Approach For Load Forecasting Research Articles

An efficient approach for load forecasting in agricultural sector using machine learning

The shortage of coal is increasing the risk of a power crisis in many states of India. In summer, the supply of electricity is less than the demand, due to which the government has to cut power. Such cuts can be avoided if the power demand is estimated correctly. It is essential to have an accurate forecast of the load, if this is not the case, then the power has to be cut, and in some places, the production is reduced. The agricultural sector is also affected by the power cuts. This paper focused on load forecasting in agriculture using machine learning and ensemble learning approaches. We first identified the various factors influencing the power load in the agriculture sector and then assessed the demand for electricity in this area. The agricultural electricity consumption data is collected from JVVNL. The results show that good forecasting accuracies are attained by including exogenous features.

Improving the Efficiency of Deep Learning Models Using Supervised Approach for Load Forecasting of Electric Vehicles

Improving the Efficiency of Deep Learning Models Using Supervised Approach for Load Forecasting of Electric Vehicles

A New Short Term Electrical Load Forecasting by Type-2 Fuzzy Neural Networks

In this study, we present a new approach for load forecasting (LF) using a recurrent fuzzy neural network (RFNN) for Kermanshah City. Imagine if there is a need for electricity in a region in the coming years, we will have to build a power plant or reinforce transmission lines, so this will be resolved if accurate forecasts are made at the right time. Furthermore, suppose that by building distributed generation plants, and predicting future consumption, we can conclude that production will be more than consumption, so we will seek to export energy to other countries and make decisions on this. In this paper, a novel combination of neural networks (NNs) and type-2 fuzzy systems (T2FSs) is used for load forecasting. Adding feedback to the fuzzy neural network can also benefit from past moments. This feedback structure is called a recurrent fuzzy neural network. In this paper, Kermanshah urban electrical load data is used. The simulation results prove the efficiency of this method for forecasting the electrical load. We found that we can accurately predict the electrical load of the city for the next day with 98% accuracy. The accuracy index is the evaluation of mean absolute percentage error (MAPE). The main contributions are: (1) Introducing a new fuzzy neural network. (2) Improving and increasing the accuracy of forecasting using the proposed fuzzy neural network. (3) Taking data from a specific area (Kermanshah City) and forecasting the electrical load for that area. (4) The ability to enter new data without calculations from the beginning.

WITHDRAWN: WITHDRAWN: Hybrid deep learning approach for load forecasting in power systems

This article has been withdrawn at the request of the editor. The Publisher apologizes for any inconvenience this may cause. The full Elsevier Policy on Article Withdrawal can be found at https://www.elsevier.com/about/our-business/policies/article-withdrawal .

Online Ensemble Learning for Load Forecasting

Traditionally, load forecasting models are trained offline and generate predictions online. However, the pure batch learning approach fails to incorporate new load information available in real-time. Conversely, online learning allows for efficient adaptation of newly incoming information. This letter advocates a novel online ensemble learning approach for load forecasting by combining batch and online learning. While the individual batch models provide an appropriate forecast basis, the online ensemble combines their predictions and ensures adaptivity for online application. In that respect, we propose a modified Passive Aggressive Regression (PAR) model to implement the online ensemble forecasting. Case studies on a real-world load dataset show that the proposed method can improve the forecasting accuracy significantly compared to a pure batch learning approach.

GA-ANN hybrid approach for load forecasting

Precise Load forecasting holds an incredible valid prospective to foretell energy expected, which fulfils the need and supply balance. It is consequently vital that the power producing establishments have preceding information on future demand with incredible accuracy. There are different strategies available for load forecasting. This research paper surveys and investigates the usage of Genetic Algorithm based Neural Networks (GA-NN) method for the analysis of load demand and comparing the results Load and Price forecasting is an important factor because of the growing load demand, and this information is used to plan maintenance, capacity requirement and planning, which in turn improves efficiency and thus reduces costs of generation and transmission. Probabilistic forecasting is gaining great attention nowadays in forecasting of load and price of electricity.

A Data-Driven Bottom-Up Approach for Spatial and Temporal Electric Load Forecasting

With the rapid urbanization, electrical infrastructure spreads to raw areas without existing loads. How to achieve accurate long-term load forecasts based on land use plans is a realistic problem. On the other hand, load forecasting (LF) should be extended to high spatial resolutions to guide middle- or low-voltage planning and time domain to consider the impacts of distribution generations and diversified users on multi-period system demands. A data-driven bottom-up spatial and temporal LF approach is proposed in this paper to solve these challenges. Land plots are treated as basic LF resolution to describe available multi-attribute data in smart grids and modern cities. Kernel density estimation and adaptive k-means are adopted to aggregate typical load densities and profiles of different land use types. Stacked auto-encoders are utilized to forecast the unknown plot load quantities. The neighbor plot loads are summed up to obtain the estimated loads of larger areas based on clustered load profiles. Case studies demonstrate that the proposed LF is more applicable than benchmark methods both in accuracy and application potential. The estimated hierarchical spatial and temporal results are of great significance to guide load balancing, power system planning, and user integration in different voltage levels.

An Ensemble Forecasting Method for the Aggregated Load With Subprofiles

With the prevalence of smart meters, fine-grained subprofiles reveal more information about the aggregated load and further help improve the forecasting accuracy. Ensemble is an effective approach for load forecasting. It either generates multiple training datasets or applies multiple forecasting models to produce multiple forecasts. In this letter, a novel ensemble method is proposed to forecast the aggregated load with subprofiles where the multiple forecasts are produced by different groupings of subprofiles. Specifically, the subprofiles are first clustered into different groups and forecasting is conducted on the grouped load profiles individually. Thus, these forecasts can be summed to form the aggregated load forecast. In this way, different aggregated load forecasts can be obtained by varying the number of clusters. Finally, an optimal weighted ensemble approach is employed to combine these forecasts and provide the final forecasting result. Case studies are conducted on two open datasets and verify the effectiveness and superiority of the proposed method.

