Daily Peak Load Research Articles

Daily peak demand forecasting using Pelican Algorithm optimised Support Vector Machine (POA-SVM)

The knowledge of daily peak load consumption is crucial for energy planning, energy management, and resource allocation, as it is an essential element of supply-side management. This knowledge is obtained from accurate predictions of which traditional methods are falling short due to constantly changing demand. Hence, there is a need for Machine Learning (ML) models that are more efficient in hybridised or enhanced forms. This paper presents the use of a novel Pelican Algorithm optimised Support Vector Machine (POA-SVM), a hybridised ML Algorithm, to predict peak load and its corresponding peak hour for proper planning. It evaluates the performance of four Support Vector Machine (SVM) models: standard SVM, SVM optimised with Bayesian Optimisation (SVMB), SVM optimised with Particle Swarm Optimization (PSO-SVM) and POA-SVM on both raw and normalised datasets. Its analysis includes a comparison of performance metrics such as Root Mean Squared Error (RMSE), Mean Squared Error (MSE), Mean Absolute Error (MAE), and R-squared (R²) to determine the models' accuracy and goodness of fit. Furthermore, we experiment with K-fold cross-validation and the hold-out validation method, finding that K-fold cross-validation yields better performance metrics for all models. Notably, the POA-SVM model shows superior performance across all metrics, particularly on the raw dataset, making it a robust choice for forecasting peak demand and its corresponding hour of the day.

Kajian Beban Harian pada Jaringan Distribusi Penyulang Jelutung

A distribution network is a conveyor network that connects the load center substation with the customer. The distribution network functions to distribute electrical energy to customers according to the power and voltage needed. The distribution network line from the Aurduri GI Transformer Bay goes to the incoming to be lowered voltage from 150kv to 20kv. After the voltage is lowered, it is directly channeled to the Jelutung Feeder to find out the daily load on the distribution network. The population in this study is "Study of Daily Burden on the Distribution Network of Jelutung Feeders". The highest daily load is the electrical load that occurs at the highest time in 24 hours. This can be affected by household, industrial, commercial activities and weather conditions. Daily peak loads occur at certain times such as the afternoon. The lowest daily load usually occurs in the morning around 06.00 to 07.30. Average lowest current (I) R,S,T 92,275A and lowest power (P) 3.29Mw. While the average highest current is 212.7A and the highest power is 7.54Mw. The highest current in September was on the 15th with a power of 8.0 mw at 18.30.

Holiday Peak Load Forecasting Using Grammatical Evolution-Based Fuzzy Regression Approach

Peak load forecasting plays an important role in electric utilities. However, the daily peak load forecasting problem, especially for holidays, is fuzzy and highly nonlinear. In order to address the nonlinearity and fuzziness of the holiday load behaviors, a grammatical evolution-based fuzzy regression approach is proposed in this paper. The proposed hybrid approach is based on the theorem that fuzzy polynomial regression can model all fuzzy functions. It employs the rules of the grammatical evolution to generate fuzzy nonlinear structures in polynomial form. Then, a two-stage fuzzy regression approach is used to determine the coefficients and calculate the fitness of the fuzzy functions. An artificial bee colony algorithm is used as the evolution system to update the elements of the grammatical evolution system. The process is repeated until a fuzzy model that best fits the load data is found. After that, the developed fuzzy nonlinear model is applied to forecast holiday peak load. Considering that different holidays possess different load patterns, a separate forecaster model is built for each holiday. Test results on real load data show that an averaged absolute percent error less than 2% can be achieved, which significantly outperforms existing methods involved in the comparison.

