—that other cities with similar available data can replicate—using New York City as a case study. Step (1) estimates the annual energy use for every building in the city using supervised machine learning algorithms. Step (2) extends these results and leverages physics-based simulation models through a convex optimization formulation that minimizes the squared difference between the aggregated building demand and the observed city-wide hourly electricity demand. Results from step (1) show that the Random Forest algorithm performs best with a mean log squared error of 0.293, while the convex optimization in step (2) results in a mean training error of 6.11% mean absolute percentage error (MAPE). To validate the stability of the produced load curves, Monte Carlo simulations are conducted, using random subsets of buildings from the city, which produce an out-of-sample error averaging 6.41% MAPE across each simulation. Particle swarm optimization is also explored—using the results from the Monte Carlo simulation—to assess if the model could be improved by relaxing certain constraints, but marginal error reductions are found, further proving the stability of the proposed model. Overall, A-UBEM is a first step towards creating highly granular urban-scale synthetic hourly load curves solely using open data. Such load curves are integral for planning sustainable cities and accelerating the adoption of low-carbon distributed energy resources (DERs) and district energy systems.

Assessment of multiple-based demand response actions for peak residential electricity reduction in Ghana

Demand-side management initiatives such as voluntary demand response provide significant energy savings in the residential sector, which is a major peak demand contributor. The potential of such savings remains unexplored in Ghanaian households due to insufficient electricity consumption data, lack of end-user behavior information and knowledge about the cost-effectiveness of such programs. This research combines 80 household survey information and energy use monitoring data of household appliances, to assess the residential demand response potential of Ghana. A bottom-up approach based on modified end-use model is used to develop aggregate hourly load curve. The estimated electricity consumption is categorized based on their degree of control to determine peak demand reduction potential for the period 2018-2050. The average daily peak load reduction ranged between 65-406 MW representing 2-14% for the considered scenarios by 2050. The results show appreciable economic viability for investment in demand response with net present value varying between 28-645 million US$. We find that price, energy security and environment signals influence end-users’ electricity use behavior. Authors observe that for energy and cost savings to be realized, utility providers and consumers need effective cooperation on information delivery and feedbacks, and consumers should be incentivized to balance the benefits.

Forecasting Daily Electric Load by Applying Artificial Neural Network with Fourier Transformation and Principal Component Analysis Technique

In this paper, we propose a hybrid forecasting model (HFM) for the short-term electric load forecasting using artificial neural network (ANN), discrete Fourier transformation (DFT) and principal component analysis (PCA) techniques in order to attain higher prediction accuracy. Firstly, we estimate Fourier coefficients by the DFT for predicting the next-day load curve with an ANN and obtain approximate load curves by applying the inverse discrete Fourier transformation. Approximate curves, together with other input variables, are given to the ANN to predict the next-day hourly load curves. Furthermore, we predict PCA scores to obtain approximate load curves in the first step, which are then given to the ANN again in the second step. Both DFT and PCA models use input variables such as calendrical and meteorological data as well as past electric loads. Applying those models for forecasting hourly electric load in the metropolitan area of Japan for January and May in 2018, we train our models using historical data since January 2008. The forecast results show that the HFM consisting of “ANN with DFT” and “ANN with PCA” predicts next-day hourly loads more accurately than the conventional three-layered ANN approach. Their corresponding mean average absolute errors show 2.7% for ANN with DFT, 2.6% for ANN with PCA and 3.0% for the conventional ANN approach. We also find that in May, when electric demand is smaller with smaller fluctuations, forecasting errors are much smaller than January for all the models. Thus, we can conclude that the HFM would contribute to attaining significantly higher forecasting accuracy.

Composite Reliability Evaluation of Load Demand Side Management and Dynamic Thermal Rating Systems

Electric power utilities across the globe are facing higher demand for electricity than ever before, while juggling to balance environmental conservation with transmission corridor expansions. Demand side management (DSM) and dynamic thermal rating systems (DTR) play an important role in alleviating some of the challenges faced by electric power utilities. In this paper, various DSM measures are explored and their interactions with the application of the DTR system in the transmission network are examined. The proposed modelling of DSM in this paper implements load shifting on load demand curves from the system, bus and load sector levels. The correlation effects of line ratings are considered in the DTR system modelling as the weather that influences line ratings is also correlated. The modelling of the line ratings was performed using the time series method, the auto regressive moving average (ARMA) model. Both the DSM and the DTR systems were implemented on the modified IEEE reliability test network. The modification was achieved by developing a load model starting from the perspective of the load sectors at each bus and a new collective hourly load curve for the system was obtained by combining the loads at all buses. Finally, the results in this paper elucidate the interaction of DSM and DTR systems.

