Load Forecasting In Mageta Island Microgrid System, Kenya, with a 5 kw Peak Load using Artificial Neural Network
|
|
- Cordelia Owen
- 5 years ago
- Views:
Transcription
1 Load Forecasting In Mageta Island Microgrid System, Kenya, with a 5 kw Peak Load using Artificial Neural Network J. M. Mbuthia*, C.M. Kiruki* *Department of Electrical and Information Engineering University of Nairobi, Kenya Abstract This paper presents an artificial neural network (ANN) algorithm for load forecasting in a small island community microgrid, in Lake Victoria in Kenya, with a peak load of less than 5kW. The aim is to investigate the suitability of neural networks for forecasting in small power installations. The paper aims to strip down the amount and variety of data used for forecasting. Only historical load data is used with the days divided into weekdays and weekends. Feed-forward ANN and the Levenberg-Marquardt algorithm is used. The results suggest that this historical data is sufficient to provide an acceptable 24-hr ahead forecast. This approach is significant in that various weather forecasts such as temperature and humidity may not be readily available for particular sites as they cover very small geographical areas that may possess variations of the regional and national weather forecasts. More so, this particular microgrid has a load composition that is highly independent of weather conditions. Air conditioners and heaters are non-typical loads in this system. A Mean Absolute Percent Error of less than 20% is achieved using historical data only. Keywords: Artificial Neural Network (ANN), Levenberg- Marquardt, low load forecasting, Microgrid I. INTRODUCTION According to World Energy Outlook 2014 report by International Energy Agency, an estimated 620 million people in sub-saharan Africa have no access to electricity. As a result, microgrids and minigrids have gained momentum due to these low levels of grid connectivity. The major sources of energy in such installations are solar and wind power. The intermittency of such major sources of energy necessitates proper management of the system. Of critical importance is the ability to forecast both the energy source and load demand in such a system. This will be essential in that it forms an integral part of the energy management system. The microgrid used in this study is located in Mageta Island in Kisumu, Kenya. It is operated by SteamCo Company and has a peak load of less than 5 kw. This paper investigates the application of an Artificial Neural Network (ANN) to such small microgrids. Load demand forecasting using ANN has been widely carried out for large regional and national power grid networks. Due to the aggregation of many loads in the order of MW spanning wide geographical areas, the demand curve tends to vary smoothly and follows a quick to learn and predict pattern. Consequently, load forecasting in such systems is highly accurate with a Mean Absolute Performance Error of less than 3% being reported. Al-Shareef et al in [1] used ANN for a 24-hr load forecast for the western area of Saudi Arabia with an average load of over 8000MW. In addition to historical data, weather elements (forecasted temperature and humidity) and day type were also used as inputs to the network. A MAPE of 1.82% was achieved. Rajasekaran and Subbaraj in [2] performed ANN based hourly load forecast for the power system of Chennai city (Tamilnadu state India) with an average load of 5000 MW. Historical data and weather information was used as the inputs to the network. A MAPE of 2.95% was achieved. Tomonobu et al in [3] used ANN to perform a 1-hr ahead load forecasting using load power data of Okinawa Electric Power Company with an average of 800 MW. Historical load data and temperature data were used as inputs. The neural network was trained by using the data of past 30 days from the day before forecast day and past 60 days before and after forecast day in previous year. A MAPE of 1.63 % was achieved. In contrast to the large grid networks presented above, the load in microgrids is small and consists of a small number of aggregated loads. Thus a variation in one load has a significant effect on the aggregated load than it would have on a larger national grid network. Load forecasting has been carried out for varying sizes of microgrids and the results suggest that as the load size decreases, the MAPE (forecasting error) increases. Two of these studies are briefly presented. Andrei et al in [4] applied a NN for forecasting for 230 houses with a peak load of 340 kw and 90 houses with a peak of 140 kw. The average NRMSE for the former was found to be 3.05% and for the latter, 3.82%. Llanos et al in [5] used a NN for load forecasting in an isolated village with a load peak of 15 kw and obtained a 13.8% MAPE. Thus it s clear that as the load/demand to be forecasted reduces and becomes noisier, the less accurate the forecasting becomes. For a microgrid of less than 5 kw peak load, the forecasting error is expected to be bigger than that in the afore-mentioned studies. However, with the right data inputs for training and data pre-processing, an ANN can be successfully applied in such small installations as will be presented in this study
2 II. NEURAL NETWORK ARCHITECTURE Neural network is a machine learning algorithm that is suitable for complex non-linear problems. Given a set of inputs, the network can predict the output whose accuracy is determined by the network parameters. In supervised learning, the net work is first trained. This implies there is a set of data with inputs and corresponding outputs known apriori. The inputs of this set of data are fed into the network and its parameters are adjusted till the predicted outputs closely match the real outputs. The parameters here include the number of inputs, hidden layers, nodes (neurons) in each layer and the weights mapping the transition from one layer to the next. The architecture of a network refers to the number of neurons, their arrangement and connectivity. Each neuron receives input from the neighbouring neurons connected to it, processes the information and produces an output. The feedforward and back propagation neural network