Multi-source and Multi-level Coordination of Energy Internet under V2G based on Particle Swarm Optimization Algorithm

In order to effectively improve the excessive load of microgrid during peak hours of urban electricity consumption, a multi-source and multi-level coordination method of energy Internet under V2G based on particle swarm optimization algorithm was proposed. First, a mathematical model for V2G energy integration in microgrids was developed and a scheduling concept based on a particle-by-particle optimization algorithm was used. Second, an improved PSO algorithm is proposed and experimentally validated, and the experimental results are compared with previous particle swarm optimization algorithms. Experiments have shown that as the number of iterations increases, the value of the objective function decreases and the optimal solution can be obtained until the maximum number of iterations is reached. The iteration speed and power processing cost of the improved PSO algorithm are better than before.The original load curve is the load trough period from 23:00 to 6:00, and two load peaks occur from 12:00 to 14:00 and 19:00 to 22:00. The V2G technology basically realizes the coordinated control of microgrid electric energy and achieves the effect of peaking and valley filling. The improved algorithm has obvious improvement compared with the original power grid state. Conclusion: The application of EV V2G technology can smooth the daily load curve of power grid and coordinate the electric energy of micro-grid to achieve ``peak cutting and valley filling'', and the effect of this algorithm is more outstanding than the previous algorithm. Finally, the future development direction and suggestions of V2G technology are put forward.The power grid with V2G discharge depth limit has the ability to basically reduce and eliminate the daily peak load, so the technology has broad research space and development prospects.

Design and simulation of a building-based off-grid photovoltaic microgrid using PVsyst

In the absence of a main or central grid, an off-grid renewable energy-based system could be a viable solution to address the electricity demand of a particular region by utilizing the available renewable energy sources (RES) of that area. This also leads society to a step toward sustainable energy development. A planned RES-integrated microgrid system not only fulfills the energy crisis and reduces electricity costs but also plays a significant role in the conservation of fossil fuels. Therefore, the design and simulation of RES integrated systems are indispensable as this type of analysis determines the vital parameters of the system e.g., the system performances including electricity output, energy conversion efficiency, yearly energy potential, and the system losses before being involved in practical implementation. Further, the cell temperature of the Photovoltaic (PV) panel based on regional weather conditions, the system life cycle based on the uncertainty factor of weather, total system installation cost with a full account of the annual savings and payback etc. can also be determined by designing and simulating such systems. This paper presents a design of a 40 kW off-grid photovoltaic (PV) microgrid system according to the load requirements at the Department of Electronics and Communication Engineering (ECE), Tezpur University, India using PVsyst software. The proposed design has been created to meet the daily peak load demand of 37 kW of the department. The energy available through the generating units of the proposed designed PV system is 45.46 MWh/Year and the system performance ratio in terms of efficiency over the year is 0.814. The overall performance analysis and simulation results would be helpful in determining the efficiency and viability of the proposed PV system while implementing it practically in the near future.

Impacts of urban air temperature and humidity on building cooling and heating energy demand in 15 cities of eastern China

Urban climates could significantly impact building energy consumption. This study aimed to investigate the impact of urban-rural disparities in temperature and humidity on building cooling and heating demand in multiple cities. Fixed stations were established using temperature/humidity loggers deployed at 17 pairs of urban and rural sites in 15 cities in eastern China. The study analyzed the differences in air temperature (ΔTa), relative humidity (ΔRH), and dew-point temperature (ΔTdew) between the urban and rural sites. These climatic data were then used as inputs for the energy simulation of a residential building. The following results were revealed: (1) ΔTa was consistently higher at night than during the day. In comparison with the rural reference sites, the RH values at all urban sites were lower; however, some urban sites displayed lower Tdew, while others exhibited higher Tdew. (2) Regarding annual total energy demands, urban buildings exhibited variable changes in energy components compared to rural buildings. Sensible cooling demands increased by 10 %–36 %, while heating demands decreased by 9 %–30 %. Latent cooling demands ranged from a decrease of 12 % to an increase of 17 %. As a result, overall changes in annual total energy demands (cooling + heating) were relatively small, ranging from −2 % to +7 %. (3) Regarding daily total energy demands, a 1 °C increase in ΔTa led to a 12.1 % increase in sensible cooling demand and an 8 % decrease in heating demand. At the same time, a 1 °C variation in ΔTdew corresponded to a 14.6 % change in latent cooling demand. (4) Regarding the daily peak load, modifications in local temperature and humidity in urban areas had a more pronounced impact on daily peak heating loads than on daily peak cooling loads. This study could enhance our understanding of how urban local climates affect building cooling and heating energy demands.