Dissimilarities of Physical Meaning between Loss of Load indices based on Load Curve Types

Abstract This paper proposes a conversion function and a method to transform LOLED (days/year) using daily peak load curve into LOLEH (hours/year) using hourly load curve and describes a relationship between LOLED and LOLEH. The indices cannot only be just transformed arithmetically but also have different characteristics physically because of using their different load curves. The conversion function is formulated as variables of capacity and a forced outage rate of generator, hourly load daily load factor, daily peak load yearly load factor and etc. Therefore, the conversion function (γ = φ(.)) is not simple. In this study, therefore, the function is formulated as linear times of separated two functions. One is an exponential formed conversion function of daily load factor and the other is a formulated with an exponential typed conversion function of daily peak load yearly load factor. Furthermore, this paper presents an algorithm and a flow chart for transforming LOLED into LOLEH. The proposed conversion function is applied to the actual KPS (Korea Power System) in 2015. The exponent coefficients of the conversion functions are assessed using the proposed method. Finally, assessment errors using the conversion function for case study of the actual system are evaluated to certify as a first proposed method.

Multi‐objective optimization model for distribution network reconfiguration in the presence of distributed generations

Summary With increasing annual energy demand and significant growth in the use of distributed generation (DG), distribution companies are exploring ways to improve the technical specifications of the distribution networks such as power losses, customer interruption costs, and voltage stability and on the other hand to increase the penetration level of DGs. In this paper, a multi-objective optimization model is proposed for distribution network reconfiguration (DNR) to decrease the total costs of energy losses and energy not supplied (ENS) and increase the voltage stability and DG penetration level. Seasonal hourly load curves are applied, and best configuration is determined for each loading condition. Uncertainties regarding the load variations and volatile power production of wind turbine are modeled using a 2-point estimate method, and a benefit-cost analysis is used to assess the effects of depreciation cost in DNR problem. The proposed model is applied on 2 typical test systems.

두 종류의 부하곡선에 관한 공급지장시간기대치(LOLE)의 상호 변환관계성

This paper develops a conversion function and method transforming from daily peak load curve used LOLED [days/year] to hourly load curve used LOLEH[hours/year]and describes relationship between LOLED [days/year] and LOLEH [hours/year]. The indices can not only be transformed just arithmetically but also have different characteristics physically because of using their different load curves. The conversion function is formulated as variables of capacity and forced outage rate of generator, hourly load daily load factor and daily peak load yearly load factor, etc. Therefore, the conversion function (γ=φ(.)) cannot be simple. In this study, therefore, the function is formulated as linear times of separated two functions. One is an exponential formed conversion function of daily load factor. Another is formulated with an exponential typed conversion function of daily peak load yearly load factor. Futhermore, this paper presents algorithm and flow chart for transforming from LOLED[days/year] to LOLEH[hours/year]. The proposed conversion function is applied to sample system and actual KPS(Korea Power System) in 2015. The exponent coefficients of the conversion functions are assessed using proposed method. Finally, assessment errors using conversion function for case studies of sample system and actual system are evaluated to certify the firstly proposed method.

Characterization of the Hourly Load Curve in the Colombian Electricity Market

This paper analyzes the hourly load curve of the total electricity demand using the most detailed information available for the Colombian electricity market. Our findings are the following: first, regulated and non-regulated markets have very different load curves; for the regulated market, the load curve has pronounced peaks and valleys while for the non-regulated market the load curve is approximately flat. All economic sectors have a flat load curve, except those whose activity is related to an hourly schedule as agriculture, livestock, fishing or commerce. Second, load factor has changed from 0.70 to 0.82; this variation is explained for the corresponding growth of the load factor in the regulated market; the non-regulated market has an approximately constant load factor. And third, cluster analysis indicates that demand can be grouped in two classes of typical days: the first cluster correspond to Sundays and holidays and the second cluster to the other days.

