One of the major problems facing the electric utility is the unknown future demand of electricity, which needs to be estimated correctly. The authors describe a neural network approach to improve short term forecasts of electricity demand. This network is based on the nonstatistical neural paradigm, back propagation, which is found to be effective for forecasting of electrical load. The load is decomposed into a daily pattern reflecting the difference in activity level during the day, a weekly pattern representing the day-of-the week effect on load, a trend component concerning the seasonal growth and a weather component reflecting the deviations in load due to weather fluctuations. The performance of this network has been compared with some commonly used conventional smoothing methods, and stochastic methods in order to demonstrate the superiority of this approach.<>
{"title":"Short term forecasting using neural network approach","authors":"D. Srinivasan, A. Liew, J.S.P. Chen","doi":"10.1109/ANN.1991.213489","DOIUrl":"https://doi.org/10.1109/ANN.1991.213489","url":null,"abstract":"One of the major problems facing the electric utility is the unknown future demand of electricity, which needs to be estimated correctly. The authors describe a neural network approach to improve short term forecasts of electricity demand. This network is based on the nonstatistical neural paradigm, back propagation, which is found to be effective for forecasting of electrical load. The load is decomposed into a daily pattern reflecting the difference in activity level during the day, a weekly pattern representing the day-of-the week effect on load, a trend component concerning the seasonal growth and a weather component reflecting the deviations in load due to weather fluctuations. The performance of this network has been compared with some commonly used conventional smoothing methods, and stochastic methods in order to demonstrate the superiority of this approach.<<ETX>>","PeriodicalId":119713,"journal":{"name":"Proceedings of the First International Forum on Applications of Neural Networks to Power Systems","volume":"40 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1991-07-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130514499","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
During the development of a (conventional) busbar protection algorithm which is able to cope with current signals distorted by current transducer saturation, the question came up, whether it would be possible to use a neural network for preprocessing the data and restoring the distorted signals. A training tool for neural networks and a set of typical distorted and undistorted current signals was selected for a verification of the idea. The test showed that the application of a neural network to this issue is possible in principal and that the signal quality is improved with respect to the needs of a busbar protection system, respectively. The ability of the neural networks to map an increasing number of input signals to reasonable output signals is investigated. Furthermore some studies were made for implementing the trained neural network in hardware.<>
{"title":"Application of neural networks in numerical busbar protection systems (NBPS)","authors":"K. Feser, U. Braun, F. Engler, A. Maier","doi":"10.1109/ANN.1991.213508","DOIUrl":"https://doi.org/10.1109/ANN.1991.213508","url":null,"abstract":"During the development of a (conventional) busbar protection algorithm which is able to cope with current signals distorted by current transducer saturation, the question came up, whether it would be possible to use a neural network for preprocessing the data and restoring the distorted signals. A training tool for neural networks and a set of typical distorted and undistorted current signals was selected for a verification of the idea. The test showed that the application of a neural network to this issue is possible in principal and that the signal quality is improved with respect to the needs of a busbar protection system, respectively. The ability of the neural networks to map an increasing number of input signals to reasonable output signals is investigated. Furthermore some studies were made for implementing the trained neural network in hardware.<<ETX>>","PeriodicalId":119713,"journal":{"name":"Proceedings of the First International Forum on Applications of Neural Networks to Power Systems","volume":"13 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1991-07-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131714658","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
The authors are concerned with the development of a neural network (NN) regulator for turbogenerator adaptive control. The NN regulator is designed based on a hierarchical architecture of neural networks. The back-propagation (BP) algorithm is used hierarchically in the NN regulator for on-line training of the turbogenerator NN model and controller. Dynamic modelling of the turbogenerator system has been investigated using the multilayer NN. The NN regulator has been implemented on a simulated complex nonlinear turbogenerator system. Simulation results evaluating the performance of the NN regulator under different operation conditions and disturbances are presented.<>
{"title":"On-line training of neural network model and controller for turbogenerators","authors":"Qinghua Wu, B. Hogg, George W. Irwin","doi":"10.1109/ANN.1991.213515","DOIUrl":"https://doi.org/10.1109/ANN.1991.213515","url":null,"abstract":"The authors are concerned with the development of a neural network (NN) regulator for turbogenerator adaptive control. The NN regulator is designed based on a hierarchical architecture of neural networks. The back-propagation (BP) algorithm is used hierarchically in the NN regulator for on-line training of the turbogenerator NN model and controller. Dynamic modelling of the turbogenerator system has been investigated using the multilayer NN. The NN regulator has been implemented on a simulated complex nonlinear turbogenerator system. Simulation results evaluating the performance of the NN regulator under different operation conditions and disturbances are presented.<<ETX>>","PeriodicalId":119713,"journal":{"name":"Proceedings of the First International Forum on Applications of Neural Networks to Power Systems","volume":"205 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1991-07-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132219810","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