Application of Seasonal Trend Decomposition using Loess and Long Short-Term Memory in Peak Load Forecasting Model in Tien Giang

Daily peak load forecasting is critical for energy providers to meet the loads of grid-connected consumers. This study proposed a Seasonal Trend decomposition using Loess combined with Long Short-Term Memory (STL-LTSM) method and compared its performance on peak forecasting of electrical energy demand with Convolutional Neural Network and LSTM (CNN-LSTM), Wavenet, and the classic approaches Artificial Neural Network (ANN) and LSTM. The study evaluated the models using demand data from the power system in Tien Giang province, Vietnam, from 2020 to 2022, considering historical demand, holidays, and weather variables as input characteristics. The results showed that the proposed STL-LSTM model can predict future demand with lower Base Mean Square Error (RMSE) and Mean Absolute Percentage Error (MAPE). Therefore, the proposed method can help energy suppliers make smart decisions and plan for future demand.

Intensive and extensive margins of the peak load: Measuring adaptation with mixed frequency panel data

In this work we investigate the response of daily electricity peak load to daily maximum temperatures across states in Europe and India. We propose a method that decomposes short- from medium/long-run effects, retains the high frequency nature of the load-weather covariation and treats economic growth as a modulating factor. By simultaneously exploiting variation in unexpected daily weather anomalies and decade-long climatic changes in each location we decompose transitory - intensive margin - adjustments from permanent - extensive margin - adjustments. We find that the shocks over the long-run differ substantially from the short-run dynamics. Furthermore, we find evidence that per capita income modulates the adjustments over the short- and long-run. We project that in response to climate change around 2050 the peak load may increase by up to 20%-30% in Southern Europe and in several states in India, depending on the degree of warming and the evolution of socio-economic conditions. Even with a limited scope to two world regions, we identify that the structure of the economy and differences in future income growth matter in shaping the adaptation to climate change. Our decomposition allows to identify how future weather anomalies can further amplify the relative increase associated to the shift in the climate norm. Assuming that the interannual variability of maximum temperatures follows the distribution observed in the past, we find a doubling of the impacts of climate change during the summer in Europe. Uncertainty around the distribution of future weather anomalies may lead to further unexpected peak load amplifications. Our results have important policy implications for power systems’ generation capacity, transmission and storage, as we show that the challenges to accommodate the peak load in days with extreme temperatures may substantially increase already around mid-century.

Testing and evaluation of a smart controller for reducing peak loads and return temperatures in district heating networks

A smart demand response control system aiming towards real-time operational optimisation of district heating (DH) network temperature levels, both in the return and supply pipes, has been developed in the TEMPO project. The return temperature is mainly dependent on the demand side. The controller optimises its value through control of the customers’ heat load. The network supply temperature, however, is directly controllable on the production side. The capabilities of supply temperature control are twofold. On the one hand, lowering the network supply temperature as close as possible to the limits determined by customer thermal demands. On the other hand, activating the intrinsic thermal capacity of the piping to temporarily store heat and thereby shifting the heat load in time. This provides additional energy flexibility potential on top of building demand response.In this study, the two features of the smart control system have been tested in a part of the DH network of Brescia (Italy). A cloud-based platform is used to collect real-time data from various sources and to communicate control signals calculated by the smart controller. The article presents the results of the tests and an evaluation of the controller performance. The analysis indicates that daily flow-weighted average return temperature reductions of almost 1 K on average can be achieved, and up to 15 K instantaneously. Using supply temperature control, the daily peak load energy supply could be reduced by 262 kWh (34%) on average, by shifting the heat load.