architecture in Fig.1 was used in this study. The neural network has 3 layers with 15 neurons in the hidden layer. This allows the possibility of implementing the solution on an embedded platform. The Mageta Island is situated along the equator and thus temperature variations are moderate throughout the year. More so, the system serves a few households and kiosks and thus barely has loads that depend on weather conditions such as AC systems. Consequently, no weather conditions were taken into account in the forecasting. This eliminates need for additional sensors in the system and knowledge of future weather conditions, which may not be readily available in some places. points in the system. This allows remote monitoring and real time data logging. Specifically, the meters are used to measure energy from the solar panels and wind turbine, energy in and out of the batteries and energy into the inverter (load). This data is relayed through the cloud to a dashboard where further data processing and analysis can be carried out. Fig.2 and Fig.3 depict the custom meters and the data visualization in the dashboard respectively. The microgrid is new and has been in operation for barely a year. It is a 24V, approximately 5 KW system with the following components: Solar: 2 arrays, each of 12 x 235W 24V panels Charge Controller: 2 x 60A MPPT controllers Batteries: 8 x 200AH 6V lead acid flooded in two strings. Inverter: 24V 4100VA continuous output Wind turbine: 1kW rectified to DC in the cabin and controlled by a dump load. Standby diesel generator Input Layer Hidden Layer Output Layer Figure 2: Custom Energy Meters x 1 a 1 (2) Different inputs for Models 1-5 x 2 x m a 2 (2) a n (2) a 1 (3) Forecast Figure 1: Neural Network Architecture MATLAB Neural Network Toolkit has been used for the training of the network. This kit uses the Levenberg- Marquardt (LM) algorithm for the back propagation training. It provides a numerical solution to the problem of minimizing a nonlinear function. LM blends the stability of the steepest descent method and the speed of the Gauss- Newton algorithm. Martin and Menhaj in [6] observed that the Marquardt algorithm was very efficient when training networks that have up to a few hundred weights. III. DATA LOGGING Custom DC meters with remote communication capabilities have been built and installed at various energy Figure 3: Dashboard Visualization 477
3 The custom remote monitoring system has been installed at Mageta Island microgrid. Data considered for load forecasting spanned five months commencing February 2016 to June Weekends and holidays have not been factored in the network training. Due to the location of the microgrid system, any holiday load would closely follow the weekend load profile. The number of weekends and holidays available in 5 months is very small to provide adequate training and testing of a neural network. Therefore, only the weekdays data has been used extensively for the forecasting phase of this study. Data pre-processing The data needs some pre-processing before being used in the ANN. The data is updated every 10 minutes and thus for every hour there are 6 data points. This translates to 144 data points for a full day. While this gives a more accurate picture of the real time operation of the site, it would be computationally expensive to train the network with such a huge number of inputs. Therefore, hourly averages are calculated resulting in 24 data points for a full day. The result of such averaging is presented in Fig. 4. Missing data is filled through extrapolation. Figure 4: Hourly Data Averaging IV. NETWORK TRAINING AND TESTING The network training and testing is done under two tests. In Test One, the goal is to determine the nature of the historical data that gives the best forecasting capability. That is, for example, whether considering the previous 5 days load data gives a better prediction for today s load or is yesterday s load enough to predict today s load. In Test Two, the goal is to determine the amount of historical data adequate to train the network. That is, for example, whether training the network with 3 months data is better than training it with a single month s data. A. Test One A month s historical data is used to train the network. 28 weekdays load data commencing 29 th Feb to 6 th April was used. As earlier noted, the weekends have been omitted in this training data. The weekday network model was tested for five different training models: Model 1: The data for the previous four consecutive weekdays and the same day one week before was used for training the network. To forecast data for a day i.e. 24 hours ahead load, the previous 5 weekdays load was used as input. Thus there are 5 neurons in the input layer. Model 2: Data for the previous five consecutive days and 1hr before, 2hrs before, 3hrs before the hour of interest for each of the five days was used. Thus there are 20 neurons in the input layer. Model 3: Data for the previous three consecutive weekdays is used. This results in 3 neurons at the input layer. Model 4: Data for the previous two consecutive weekdays is used. This results in 2 neurons at the input layer. Model 5: Data for the previous two consecutive weekdays and 1hr before, 2hrs before, 3hrs before the hour of interest for each of the two days was used. Thus there are 8 neurons in the input layer. For testing the network, data that was not used in the training was considered. E.g. to forecast the data for 14 th March, the neural network would require data for the previous 5 weekdays (Model 1 and 2) dating back to 9 th March. Since the training phase used data for 28 weekdays starting 29 th Feb to 6 th April, the model was tested for dates starting 14 th April. Thus to forecast the load for 14 th April (Thursday) using model 1, the load for the previous 5 weekdays was required as the input to the neural network. These weekdays are: 13 th, 12 th, 11 th, 8 th, 7th; with 9 th and 10 th omitted since they fall on a weekend. This ensured that no data used in the training process was used for the testing phase. B. Test Two Test One is only concerned with how the various input models affect the accuracy of the neural network. The network was trained with a constant number of historical load data, that is, 28 weekdays. In this test, the additional effect of the number of days used to train the network is considered. In addition to evaluating the various models, the network will be trained using 30, 40 and 60 days. The network is tested by forecasting the load for 13 th, 14 th and 15 th June V. RESULTS AND DISCUSSION The forecasting results are evaluated using the Mean Absolute Percent Error of the formula: MAPE = ( 1 N Actual i Forecast i N i=1 ) 100 (1) Actual i where N is the number of forecasting points. The table below shows the results obtained for test one: 478
