Development of feed-forward neural network models for gas short-term load forecasting

Brown, R H; Matin, I; Feng, X

Title: Development of feed-forward neural network models for gas short-term load forecasting

Full Record
Other Related Research

Abstract

The development of feed-forward artificial neural network based models to predict gas consumption on a daily basis is the subject of this paper. The discussion concerns an iterative process based on network sensitivities and intuition regarding selection of proper input factors. The method is applied in forecasting gas consumption for two regions in the State of Wisconsin, namely, a portion of metropolitan Milwaukee and a region near Fond du Lac. The investigation includes a study of the effects of various network sizes and training algorithms given a limited availability of relevant historical data. The effects of using multiple sources of weather data are also investigated for regions without a centrally located weather recording station. The training results indicate that feed-forward artificial neural network based models reduce the residual predicted consumption root mean squared errors by more than half when compared to models based on linear regression using identical input factors. The inclusion of weather data from multiple sources leads to further error reduction. The models were then used for load forecasting starting in Summer 1994. The static models performed well from the Summer 1993 until the severe cold weather of mid-January 1994. This poor performance was scrutinized; investigations point tomore »« less

Authors:

Brown, R H; Matin, I; Feng, X ^[1]

Marquette Univ., Milwaukee, WI (United States)

Publication Date:: Sat Dec 31 00:00:00 EST 1994

Research Org.:: USDOE Pittsburgh Energy Technology Center (PETC), PA (United States); Oregon State Univ., Corvallis, OR (United States). Dept. of Computer Science; Naval Research Lab., Washington, DC (United States); Electric Power Research Inst. (EPRI), Palo Alto, CA (United States); Bureau of Mines, Washington, DC (United States)

OSTI Identifier:: 81604

Report Number(s):: CONF-9410335-
ON: DE95011702; TRN: 95:004731-0019

Resource Type:: Conference

Resource Relation:: Conference: Adaptive control systems technology symposium, Pittsburgh, PA (United States), 24-25 Oct 1994; Other Information: PBD: [1994]; Related Information: Is Part Of High-tech controls for energy and environment. Proceedings; Biondo, S.J.; Drummond, C.J. [eds.] [USDOE Pittsburgh Energy Technology Center, PA (United States)]; PB: 287 p.

Country of Publication:: United States

Language:: English

Subject:: 03 NATURAL GAS; 99 MATHEMATICS, COMPUTERS, INFORMATION SCIENCE, MANAGEMENT, LAW, MISCELLANEOUS; NEURAL NETWORKS; USES; WISCONSIN; ENERGY CONSUMPTION; NATURAL GAS; CONSUMPTION RATES; ALGORITHMS; SUPPLY AND DEMAND; WEATHER; ARTIFICIAL INTELLIGENCE; FORECASTING

Citation Formats


                    Brown, R H, Matin, I, and Feng, X. Development of feed-forward neural network models for gas short-term load forecasting.  United States: N. p., 1994. 
        Web.

Copy to clipboard


                    Brown, R H, Matin, I, & Feng, X. Development of feed-forward neural network models for gas short-term load forecasting.  United States.

Copy to clipboard


                    Brown, R H, Matin, I, and Feng, X. 1994.  
        "Development of feed-forward neural network models for gas short-term load forecasting".  United States.

Copy to clipboard


                    
@article{osti_81604,

  title        = {Development of feed-forward neural network models for gas short-term load forecasting},

  author       = {Brown, R H and Matin, I and Feng, X},

  abstractNote = {The development of feed-forward artificial neural network based models to predict gas consumption on a daily basis is the subject of this paper. The discussion concerns an iterative process based on network sensitivities and intuition regarding selection of proper input factors. The method is applied in forecasting gas consumption for two regions in the State of Wisconsin, namely, a portion of metropolitan Milwaukee and a region near Fond du Lac. The investigation includes a study of the effects of various network sizes and training algorithms given a limited availability of relevant historical data. The effects of using multiple sources of weather data are also investigated for regions without a centrally located weather recording station. The training results indicate that feed-forward artificial neural network based models reduce the residual predicted consumption root mean squared errors by more than half when compared to models based on linear regression using identical input factors. The inclusion of weather data from multiple sources leads to further error reduction. The models were then used for load forecasting starting in Summer 1994. The static models performed well from the Summer 1993 until the severe cold weather of mid-January 1994. This poor performance was scrutinized; investigations point to probable causes that include severe weather patterns non-existent in the training data and the excitation of consumer behavioral modes either not evident in the training data or not well represented by time-of-the-year input factor proxies. The search for a solution lead to dynamic models that were implemented by updating the neural networks with new data as it became available. These models performed significantly better but still poorly during and after the severe weather of January 1994.},