RAID: Robust and Interpretable Daily Peak Load Forecasting via Multiple Deep Neural Networks and Shapley Values

Accurate daily peak load forecasting (DPLF) is crucial for informed decision-making in energy management. Deep neural networks (DNNs) are particularly apt for DPLF because they can analyze multiple factors, such as timestamps, weather conditions, and historical electric loads. Interpretability of machine learning models is essential for ensuring stakeholders understand and trust the decision-making process. We proposed the RAID (robust and interpretable DPLF) model, which enhances DPLF accuracy by recognizing daily peak load patterns and building separate DNN models for each day of the week. This approach was accessible for energy providers with limited computational resources, as the DNN models could be configured without a graphics processing unit (GPU). We utilized scikit-learn’s MLPRegressor for streamlined implementation, Optuna for hyperparameter optimization, and the Shapley additive explanations (SHAP) method to ensure interpretability. Applied to a dataset from two commercial office buildings in Richland, Washington, RAID outperformed existing methods like recurrent neural networks, Cubist, and HYTREM, achieving the lowest mean absolute percentage error values: 14.67% for Building 1 and 12.74% for Building 2. The kernel SHAP method revealed the influence of the previous day’s peak load and temperature-related variables on the prediction. The RAID model substantially improved energy management through enhanced DPLF accuracy, outperforming competing methods, providing a GPU-free configuration, and ensuring interpretable decision-making, with the potential to influence energy providers’ choices and promote overall energy system sustainability.

Incorporating direct load control demand response into active distribution system planning

Nowadays, moving toward new technologies and smart grids is a successful way of overcoming distribution system challenges. Smart grids facilitate demand-side management (DSM) which affects the expansion of network resources. In this paper, distribution system planning (DSP) considering demand response (DR), as an effective DSM tool, is presented. The proposed model is based on direct load control (DLC) DR in which a utility offers financial incentives to the customers to control their consumption through installing smart meters and switches. The utility deploys the DR program to reduce network daily peak load, which is formulated as a linear programming model, where uncertainties associated with the DR program are captured by the spherical simplex unscented transformation method. Although the DR is implemented on the basis of hourly data, integrating its high-resolution results into the DSP can cause intractability. To mitigate this issue, scenario definition and scenario reduction models using the K-means clustering algorithm are proposed. Besides, feeder reinforcement as well as renewable and conventional distributed generation (DG) installation are considered as the other network expansion alternatives. The proposed method is applied to the modified IEEE 33 Bus test system and effects of the DR program and DG installation on DSP are investigated. The results reveals that the implementation of the DLC program in the DSP either reduces or postpones feeder reinforcement actions thereby resulting in lower overall DSP costs. Furthermore, DG installation can significantly reduce operating costs, particularly purchased energy from the main grid and energy loss. Eventually, the most cost-efficient solution is obtained by incorporating both DLC and DG into the DSP.

Energy consumption calculation of campus buildings based on the law of personnel flow

AbstractFor a certain functional building group, the coupling relationship between the load of a single unit and entire building groups lies in the spatial and temporal patterns personnel distribution. Using the DeST (Designer's Simulation Toolkit) method, the personnel flow trajectory in a university campus in Shanghai was analyzed and the established energy consumption calculation model based on personnel flow trajectory of university buildings. It is proposed that personnel mobility should be included in the application of DeST to predict the total load of building groups. The results show that after the people flow trajectory was included, the total annual accumulated load decreased from 27.01 to 19.39 GWh, reducing by 28.20%; the hourly load fluctuation became more stable, the daily peak load decreased by 22.57%, and the daily peak valley difference decreased by 43.74%. The results suggest that the prediction for the air conditioning load of building complexes would be better when the trajectory of human flow is considered, minimizing unnecessary system scale and investments. Under the “double carbon” policy, the personnel flow trajectory model can effectively reduce waste in air conditioning energy consumption in campus buildings.

A Hybrid Model of VMD-EMD-FFT, Similar Days Selection Method, Stepwise Regression, and Artificial Neural Network for Daily Electricity Peak Load Forecasting

Daily electricity peak load forecasting is important for electricity generation capacity planning. Accurate forecasting leads to saving on excessive electricity generating capacity, while maintaining the stability of the power system. The main challenging tasks in this research field include improving forecasting accuracy and reducing computational time. This paper proposes a hybrid model involving variational mode decomposition (VMD), empirical mode decomposition (EMD), fast Fourier transform (FFT), stepwise regression, similar days selection (SD) method, and artificial neural network (ANN) for daily electricity peak load forecasting. Stepwise regression and similar days selection method are used for input variable selection. VMD and FFT are applied for data decomposition and seasonality capturing, while EMD is employed for determining an appropriate decomposition level for VMD. The hybrid model is constructed to effectively forecast special holidays, which have different patterns from other normal weekdays and weekends. The performance of the hybrid model is tested with real electricity peak load data provided by the Electricity Generating Authority of Thailand, the leading power utility state enterprise under the Ministry of Energy. Experimental results show that the hybrid model gives the best performance while saving computation time by solving the problems in input variable selection, data decomposition, and imbalance data of normal and special days in the training process.