4 Table 1: Test One Results Date Model 1 Model 2 Model 3 Model 4 Model 5 14/4/ /4/ /4/ /4/ /4/ Weekdays Average Weekend Average The data highlighted in purple indicates the weekend data and as expected, the forecasting error as calculated using equation (1) is higher as compared to that of the weekdays. This is because the neural network has been trained using weekdays data only and thus forecasting a weekend load using this network would result in higher errors. This shows that it s better to use two separate networks for weekdays and weekends load forecasting. As earlier noted, there weren t enough weekend data to sufficiently train the network. It s evident that Model 2 which considers the previous 5 days and 1hr before, 2hrs before, 3hrs before the hour of interest performed relatively poorly. Model 3 which considers data for only the previous 3 days had the best performance. The graphs below show the results for 24 hr ahead load forecasting for 14 th, 17 th and 18 th March 2016 using model 1. Figure 6: Actual and Forecast Load for 15 April, 2016 In Test Two, the following tables show the results for every model. Table 2 : Model 1 No. of Training 30 Days 40 Days 60 Days Days 13/06/ /06/ /06/ AVERAGE Table 3: Model 2 13/06/ /06/ /06/ AVERAGE Table 4: Model 3 13/06/ /06/ /06/ AVERAGE Table 5: Model 4 13/06/ /06/ /06/ AVERAGE Figure 5: Actual and Forecast Load for Apr 14, 2016 Table 6 : Model 5 13/06/ /06/ /06/ AVERAGE
5 The following graphs show the results for 15 th June using Model 5 with various training days: network is accumulated and thus the more accurate the forecasting gets. Figure 7: Actual and Forecast Load with 30 Days Training VI. CONCLUSION A MAPE performance of less than 20 % has been achieved in the load forecasting. Given that the peak load recorded in the period of observation is less than 4 kw, this is an acceptable result. From the literature review, it is noted that the accuracy of forecasting decreases as the load size decreases. Most forecasting has been carried out for large systems in the range of tens of MW with a MAPE of less than 3%. This shows that the forecasting approach taken in this study will give even better results if undertaken in a larger microgrid system. From this study, it s evident that models 3, 4 and 5 are generally more suited in making 24 hr ahead load forecast. Therefore, knowledge of the previous two or three days load is sufficient to make adequately accurate forecast of the present day load profile. In addition, the study gives an insight that the more historical data used in training the network, the better the accuracy in forecasting. This is evidenced by the relatively better accuracy reported when using 60 training days in all the five models. Consequently, the older the microgrid monitoring system gets, the more the historical data for training the network is accumulated and thus the more accurate the forecasting gets. Figure 8: Actual and Forecast Load with 40 Days Training Figure 9: Actual and Forecast Load with 60 Days Training From this Test 2, it s evident that models 3, 4 and 5 are generally more suited in making 24 hr ahead load forecast. This is supported by the results obtained in Test 1. Therefore, knowledge of the previous two or three days load is sufficient to make adequately accurate forecast of the present day load profile. In addition, Test 2 reveals that the more historical data used in training the network, the better the accuracy in forecasting. This is evidenced by the relatively better accuracy reported when using 60 training days in all the five models. Consequently, the older the microgrid system gets, the more the historical data for training the REFERENCES [1] A.J. Al-Shareef, E.A. Mohamed, and E. Al-Judaibi, "Next 24-Hours Load Forecasting Using Artificial Neural Network (ANN) for the Western Area of Saudi Arabia," JKAU: Engineering Sciences, pp , [2] Dr. P. Subbaraj and V. Rajasekaran, "Short Term Hourly Load Forecasting Using Combined Artificial Neural Networks," in IEEE International Conference on Computational Intelligence and Multimedia Applications, [3] Tomonobu Senjyu, Hitoshi Takara, Katsumi Uezato, and Toshihisa Funabashi, "One-Hour-Ahead Load Forecasting Using Neural Network," IEEE Transactions On Power Systems, vol. 17, no. 1, pp , February [4] Andrei Marinescu, Colin Harris, Ivana Dusparic, Siobhan Clarke, and Vinny Cahill, "Residential Electrical Demand Forecasting in Very Small Scale: An Evaluation of Forecasting Methods," in Proceeds of IEEE 2nd International Workshop on Software Engineering Challenges for the Smart Grid (SE4SG), San Francisco, CA, 2013, pp [5] J. Llanos, D. Sáez, R. Palma-Behnke, A. Núñez, and G. Jiménez-Estévez, "Load profile generator and load forecasting for a renewable based microgrid using Self Organizing Maps and neural networks," in Proceeds of IEEE 2012 International Joint Conference onneural Networks (IJCNN), Brisbane, QLD, 2012, pp [6] Martin T. Hagan and Mohammad B. Menhaj, "Training Feedforward Networks with the Marquardt Algorithm," IEEE Transactions On Neural Networks, vol. 5, no. 6, pp , November [7] Maryam Ramezani, Hamid Falaghi, and Mahmood-Reza Haghifam, "Short-Term Electric Load Forecasting Using Neural Networks," in Proceeds of EUROCON 2005.The International Conference on Computer as a Tool, Belgrade, Serbia, Novemeber,2005, pp