Lambda-Based Data Processing Architecture for Two-Level Load Forecasting in Residential Buildings

Building energy management systems (BEMS) have been intensively used to manage the electricity consumption of residential buildings more efficiently. However, the dynamic behavior of the occupants introduces uncertainty problems that affect the performance of the BEMS. To address this uncertainty problem, the BEMS may implement load forecasting as one of the BEMS modules. Load forecasting utilizes historical load data to compute model predictions for a specific time in the future. Recently, smart meters have been introduced to collect electricity consumption data. Smart meters not only capture aggregation data, but also individual data that is more frequently close to real-time. The processing of both smart meter data types for load forecasting can enhance the performance of the BEMS when confronted with uncertainty problems. The collection of smart meter data can be processed using a batch approach for short-term load forecasting, while the real-time smart meter data can be processed for very short-term load forecasting, which adjusts the short-term load forecasting to adapt to the dynamic behavior of the occupants. This approach requires different data processing techniques for aggregation and individual of smart meter data. In this paper, we propose Lambda-based data processing architecture to process the different types of smart meter data and implement the two-level load forecasting approach, which combines short-term and very short-term load forecasting techniques on top of our proposed data processing architecture. The proposed approach is expected to enhance the BEMS to address the uncertainty problem in order to process data in less time. Our experiment showed that the proposed approaches improved the accuracy by 7% compared to a typical BEMS with only one load forecasting technique, and had the lowest computation time when processing the smart meter data.

A two-stage short-term load forecasting approach using temperature daily profiles estimation

Electrical load forecasting plays an important role in the regular planning of power systems, in which load is influenced by several factors that must be analysed and identified prior to modelling in order to ensure better and instant load balancing between supply and demand. This paper proposes a two-stage approach for short-term electricity load forecasting. In the first stage, a set of day classes of load profiles are identified using K-means clustering algorithm alongside daily temperature estimation profiles. The proposed estimation method is particularly useful in case of lack of historical regular temperature data. While in the second stage, the stacked denoising autoencoders approach is used to build regression models able to forecast each day type independently. The obtained models are trained and evaluated using hourly electricity power data offered by Algeria’s National Electricity and Gas Company. Several models are investigated to substantiate the accuracy and effectiveness of the proposed approach.

A power load forecast approach based on spatial‐temporal clustering of load data

SummaryLoad forecast is very important for power system operation. Pursuing higher accuracy of load forecast is always the major target of this field. The existence of bad historical load data could badly affect the forecast accuracies of time series–based load forecast techniques. Within the multi‐node load data, the outliers of them are merely appeared instantaneously; the bad impact of the outliers could be decreased by making use of the spatial relativity of multi‐node load data. A novel load forecasting approach based on spatial‐temporal feature clustering is proposed in this paper. The temporal regular load pattern is extracted from the total load for an individual node. The spatial distribution characteristics of the individual incremental load have been categorized by the k‐medoids clustering algorithm. The MapReduce computing mode is used to manipulate multi‐node load data to raise computation efficiency. This new forecasting approach could reduce the influence of outliers and provide reliable and efficient load forecast with high accuracy. Testing in a real power system data set with up to 6.6% of outliers, the results of this approach show a lower forecasting error, about 6.2%, compared with three other time series–based methods (between 9.8% and 10.6%).

Short-term load forecasting in electrical power systems via trajectory tracking and error correcting approach

Rapid and accurate load forecasting is essential for renewable yet highly stochastic power (such as wind and solar power) to be massively utilized in practice. While there are many load forecasting methods reported in the literature, most of which, however, do not literally guarantee the convergence of forecasting error. This paper proposes a new error correcting approach for load forecasting in power systems by using trajectory tracking stability theory. In principle, the proposed method is not an autonomous but heuristic correcting approach to assess and improve the results of other existing models. This method is able to ensure the convergence of forecasting error in theory and is independent of system model, making it more feasible and cost-effective for forecasting performance improvement. Simulation experiments confirm the effectiveness of the proposed method for multiple existing models and forecasting horizons.

A novel economy reflecting short-term load forecasting approach

The global economic downturn in 2008 and 2009, which was spurred by the bankruptcy of Lehman Brothers, sharply reduced the demand for electricity load. Conventional load-forecasting approaches were unable to respond to sudden changes in the economy, because these approaches do not consider the effect of economic factors. Therefore, the over-prediction problem occurred. To overcome this problem, this paper proposes a novel, economy-reflecting, short-term load forecasting (STLF) approach based on theories of moving average (MA) line of stock index and machine learning. In this approach, the stock indices decision model is designed to reflect fluctuations in the Taiwan Stock Exchange Capitalization Weighted Stock Index (TAIEX) series, which is selected as an optimal input variable in support vector regression load forecasting model at an appropriate timing. The Taiwan island-wide hourly electricity load demands from 2008 to 2010 are used as the case study for performance benchmarking. Results show that the proposed approach with a 60-day MA of the TAIEX as economic learning pattern achieves good forecasting performance. It outperforms the conventional approach by 29.16% on average during economic downturn-affected days. Overall, the proposed approach successfully overcomes the over-prediction problems caused by the economic downturn. To the best of our knowledge, this paper is the first attempt to apply MA line theory of stock index on STLF.