The shape of future electricity demand: Exploring load curves in 2050s Germany and Britain

National demand for electricity follows a regular and predictable daily pattern. This pattern is set to change due to efficiency improvements, de-industrialisation and electrification of heat and transport. These changes are independent of renewable infeed and are not well understood: contemporary studies assume that electricity load curves will retain their current shape, scaling equally in all hours. Changes to this shape will profoundly affect the electricity industry: increasing the requirements for flexible and peaking capacity, and reducing asset utilisation and profitability.This paper explores the evolution of load curves to 2050 in Germany and Britain: two countries undergoing radically different energy transformations. It reviews recent developments in Europe's electricity demand, and introduces two models for synthesising future hourly load curves: eLOAD (electricity LOad curve ADjustment) and DESSTinEE (Demand for Energy Services, Supply and Transmission in EuropE). Both models are applied to a decarbonisation scenario for 2050, and consistently show peak loads increasing by about 23% points above the change in annual demand, to 103 GW in Germany and 92 GW in Britain. Sensitivities around electrification show that a million extra heat pumps or electric vehicles add up to 1.5 GW to peak demand.The structure and shape of the future load curves are analysed, and impacts on the national electricity systems are drawn.

Construction of the hourly load curves and detecting the annual peak load of future Syrian electric power demand using bottom-up approach

A comprehensive analysis has been carried out to calculate the hourly electricity demand and the load factors (LF) of the various consumption sectors, construct the load duration curves (LDC) and predict the annual peak load at final and secondary levels with the ultimate task of formulating the expansion plan of future generation system. For this purpose the actual hourly electricity demand (distributed by day types and seasons) of selected typical clients representing the consumption behaviour of household, service, transport and industry sectors in the base year have been collected and evaluated. Starting from these data and using the reference scenario results of future electricity demand projections – in term of annual demand and relative shares of consumption sectors, the future annual LDC’s and the total peak load for the next three decades have been constructed using the bottom-up approach MAED_EL. The results indicate that the current residential behaviour of Syrian power system will shift in the reference scenario more and more to the typical industry behaviour characterized by higher load factors. In the study period 1999–2030 the LF will increase from 0.64 to 0.71 and the peak load will grow annually at average rates of 5.2% in the reference scenario.

Security-Constrained Optimal Coordination of Generation and Transmission Maintenance Outage Scheduling

The proposed approach to the optimal coordination of midterm planning can be used by vertically integrated utilities as well as the ISOs in electricity markets. The approach coordinates the optimal maintenance scheduling of generating units and transmission lines, within the outage scheduling horizon of weeks to months, with the security-constrained unit commitment of available components. The proposed model provides hourly schedules for maintenance outages, generation unit commitment, and transmission flows based on hourly load curves. In addition, the approach calculates the optimal allocation of fuel and emission allowance for thermal units in a vertically integrated utility, and energy allocation in electricity markets. This paper discusses that a seamless coordination of midterm and short-term optimization schemes will effectively improve the security of electricity services and reduce the chance of blackouts in aging power systems. The proposed approach highlights a trade-off between calculation speed and accuracy for large-scale implementations. A six-bus system is used for detailed analyses, and further results presented for the IEEE-118 system exhibit the effectiveness of the proposed midterm planning approach.

Transformer real-time reliability model based on operating conditions

Operational reliability evaluation theory reflects real-time reliability level of power system. The component failure rate varies operating conditions. The impact of real-time operating conditions such as ambient temperature and transformer MVA (megavolt-ampere) loading on transformer insulation life is studied in this paper. The formula of transformer failure rate based on the winding hottest-spot temperature (HST) is given. Thus the real-time reliability model of transformer based on operating conditions is presented. The work is illustrated using the 1979 IEEE Reliability Test System. The changes of operating conditions are simulated by using hourly load curve and temperature curve, so the curves of real-time reliability indices are obtained by using operational reliability evaluation.

Estimating Hourly Electric Load with Generalized Least Square Procedures

Although electricity demand receives much attention in the empiri- cal literature (see Taylor (1975) and EPRI (1982b) for excellent surveys on the topic), hourly load demand analysis has only recently begun. Notable contributions are a series of studies sponsored by the Electric Power Research Institute (EPRI (1977. 1979a, 1979b, 1981a, 19816. 1982a) and Platt (1983)). These studies estimate load curve models for regions of the United States. Unfortunately, from a utility perspective, the empirical results presented in these studies are not directly applicable. Further, because the data used in these studies are not generally available at the geographic level of a utility service area, applying their methodology is problematic. This paper presents a practical method for an electric utility to produce an hourly load curve model similar in overall framework to these studies. Our procedure is innovative in that it produces statistically efficient estimates, which the above papers do not. We also demonstrate a method that uses supplementary forecasts to enhance the forecasting performance of the hourly load model.