6 [8] Albert Molderink, Vincent Bakker, Maurice G.C. Bosman, Johann L. Hurink, and Gerard J.M. Smit, "Management and control of domestic smart grid technology," IEEE Transactions on Smart Grid, vol. 1, no. 2, pp , September [9] Mara Lficia M. Lopes, Carlos R. Minussi, and Anna Diva P. htufo, "A Fast Electric Load Forecasting Using Neural Networks," in Proceedings of the 43rd IEEE Midwest Symposium on Circuits and Systems, Lansing MI, 2000, pp Mwangi Mbuthia is an Associate Professor and currently the Dean, School of Engineering at the University of Nairobi. He received a B.Sc. (First Class Honors) in Electrical and Electronic Engineering in 1976 from the University Nairobi, An M.Sc. and DIC in 1978 from Imperial College of Science Technology and Medicine in London, and a Ph.D in 1985 from the University of Manchester in England. His research interests include broadband last mile connectivity devices, special materials and devices for solar power applications, IoT devices, smart sensor and artificial intelligence applications in energy delivery automation and smart grids. He has developed many products and systems and published widely. Kiruki Cosmas Raymond received a B.Sc. (First Class Honors) in Electrical and Electronic Engineering in 2014 from the University of Nairobi, Kenya. Currently, he is a Research Assistant at the Electrical and Information Engineering department at the University of Nairobi, where he is pursuing an M.sc. in Eletrical and Electronic Engineering. His research interests include IoT and embedded systems design, artificial intelligence and machine learning and their applications to renewable energy technologies and smart grids
One-Hour-Ahead Load Forecasting Using Neural Network
IEEE TRANSACTIONS ON POWER SYSTEMS, VOL. 17, NO. 1, FEBRUARY 2002 113 One-Hour-Ahead Load Forecasting Using Neural Network Tomonobu Senjyu, Member, IEEE, Hitoshi Takara, Katsumi Uezato, and Toshihisa Funabashi,
More informationApplication of Artificial Neural Network for Short Term Load Forecasting
aerd Scientific Journal of Impact Factor(SJIF): 3.134 e-issn(o): 2348-4470 p-issn(p): 2348-6406 International Journal of Advance Engineering and Research Development Volume 2,Issue 4, April -2015 Application
More informationShort Term Load Forecasting Based Artificial Neural Network
Short Term Load Forecasting Based Artificial Neural Network Dr. Adel M. Dakhil Department of Electrical Engineering Misan University Iraq- Misan Dr.adelmanaa@gmail.com Abstract Present study develops short
More informationApplication of Artificial Neural Networks in Evaluation and Identification of Electrical Loss in Transformers According to the Energy Consumption
Application of Artificial Neural Networks in Evaluation and Identification of Electrical Loss in Transformers According to the Energy Consumption ANDRÉ NUNES DE SOUZA, JOSÉ ALFREDO C. ULSON, IVAN NUNES
More informationDay Ahead Hourly Load and Price Forecast in ISO New England Market using ANN
23 Annual IEEE India Conference (INDICON) Day Ahead Hourly Load and Price Forecast in ISO New England Market using ANN Kishan Bhushan Sahay Department of Electrical Engineering Delhi Technological University
More informationArtificial Neural Network for Energy Demand Forecast
International Journal of Electrical and Electronic Science 2018; 5(1): 8-13 http://www.aascit.org/journal/ijees ISSN: 2375-2998 Artificial Neural Network for Energy Demand Forecast Akpama Eko James, Vincent
More informationA Hybrid Model of Wavelet and Neural Network for Short Term Load Forecasting
International Journal of Electronic and Electrical Engineering. ISSN 0974-2174, Volume 7, Number 4 (2014), pp. 387-394 International Research Publication House http://www.irphouse.com A Hybrid Model of
More informationSMART GRID FORECASTING
SMART GRID FORECASTING AND FINANCIAL ANALYTICS Itron Forecasting Brown Bag December 11, 2012 PLEASE REMEMBER» Phones are Muted: In order to help this session run smoothly, your phones are muted.» Full
More informationANN and Statistical Theory Based Forecasting and Analysis of Power System Variables
ANN and Statistical Theory Based Forecasting and Analysis of Power System Variables Sruthi V. Nair 1, Poonam Kothari 2, Kushal Lodha 3 1,2,3 Lecturer, G. H. Raisoni Institute of Engineering & Technology,
More informationWind Power Forecasting using Artificial Neural Networks
Wind Power Forecasting using Artificial Neural Networks This paper aims at predicting the power output of wind turbines using artificial neural networks,two different algorithms and models were trained
More informationShort Term Load Forecasting Using Multi Layer Perceptron
International OPEN ACCESS Journal Of Modern Engineering Research (IJMER) Short Term Load Forecasting Using Multi Layer Perceptron S.Hema Chandra 1, B.Tejaswini 2, B.suneetha 3, N.chandi Priya 4, P.Prathima
More informationESTIMATION OF HOURLY MEAN AMBIENT TEMPERATURES WITH ARTIFICIAL NEURAL NETWORKS 1. INTRODUCTION
Mathematical and Computational Applications, Vol. 11, No. 3, pp. 215-224, 2006. Association for Scientific Research ESTIMATION OF HOURLY MEAN AMBIENT TEMPERATURES WITH ARTIFICIAL NEURAL NETWORKS Ömer Altan
More informationA Novel 2-D Model Approach for the Prediction of Hourly Solar Radiation
A Novel 2-D Model Approach for the Prediction of Hourly Solar Radiation F Onur Hocao glu, Ö Nezih Gerek, and Mehmet Kurban Anadolu University, Dept of Electrical and Electronics Eng, Eskisehir, Turkey
More informationImplementation of Artificial Neural Network for Short Term Load Forecasting
Implementation of Artificial Neural Network for Short Term Load Forecasting Anju Bala * * Research Scholar, E.E. Department, D.C.R. University of Sci. & Technology, Murthal, anju8305@gmail.com N. K. Yadav
More informationShort Term Load Forecasting Of Chhattisgarh Grid Using Artificial Neural Network
Short Term Load Forecasting Of Chhattisgarh Grid Using Artificial Neural Network 1 Saurabh Ghore, 2 Amit Goswami 1 M.Tech. Student, 2 Assistant Professor Department of Electrical and Electronics Engineering,
More informationReal Time wave forecasting using artificial neural network with varying input parameter
82 Indian Journal of Geo-Marine SciencesINDIAN J MAR SCI VOL. 43(1), JANUARY 2014 Vol. 43(1), January 2014, pp. 82-87 Real Time wave forecasting using artificial neural network with varying input parameter
More informationHow Accurate is My Forecast?