Week-ahead Daily Peak Load Forecasting Using Hybrid Convolutional Neural Network

Peak load forecasting (PLF) is essential for the operation of large-scale power systems, such as security-constrained economic dispatch and rolling unit commitment. PLF is critical for peak load shaving and demand side management in distribution systems, too. Results obtained by PLF can assist market participants to develop their bidding schemes in power markets. This work presents a novel method, based on a hybrid convolutional neural network (CNN) that is cascaded with a fully-connected network, to explore week-ahead daily PLF. The proposed method uses a systematic grid search along with Adaptive Moment Estimation (Adam) optimizer to design the hybrid model. The grid search tunes the network structure and hyperparameters (such as kernel size) of the hybrid CNN while Adam optimizer adjusts the synaptic weights and parameters (e.g., values of a kernel). Realistic daily peak load data and meteorological (temperature) data in Taiwan are investigated. Simulation results obtained by the proposed hybrid deep learning model are better than those obtained by the traditional multi-layer neural network, recurrent neural network, vector auto-regressive moving average model and support vector machine.

Technical, Economic and Environmental Comparison of Three Different Grid-Connected PV Tracking Systems Power Plant Under Kurdistan Region/Iraq Climate Condition

In this paper, a 1 MW grid-connected PV system was performed and simulated numerically in Zakho city, using hourly meteorological data for three systems; fixed, single-axis, and dual-axis tracking systems for solar module. The analysis of this study is based on technical, economic, and environmental feasibility. The analysis based on the actual 3 PV solar panels using different tracking systems which installed on the roof of the Engineering college of Zakho University. The evaluation findings suggest that PV technology is quite promising in this location, with annual yield factors of (1416 kWh/kW), (1694 kWh/kW), and (1902 kWh/kW) for the three systems, respectively. Furthermore, the proposed system's capacity factors are 16.2 percent, 19.3 percent, and 21.70 percent. The economic growth of a 1MW grid-connected photovoltaic (PV) system adjusted for meeting the daily peak load in Zakho is analyzed and compared based on the cost of electricity (COE), net present value (NPV), payback period (PBP), and the energy payback time (EPBT) for three systems in the present work. The COE for the three proposed systems, fixed, 1st axis, and dual axes solar tracking systems, was 0.0826 USD/kWh, 0.0489 USD/kWh, and 0.0441 USD/kWh, respectively, which indicated the tracking system is economically feasible. The findings indicate a favorable trend, implying that large-scale photovoltaics of dual-axis systems might be a feasible option for addressing future power needs.

Regression-Based Methods for Daily Peak Load Forecasting in South Korea

This study examines the daily peak load forecasting problem in South Korea. This problem has become increasingly important due to the continually changing energy environment. As such, it has been studied by many researchers over the decades. South Korea is geographically located such that it experiences four distinct seasons. Seasonal changes are among the main factors affecting electricity demand. In addition, much of the electricity consumption in a strong manufacturing country like South Korea is driven by industry rather than by residential customers. In order to forecast daily peak loads of South Korea, in this study we proposed multiple linear regression-based methods where several season-specific regression models (i.e., summer, winter, and all-season models) were included. The most appropriate model among the three models was selected considering the characteristics of the electricity demand, and was then applied to daily forecasting. The performance of the proposed methods were evaluated through computational experiments. Forecasts obtained by the proposed methods were compared with those obtained by existing forecasting methods, including a machine learning method. The results showed that the proposed methods had mean absolute percentage errors around 1.95% and outperformed all benchmarks.