  doi          = {},

  url          = {https://www.osti.gov/biblio/81604},
  journal      = {},
number       = ,

  volume       = ,

  place        = {United States},

  year         = {Sat Dec 31 00:00:00 EST 1994},

  month        = {Sat Dec 31 00:00:00 EST 1994}

}

Copy to clipboard

Conference:

Other availability

Please see Document Availability for additional information on obtaining the full-text document. Library patrons may search WorldCat to identify libraries that hold this conference proceeding.

Save / Share:

Export Metadata

Save to My Library

Similar records in OSTI.GOV collections:

Hourly load forecasting using artificial neural networks. Final report

Technical Report Khotanzad, A

An artificial neural network short-term load forecaster (ANNSTLF) and an artificial neural network (ANN) based temperature forecaster have been developed by Southern Methodist University under contracts RP2473-44 and RP3573-4. ANNSTLF can produce hourly load forecasts for one to 168 hours ahead (one to seven days ahead) with errors ranging from 2 to 4% depending on utility size and characteristics. Implementation of ANNSTLF requires an initial training with historical hourly load and weather data. Two weather parameters, temperature and relative humidity, from either one or multiple locations can be utilized. In the operational phase, the previous day`s load and weather datamore »« less
An adaptive modular artificial neural network hourly load forecaster and its implementation at electric utilities

Journal Article Khotanzad, A; Hwang, R; Abaye, A; ... - IEEE Transactions on Power Systems

This paper describes a modular artificial neural network (ANN) based hourly load forecaster which has already been implemented at 20 electric utilities across the US and is being used on-line by several of them. The behavior of the load and its correlation with parameters affecting it (e.g. weather variables) are decomposed into three distinct trends of weekly, daily, and hourly. Each trend is modeled by a separate module containing several multi-layer feed-forward ANNs trained by the back-propagation learning rule. The forecasts produced by each module are then combined by adaptive filters to arrive at the final forecast. During the forecastingmore »« less
https://doi.org/10.1109/59.466468
Predicting effects of climate fluctuations for water management by applying neural network

Journal Article Zhang, E; Trimble, P - World Resource Review

This paper investigates the applicability of neural networks in climate based forecasting for regional water resources management. A neural network is a computational method inspired by studies of the brain and nerve systems in biological organisms. Neural networks represent highly idealized mathematical models of the authors present understanding of such complex systems. Typically, a neural network consists of a set of layered processing units and weighted interconnections between the units. There exists a variety of neural network models and learning procedures. This paper applies the most widely used Back Propagation model to the climate forecasting. While the architecture of themore »« less
An artificial neural network based groundwater flow and transport simulator

Conference Krom, T; Rosbjerg, D

Artificial neural networks are investigated as a tool for the simulation of contaminant loss and recovery in three-dimensional heterogeneous groundwater flow and contaminant transport modeling. These methods have useful applications in expert system development, knowledge base development and optimization of groundwater pollution remediation. The numerical model runs used to develop the artificial neural networks can be re-used to develop artificial neural networks to address alternative optimization problems or changed formulations of the constraints and or objective function under optimization. Artificial neural networks have been analyzed with the goal of estimating objectives which normally require the use of traditional flow andmore »« less
Weather sensitive short-term load forecasting using non-fully connected artificial neural networks

Miscellaneous Chen, Shintzo

Artificial Neural Networks (ANN) are parallel distributed models which are capable of performing nonlinear modeling and adaptation without making functional assumptions. This thesis presents a non-fully connected ANN model with cascaded training strategy for forecasting weather sensitive electric loads. The proposed model is capable of forecasting the hourly load for an entire week and has been implemented on real load data. The average absolute forecast errors of the one hour ahead forecast in the Winter and Summer test cases are shown to be 3.42% and 3.19% respectively; the average absolute forecast errors at peak in both test cases are shownmore »« less

Similar Records