How Accurate is My Forecast? Tao Hong, PhD Utilities Business Unit, SAS 15 May 2012 PLEASE STAND BY Today s event will begin at 11:00am EDT The audio portion of the presentation will be heard through your
More informationModelling and Prediction of 150KW PV Array System in Northern India using Artificial Neural Network
International Journal of Engineering Science Invention ISSN (Online): 2319 6734, ISSN (Print): 2319 6726 Volume 5 Issue 5 May 2016 PP.18-25 Modelling and Prediction of 150KW PV Array System in Northern
More informationShort Term Load Forecasting for Bakhtar Region Electric Co. Using Multi Layer Perceptron and Fuzzy Inference systems
Short Term Load Forecasting for Bakhtar Region Electric Co. Using Multi Layer Perceptron and Fuzzy Inference systems R. Barzamini, M.B. Menhaj, A. Khosravi, SH. Kamalvand Department of Electrical Engineering,
More informationShort-Term Demand Forecasting Methodology for Scheduling and Dispatch
Short-Term Demand Forecasting Methodology for Scheduling and Dispatch V1.0 March 2018 Table of Contents 1 Introduction... 3 2 Historical Jurisdictional Demand Data... 3 3 EMS Demand Forecast... 4 3.1 Manual
More informationShort-term wind forecasting using artificial neural networks (ANNs)
Energy and Sustainability II 197 Short-term wind forecasting using artificial neural networks (ANNs) M. G. De Giorgi, A. Ficarella & M. G. Russo Department of Engineering Innovation, Centro Ricerche Energia
More informationElectric Load Forecasting Using Wavelet Transform and Extreme Learning Machine
Electric Load Forecasting Using Wavelet Transform and Extreme Learning Machine Song Li 1, Peng Wang 1 and Lalit Goel 1 1 School of Electrical and Electronic Engineering Nanyang Technological University
More informationSingle Home Electricity Power Consumption Forecast Using Neural Networks Model
Single Home Electricity Power Consumption Forecast Using Neural Networks Model Naser Farag Abed 1 and Milan M.Milosavljevic 1,2 1 Singidunum University Belgrade, 11000, Serbia 2 School of Electrical Engineering,
More informationDemand Forecasting in Deregulated Electricity Markets
International Journal of Computer Applications (975 8887) Demand Forecasting in Deregulated Electricity Marets Anamia Electrical & Electronics Engineering Department National Institute of Technology Jamshedpur
More information1. Introduction. 2. Artificial Neural Networks and Fuzzy Time Series
382 IJCSNS International Journal of Computer Science and Network Security, VOL.8 No.9, September 2008 A Comparative Study of Neural-Network & Fuzzy Time Series Forecasting Techniques Case Study: Wheat
More informationMarkovian Models for Electrical Load Prediction in Smart Buildings
Markovian Models for Electrical Load Prediction in Smart Buildings Muhammad Kumail Haider, Asad Khalid Ismail, and Ihsan Ayyub Qazi LUMS School of Science and Engineering, Lahore, Pakistan {8,,ihsan.qazi}@lums.edu.pk
More informationAbout Nnergix +2, More than 2,5 GW forecasted. Forecasting in 5 countries. 4 predictive technologies. More than power facilities
About Nnergix +2,5 5 4 +20.000 More than 2,5 GW forecasted Forecasting in 5 countries 4 predictive technologies More than 20.000 power facilities Nnergix s Timeline 2012 First Solar Photovoltaic energy
More informationA Feature Based Neural Network Model for Weather Forecasting
World Academy of Science, Engineering and Technology 4 2 A Feature Based Neural Network Model for Weather Forecasting Paras, Sanjay Mathur, Avinash Kumar, and Mahesh Chandra Abstract Weather forecasting
More informationURD Cable Fault Prediction Model
1 URD Cable Fault Prediction Model Christopher Gubala ComEd General Engineer Reliability Analysis 2014 IEEE PES General Meeting Utility Current Practices & Challenges of Predictive Distribution Reliability
More informationDevising Hourly Forecasting Solutions Regarding Electricity Consumption in the Case of Commercial Center Type Consumers
energies Article Devising Hourly Forecasting Solutions Regarding Electricity Consumption in Case of Commercial Center Type Consumers Alexandru Pîrjan 1, * ID, Simona-Vasilica Oprea 2, George Căruțașu 1
More informationA SEASONAL FUZZY TIME SERIES FORECASTING METHOD BASED ON GUSTAFSON-KESSEL FUZZY CLUSTERING *
No.2, Vol.1, Winter 2012 2012 Published by JSES. A SEASONAL FUZZY TIME SERIES FORECASTING METHOD BASED ON GUSTAFSON-KESSEL * Faruk ALPASLAN a, Ozge CAGCAG b Abstract Fuzzy time series forecasting methods
More informationThunderstorm Forecasting by using Artificial Neural Network
Thunderstorm Forecasting by using Artificial Neural Network N.F Nik Ismail, D. Johari, A.F Ali, Faculty of Electrical Engineering Universiti Teknologi MARA 40450 Shah Alam Malaysia nikfasdi@yahoo.com.my