Week‐ahead daily peak load forecasting using genetic algorithm‐based hybrid convolutional neural network

Daily peak load forecasting is crucial for the operation of bulk power systems, including economic dispatch and unit commitment. It is also essential for peak load shaving and load management in distribution systems. In power markets, peak load forecasting helps participants develop bidding strategies. This paper proposes a new method, using a hybrid convolutional neural network (CNN) that is cascaded with a fully-connected network, for making week-ahead daily peak load forecasts. The proposed method uses three loops to obtain the optimal CNN: The outer loop performs crossover/mutation operations and tournament selection to produce chromosomes to optimize the network topology and hyperparameters (such as kernel size) of the hybrid CNN by genetic algorithms; the middle loop deals with the order of chromosomes; the inner loop optimizes the synaptic weights and parameters (e.g. values of a kernel) using Adam optimizer. Daily peak load data and corresponding meteorological data for Taiwan are explored. Simulation results show that the proposed method outperforms the traditional CNN, multi-layer neural network, recurrent neural network, support vector regression and vector autoregressive moving average model.

LSTM enhanced by dual-attention-based encoder-decoder for daily peak load forecasting

Daily peak load forecasting is a challenging problem in the filed of electric power load forecasting. Since the nonlinear and dynamic of influence factors and their sequential dependencies are significant for modeling daily peak load, a prediction model based on long short-term memory (LSTM) enhanced by dual-attention-based encoder-decoder is presented. Functioned as the specific encoder and decoder, LSTM is utilized to participate in the nonlinear dynamic temporal modeling. The encoder-decoder is used for information utilization of both the influence factors and daily peak load. Moreover, a dual-attention mechanism, which is inserted into the encoder-decoder, is designed to take into account the effects of different influence factors and time nodes on the daily peak load simultaneously. It is benefit for the above mechanism design to analyze the characteristics of daily peak load precisely and to achieve more accurate prediction results. Comprehensive experiments are performed based on a real set of one provincial capital city in eastern China. The case study shows that the proposed methodology provides the most accurate results with an average MAPE 2.07%, an average RMSE 133 MW and an average MAE 326.6 MW.

Interpretable Short-Term Electrical Load Forecasting Scheme Using Cubist

Daily peak load forecasting (DPLF) and total daily load forecasting (TDLF) are essential for optimal power system operation from one day to one week later. This study develops a Cubist-based incremental learning model to perform accurate and interpretable DPLF and TDLF. To this end, we employ time-series cross-validation to effectively reflect recent electrical load trends and patterns when constructing the model. We also analyze variable importance to identify the most crucial factors in the Cubist model. In the experiments, we used two publicly available building datasets and three educational building cluster datasets. The results showed that the proposed model yielded averages of 7.77 and 10.06 in mean absolute percentage error and coefficient of variation of the root mean square error, respectively. We also confirmed that temperature and holiday information are significant external factors, and electrical loads one day and one week ago are significant internal factors.

Membership Function Optimization of Fuzzy Inference System Using Cuckoo Search Algorithm for Peak Load Forecasting in National Holiday

This study aims to analysis peak load prediction of Indonesian national holidays for Jawa-Bali electricity system. Forecasting applied using the Fuzzy Logic System (FLS) method combined with the Cuckoo Search Algorithm (CSA). CSA is used to determine the optimal membership function in fuzzy logic. Cuckoo search algorithm has a very good performance in terms of optimization. This method is applied for short-term load estimates on holidays/special days on the Jawa-Bali electricity system, Indonesia. The study used data from daily peak loads during Indonesian national holidays in 2014 on the Jawa-Bali electricity system. The data analyzed is the daily peak load documentation data for 4 days before national holidays and during national holidays in 2014. Testing the simulation results, it was found that the Fuzzy Logic System - Cuckoo Search Algorithm (FLS-CSA) method gives good forecasting results, this is evidenced by using the mean absolute percentage error (MAPE). Forecasting results using the Cuckoo Search Algorithm (CSA) optimization method on fuzzy logic membership functions for peak loads on national holidays on the Java-Bali 500kV electrical system give satisfactory results with an average forecasting error of 1.511314562%.