A study is made on the application of the artificial neural network (ANN) method to forecast the short-term load for a large power system. The load has two distinct patterns: weekday and weekend-day patterns. The weekend-day pattern include Saturday, Sunday, and Monday loads. Three different ANN models are proposed, including two feedforward neural networks and one recurrent neural network. Inputs to the ANN are past loads and the output is the predicted load for a given day. The standard deviation and percent error of each model are compared.<>
{"title":"A study on neural networks for short-term load forecasting","authors":"K.Y. Lee, Y. T. Cha, C. Ku","doi":"10.1109/ANN.1991.213492","DOIUrl":"https://doi.org/10.1109/ANN.1991.213492","url":null,"abstract":"A study is made on the application of the artificial neural network (ANN) method to forecast the short-term load for a large power system. The load has two distinct patterns: weekday and weekend-day patterns. The weekend-day pattern include Saturday, Sunday, and Monday loads. Three different ANN models are proposed, including two feedforward neural networks and one recurrent neural network. Inputs to the ANN are past loads and the output is the predicted load for a given day. The standard deviation and percent error of each model are compared.<<ETX>>","PeriodicalId":119713,"journal":{"name":"Proceedings of the First International Forum on Applications of Neural Networks to Power Systems","volume":"22 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1991-07-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115769189","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
The static and dynamic characteristics of power system loads are critical to obtaining quality operating point predictions or stability calculations. The composite behavior of components at load buses are usually too complicated to be expressed in a simple form. Based on the approximation capability of artificial neural networks the authors explore the possibility of using neural networks to emulate load behaviours. The results verify the potential of load representation by neural networks.<>
{"title":"Approximations of power system dynamic load characteristics by artificial neural networks","authors":"R. J. Thomas, B. Ku","doi":"10.1109/ANN.1991.213482","DOIUrl":"https://doi.org/10.1109/ANN.1991.213482","url":null,"abstract":"The static and dynamic characteristics of power system loads are critical to obtaining quality operating point predictions or stability calculations. The composite behavior of components at load buses are usually too complicated to be expressed in a simple form. Based on the approximation capability of artificial neural networks the authors explore the possibility of using neural networks to emulate load behaviours. The results verify the potential of load representation by neural networks.<<ETX>>","PeriodicalId":119713,"journal":{"name":"Proceedings of the First International Forum on Applications of Neural Networks to Power Systems","volume":"480 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1991-07-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123398757","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
That the artificial neural network (ANN) can perform the pattern classification in excellent fashion is already established in the literature. The authors envisage the relay as a pattern classifying device. This opens a new dimension in relay philosophy which needs wide investigations. Keeping the microprocessor relay framework intact, the authors report the findings about the feasibility of using ANN in protection of transmission lines. The ADALINE model is explored for the application and is found to yield encouraging results. The input variables are quantified over the operating range which eases the arithmetics of the microprocessor. The training is performed in off-line mode and the converged weight matrix is stored for on-line use.<>
{"title":"Application of artificial neural network in protective relaying of transmission lines","authors":"S. Khaparde, P. Kale, S. Agarwal","doi":"10.1109/ANN.1991.213509","DOIUrl":"https://doi.org/10.1109/ANN.1991.213509","url":null,"abstract":"That the artificial neural network (ANN) can perform the pattern classification in excellent fashion is already established in the literature. The authors envisage the relay as a pattern classifying device. This opens a new dimension in relay philosophy which needs wide investigations. Keeping the microprocessor relay framework intact, the authors report the findings about the feasibility of using ANN in protection of transmission lines. The ADALINE model is explored for the application and is found to yield encouraging results. The input variables are quantified over the operating range which eases the arithmetics of the microprocessor. The training is performed in off-line mode and the converged weight matrix is stored for on-line use.<<ETX>>","PeriodicalId":119713,"journal":{"name":"Proceedings of the First International Forum on Applications of Neural Networks to Power Systems","volume":"36 1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1991-07-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124502914","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
The backpropagation algorithm for neural network training is formulated as the solution of a set of sparse differential algebraic equations (DAE). These equations are then solved as a function of time. The solution of the differential equations is performed using an implicit integrator with adjustable time step. The topology of the Jacobian matrix associated with the DAE's is illustrated. A training example is included.<>
{"title":"Back-propagation as the solution of differential-algebraic equations for artificial neural network training","authors":"J. Sanchez-Gasca, D. Klapper, J. Yoshizawa","doi":"10.1109/ANN.1991.213470","DOIUrl":"https://doi.org/10.1109/ANN.1991.213470","url":null,"abstract":"The backpropagation algorithm for neural network training is formulated as the solution of a set of sparse differential algebraic equations (DAE). These equations are then solved as a function of time. The solution of the differential equations is performed using an implicit integrator with adjustable time step. The topology of the Jacobian matrix associated with the DAE's is illustrated. A training example is included.<<ETX>>","PeriodicalId":119713,"journal":{"name":"Proceedings of the First International Forum on Applications of Neural Networks to Power Systems","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1991-07-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129423289","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