More information2013 FORECAST ACCURACY BENCHMARKING SURVEY AND ENERGY
2013 FORECAST ACCURACY BENCHMARKING SURVEY AND ENERGY Itron Forecasting Brown Bag June 4, 2013 Please Remember» Phones are Muted: In order to help this session run smoothly, your phones are muted.» Full
More informationSYSTEM OPERATIONS. Dr. Frank A. Monforte
SYSTEM OPERATIONS FORECASTING Dr. Frank A. Monforte Itron s Forecasting Brown Bag Seminar September 13, 2011 PLEASE REMEMBER» In order to help this session run smoothly, your phones are muted.» To make
More informationMs. Cheryl Blundon Director Corporate Services & Board Secretary
Ai\I or newfoundland!abrader k hydro a nalcor energy company Hydro Place. 500 Columbus Drive. P.O. Box 12400. St. John's. NI. Canada Al 4K7 t. 709.737.1400 f. 709.737.1800 www.n1h.nl.ca May 13, 2015 The
More informationSmart Meter Based Short-Term Load Forecasting for Residential Customers
Smart Meter Based Short-Term Load Forecasting for Residential Customers M. Ghofrani, SM IEEE, M. Hassanzadeh, SM IEEE, M. Etezadi-Amoli, life Sr. Member IEEE, M. S. Fadali, Sr. Member IEEE Department of
More informationResearch Article Weather Forecasting Using Sliding Window Algorithm
ISRN Signal Processing Volume 23, Article ID 5654, 5 pages http://dx.doi.org/.55/23/5654 Research Article Weather Forecasting Using Sliding Window Algorithm Piyush Kapoor and Sarabjeet Singh Bedi 2 KvantumInc.,Gurgaon22,India
More informationDirector Corporate Services & Board Secretary
March, The Board of Commissioners of Public Utilities Prince Charles Building Torbay Road, P.O. Box 0 St. John s, NL AA B Attention: Ms. Cheryl Blundon Director Corporate Services & Board Secretary Dear
More informationPrediction of Hourly Solar Radiation in Amman-Jordan by Using Artificial Neural Networks
Int. J. of Thermal & Environmental Engineering Volume 14, No. 2 (2017) 103-108 Prediction of Hourly Solar Radiation in Amman-Jordan by Using Artificial Neural Networks M. A. Hamdan a*, E. Abdelhafez b
More informationFeature Selection Optimization Solar Insolation Prediction Using Artificial Neural Network: Perspective Bangladesh
American Journal of Engineering Research (AJER) 2016 American Journal of Engineering Research (AJER) e-issn: 2320-0847 p-issn : 2320-0936 Volume-5, Issue-8, pp-261-265 www.ajer.org Research Paper Open
More informationA Fuzzy Logic Based Short Term Load Forecast for the Holidays
A Fuzzy Logic Based Short Term Load Forecast for the Holidays Hasan H. Çevik and Mehmet Çunkaş Abstract Electric load forecasting is important for economic operation and planning. Holiday load consumptions
More informationData and prognosis for renewable energy
The Hong Kong Polytechnic University Department of Electrical Engineering Project code: FYP_27 Data and prognosis for renewable energy by Choi Man Hin 14072258D Final Report Bachelor of Engineering (Honours)
More informationMODELLING TRAFFIC FLOW ON MOTORWAYS: A HYBRID MACROSCOPIC APPROACH
Proceedings ITRN2013 5-6th September, FITZGERALD, MOUTARI, MARSHALL: Hybrid Aidan Fitzgerald MODELLING TRAFFIC FLOW ON MOTORWAYS: A HYBRID MACROSCOPIC APPROACH Centre for Statistical Science and Operational
More informationBetter Weather Data Equals Better Results: The Proof is in EE and DR!
Better Weather Data Equals Better Results: The Proof is in EE and DR! www.weatherbughome.com Today s Speakers: Amena Ali SVP and General Manager WeatherBug Home Michael Siemann, PhD Senior Research Scientist
More informationNEURO-FUZZY CONTROL IN LOAD FORECASTING OF POWER SECTOR
NEURO-FUZZY CONTROL IN LOAD FORECASTING OF POWER SECTOR S.HEMA CHANDRA 1,VENDOTI MOUNIKA 2 & GUNTURU VANDANA 3 1,2,3 Department of Electronics and Control Engineering, Sree Vidyanikethan Engineering College
More informationApplication of Fully Recurrent (FRNN) and Radial Basis Function (RBFNN) Neural Networks for Simulating Solar Radiation
Bulletin of Environment, Pharmacology and Life Sciences Bull. Env. Pharmacol. Life Sci., Vol 3 () January 04: 3-39 04 Academy for Environment and Life Sciences, India Online ISSN 77-808 Journal s URL:http://www.bepls.com
More informationLoad Forecasting Using Artificial Neural Networks and Support Vector Regression
Proceedings of the 7th WSEAS International Conference on Power Systems, Beijing, China, September -7, 2007 3 Load Forecasting Using Artificial Neural Networks and Support Vector Regression SILVIO MICHEL
More informationDAY AHEAD FORECAST OF SOLAR POWER FOR OPTIMAL GRID OPERATION
DAY AHEAD FORECAST OF SOLAR POWER FOR OPTIMAL GRID OPERATION Jeenu Jose 1, Vijaya Margaret 2 1 PG Scholar, Department of Electrical & Electronics Engineering, Christ Uinversity, India. 2 Assistant Professor,
More informationCOMPARISON OF CLEAR-SKY MODELS FOR EVALUATING SOLAR FORECASTING SKILL
COMPARISON OF CLEAR-SKY MODELS FOR EVALUATING SOLAR FORECASTING SKILL Ricardo Marquez Mechanical Engineering and Applied Mechanics School of Engineering University of California Merced Carlos F. M. Coimbra
More informationA new method for short-term load forecasting based on chaotic time series and neural network
A new method for short-term load forecasting based on chaotic time series and neural network Sajjad Kouhi*, Navid Taghizadegan Electrical Engineering Department, Azarbaijan Shahid Madani University, Tabriz,
More informationRenewables and the Smart Grid. Trip Doggett President & CEO Electric Reliability Council of Texas
Renewables and the Smart Grid Trip Doggett President & CEO Electric Reliability Council of Texas North American Interconnected Grids The ERCOT Region is one of 3 North American grid interconnections. The
More informationHYBRID ARTIFICIAL NEURAL NETWORK SYSTEM FOR SHORT-TERM LOAD FORECASTING
Ilić, S. A. et al.: Hybrid Artificial Neural Network System for Short-Term... S215 HYBRID ARTIFICIAL NEURAL NETWORK SYSTEM FOR SHORT-TERM LOAD FORECASTING by Slobodan A. ILIĆ*, Srdjan M. VUKMIROVIĆ, Aleksandar
More informationIN recent years, introduction of an alternative energy source
Application of Neural Network to One-Day-Ahead 24 hours Generating Power Forecasting for Photovoltaic System Atsushi Yona, Student Member, IEEE, Tomonobu Senjyu, Senior Member, IEEE, Ahmed Yousuf Saber,
More informationWEATHER NORMALIZATION METHODS AND ISSUES. Stuart McMenamin Mark Quan David Simons
WEATHER NORMALIZATION METHODS AND ISSUES Stuart McMenamin Mark Quan David Simons Itron Forecasting Brown Bag September 17, 2013 Please Remember» Phones are Muted: In order to help this session run smoothly,
More informationPredicting the Electricity Demand Response via Data-driven Inverse Optimization
Predicting the Electricity Demand Response via Data-driven Inverse Optimization Workshop on Demand Response and Energy Storage Modeling Zagreb, Croatia Juan M. Morales 1 1 Department of Applied Mathematics,
More informationMODELLING ENERGY DEMAND FORECASTING USING NEURAL NETWORKS WITH UNIVARIATE TIME SERIES
MODELLING ENERGY DEMAND FORECASTING USING NEURAL NETWORKS WITH UNIVARIATE TIME SERIES S. Cankurt 1, M. Yasin 2 1&2 Ishik University Erbil, Iraq 1 s.cankurt@ishik.edu.iq, 2 m.yasin@ishik.edu.iq doi:10.23918/iec2018.26
More informationDemand Forecasting Reporting Period: 19 st Jun th Sep 2017
N A T I O N A L G R I D P A G E 1 O F 21 C O M M E R C I A L, E L E C T R I C I T Y C O M M E R C I A L O P E R A T I O N S Demand Forecasting Reporting Period: 19 st Jun 2017 10 th Sep 2017 EXECUTIVE
More informationReport on System-Level Estimation of Demand Response Program Impact
Report on System-Level Estimation of Demand Response Program Impact System & Resource Planning Department New York Independent System Operator April 2012 1 2 Introduction This report provides the details
More informationMonthly Long Range Weather Commentary Issued: February 15, 2015 Steven A. Root, CCM, President/CEO
Monthly Long Range Weather Commentary Issued: February 15, 2015 Steven A. Root, CCM, President/CEO sroot@weatherbank.com JANUARY 2015 Climate Highlights The Month in Review During January, the average
More informationpeak half-hourly Tasmania
Forecasting long-term peak half-hourly electricity demand for Tasmania Dr Shu Fan B.S., M.S., Ph.D. Professor Rob J Hyndman B.Sc. (Hons), Ph.D., A.Stat. Business & Economic Forecasting Unit Report for
More informationAn Operational Solar Forecast Model For PV Fleet Simulation. Richard Perez & Skip Dise Jim Schlemmer Sergey Kivalov Karl Hemker, Jr.
An Operational Solar Forecast Model For PV Fleet Simulation Richard Perez & Skip Dise Jim Schlemmer Sergey Kivalov Karl Hemker, Jr. Adam Kankiewicz Historical and forecast platform Blended forecast approach
More informationUSE OF FUZZY LOGIC TO INVESTIGATE WEATHER PARAMETER IMPACT ON ELECTRICAL LOAD BASED ON SHORT TERM FORECASTING
Nigerian Journal of Technology (NIJOTECH) Vol. 35, No. 3, July 2016, pp. 562 567 Copyright Faculty of Engineering, University of Nigeria, Nsukka, Print ISSN: 0331-8443, Electronic ISSN: 2467-8821 www.nijotech.com
More informationForecasting Hourly Electricity Load Profile Using Neural Networks
Forecasting Hourly Electricity Load Profile Using Neural Networks Mashud Rana and Irena Koprinska School of Information Technologies University of Sydney Sydney, Australia {mashud.rana, irena.koprinska}@sydney.edu.au
More informationSOIL MOISTURE MODELING USING ARTIFICIAL NEURAL NETWORKS
Int'l Conf. Artificial Intelligence ICAI'17 241 SOIL MOISTURE MODELING USING ARTIFICIAL NEURAL NETWORKS Dr. Jayachander R. Gangasani Instructor, Department of Computer Science, jay.gangasani@aamu.edu Dr.