A comparison of the forecast accuracy of artificial neural networks is made to other more established forecasting methodologies. Eight different types of forecasts were developed on a daily basis for five months and results analyzed. The MAPE (mean absolute percent error) was computed for each model. The series being forecast was the total system load for the Puget Sound Power and Light Company. The performance of the neural nets was disappointing with all but one of the other techniques outperforming them. Although the neural nets did not do well in this competition, this may be caused by a lack of forecasting experience by the neural net developers rather than limitations in the abilities of nets themselves. Forecasts made with neural nets using the same inputs showed dramatic improvements but the performance was still not as good as the best regression forecast.<>
{"title":"Comparison of the forecasting accuracy of neural networks with other established techniques","authors":"M. Casey Brace, J. Schmidt, M. Hadlin","doi":"10.1109/ANN.1991.213493","DOIUrl":"https://doi.org/10.1109/ANN.1991.213493","url":null,"abstract":"A comparison of the forecast accuracy of artificial neural networks is made to other more established forecasting methodologies. Eight different types of forecasts were developed on a daily basis for five months and results analyzed. The MAPE (mean absolute percent error) was computed for each model. The series being forecast was the total system load for the Puget Sound Power and Light Company. The performance of the neural nets was disappointing with all but one of the other techniques outperforming them. Although the neural nets did not do well in this competition, this may be caused by a lack of forecasting experience by the neural net developers rather than limitations in the abilities of nets themselves. Forecasts made with neural nets using the same inputs showed dramatic improvements but the performance was still not as good as the best regression forecast.<<ETX>>","PeriodicalId":119713,"journal":{"name":"Proceedings of the First International Forum on Applications of Neural Networks to Power Systems","volume":"27 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1991-07-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127671045","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
The authors have developed a nuclear power plant diagnosis technique, transient phenomena analysis that uses neural networks. Neural networks identify failed equipment by recognizing the pattern of main plant parameters. It is possible to obtain the cause of an abnormality when a nuclear power plant is in a transient state. The neural network has 49 units on its input layer, 20 units on its hidden layer and 100 units on its output layer. Testing of the neural network was carried out with patterns that have been accumulated from past incident data by a backpropagation procedure.<>
{"title":"Development of nuclear power plant diagnosis technique using neural networks","authors":"M. Horiguchi, N. Fukawa, K. Nishimura","doi":"10.1109/ANN.1991.213463","DOIUrl":"https://doi.org/10.1109/ANN.1991.213463","url":null,"abstract":"The authors have developed a nuclear power plant diagnosis technique, transient phenomena analysis that uses neural networks. Neural networks identify failed equipment by recognizing the pattern of main plant parameters. It is possible to obtain the cause of an abnormality when a nuclear power plant is in a transient state. The neural network has 49 units on its input layer, 20 units on its hidden layer and 100 units on its output layer. Testing of the neural network was carried out with patterns that have been accumulated from past incident data by a backpropagation procedure.<<ETX>>","PeriodicalId":119713,"journal":{"name":"Proceedings of the First International Forum on Applications of Neural Networks to Power Systems","volume":"2 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1991-07-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132861079","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
The generalization ability of a neural network in a specific application is of interest to many neural network designers. Learning theory, derived from maximum entropy, is applied to a neural network used for incipient fault detection in single-phase induction motors. The authors use learning theory to predict the proper number of training examples needed to reach a specific accuracy level (before actually training the network), so that excessive and unnecessary training examples and training time can be avoided. The results of learning theory are compared to actual training results to show the efficiency and reliability of the use of learning theory.<>
{"title":"Application of learning theory to a single phase induction motor incipient fault detector artificial neural network","authors":"M. Chow, G. Bilbro, S. Yee","doi":"10.1109/ANN.1991.213504","DOIUrl":"https://doi.org/10.1109/ANN.1991.213504","url":null,"abstract":"The generalization ability of a neural network in a specific application is of interest to many neural network designers. Learning theory, derived from maximum entropy, is applied to a neural network used for incipient fault detection in single-phase induction motors. The authors use learning theory to predict the proper number of training examples needed to reach a specific accuracy level (before actually training the network), so that excessive and unnecessary training examples and training time can be avoided. The results of learning theory are compared to actual training results to show the efficiency and reliability of the use of learning theory.<<ETX>>","PeriodicalId":119713,"journal":{"name":"Proceedings of the First International Forum on Applications of Neural Networks to Power Systems","volume":"16 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1991-07-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131733078","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：