More informationTRANSMISSION BUSINESS LOAD FORECAST AND METHODOLOGY
Filed: September, 00 EB-00-00 Tab Schedule Page of 0 TRANSMISSION BUSINESS LOAD FORECAST AND METHODOLOGY.0 INTRODUCTION 0 This exhibit discusses Hydro One Networks transmission system load forecast and
More informationA Unified Framework for Near-term and Short-term System Load Forecasting
Forecasting / Load Research A Unified Framework for Near-term and Short-term System Load Forecasting Dr. Frank A. Monforte Director, Forecasting Solutions 2009, Itron Inc. All rights reserved. 1 A Unified
More informationAddress for Correspondence
Research Article APPLICATION OF ARTIFICIAL NEURAL NETWORK FOR INTERFERENCE STUDIES OF LOW-RISE BUILDINGS 1 Narayan K*, 2 Gairola A Address for Correspondence 1 Associate Professor, Department of Civil
More informationAustralian Journal of Basic and Applied Sciences. A Comparative Analysis of Neural Network based Short Term Load Forecast for Seasonal Prediction
Australian Journal of Basic and Applied Sciences, 7() Sep 03, Pages: 49-48 AENSI Journals Australian Journal of Basic and Applied Sciences Journal home page: www.ajbasweb.com A Comparative Analysis of
More information2018 Annual Review of Availability Assessment Hours
2018 Annual Review of Availability Assessment Hours Amber Motley Manager, Short Term Forecasting Clyde Loutan Principal, Renewable Energy Integration Karl Meeusen Senior Advisor, Infrastructure & Regulatory
More informationSolar Irradiance Prediction using Neural Model
Volume-8, Issue-3, June 2018 International Journal of Engineering and Management Research Page Number: 241-245 DOI: doi.org/10.31033/ijemr.8.3.32 Solar Irradiance Prediction using Neural Model Raj Kumar
More informationAnalyzing the effect of Weather on Uber Ridership
ABSTRACT MWSUG 2016 Paper AA22 Analyzing the effect of Weather on Uber Ridership Snigdha Gutha, Oklahoma State University Anusha Mamillapalli, Oklahoma State University Uber has changed the face of taxi
More informationCARLOS F. M. COIMBRA (PI) HUGO T. C. PEDRO (CO-PI)
HIGH-FIDELITY SOLAR POWER FORECASTING SYSTEMS FOR THE 392 MW IVANPAH SOLAR PLANT (CSP) AND THE 250 MW CALIFORNIA VALLEY SOLAR RANCH (PV) PROJECT CEC EPC-14-008 CARLOS F. M. COIMBRA (PI) HUGO T. C. PEDRO
More informationOptimum Neural Network Architecture for Precipitation Prediction of Myanmar
Optimum Neural Network Architecture for Precipitation Prediction of Myanmar Khaing Win Mar, Thinn Thu Naing Abstract Nowadays, precipitation prediction is required for proper planning and management of
More informationJournal of of Computer Applications Research Research and Development and Development (JCARD), ISSN (Print), ISSN
JCARD Journal of of Computer Applications Research Research and Development and Development (JCARD), ISSN 2248-9304(Print), ISSN 2248-9312 (JCARD),(Online) ISSN 2248-9304(Print), Volume 1, Number ISSN
More informationEE-588 ADVANCED TOPICS IN NEURAL NETWORK
CUKUROVA UNIVERSITY DEPARTMENT OF ELECTRICAL&ELECTRONICS ENGINEERING EE-588 ADVANCED TOPICS IN NEURAL NETWORK THE PROJECT PROPOSAL AN APPLICATION OF NEURAL NETWORKS FOR WEATHER TEMPERATURE FORECASTING
More informationCOMPARISON OF PEAK FORECASTING METHODS. Stuart McMenamin David Simons
COMPARISON OF PEAK FORECASTING METHODS Stuart McMenamin David Simons Itron Forecasting Brown Bag March 24, 2015 PLEASE REMEMBER» Phones are Muted: In order to help this session run smoothly, your phones
More informationLoad Forecasting at Newfoundland and Labrador Hydro Monthly Report: October 2015.
\ I newfoundland labrador.hydro a nalcor energy company Hydro Place. 00 Columbus Drive. P.O. Box 00. St. John's. NI Canada AB ( t. 0..00 f. 0..00 www.nh.nl.ca November, 0 The Board of Commissioners of
More informationDefining Normal Weather for Energy and Peak Normalization
Itron White Paper Energy Forecasting Defining Normal Weather for Energy and Peak Normalization J. Stuart McMenamin, Ph.D Managing Director, Itron Forecasting 2008, Itron Inc. All rights reserved. 1 Introduction
More informationSTATISTICAL LOAD MODELING
STATISTICAL LOAD MODELING Eugene A. Feinberg, Dora Genethliou Department of Applied Mathematics and Statistics State University of New York at Stony Brook Stony Brook, NY 11794-3600, USA Janos T. Hajagos
More informationWind Assessment & Forecasting
Wind Assessment & Forecasting GCEP Energy Workshop Stanford University April 26, 2004 Mark Ahlstrom CEO, WindLogics Inc. mark@windlogics.com WindLogics Background Founders from supercomputing industry
More informationThis paper presents the
ISESCO JOURNAL of Science and Technology Volume 8 - Number 14 - November 2012 (2-8) A Novel Ensemble Neural Network based Short-term Wind Power Generation Forecasting in a Microgrid Aymen Chaouachi and
More informationpeak half-hourly New South Wales
Forecasting long-term peak half-hourly electricity demand for New South Wales Dr Shu Fan B.S., M.S., Ph.D. Professor Rob J Hyndman B.Sc. (Hons), Ph.D., A.Stat. Business & Economic Forecasting Unit Report
More informationThe Kentucky Mesonet: Entering a New Phase
The Kentucky Mesonet: Entering a New Phase Stuart A. Foster State Climatologist Kentucky Climate Center Western Kentucky University KCJEA Winter Conference Lexington, Kentucky February 9, 2017 Kentucky
More informationWEATHER DEPENENT ELECTRICITY MARKET FORECASTING WITH NEURAL NETWORKS, WAVELET AND DATA MINING TECHNIQUES. Z.Y. Dong X. Li Z. Xu K. L.
WEATHER DEPENENT ELECTRICITY MARKET FORECASTING WITH NEURAL NETWORKS, WAVELET AND DATA MINING TECHNIQUES Abstract Z.Y. Dong X. Li Z. Xu K. L. Teo School of Information Technology and Electrical Engineering
More informationVariables For Each Time Horizon
Variables For Each Time Horizon Andy Sukenik Itron s Forecasting Brown Bag Seminar December 13th, 2011 Please Remember In order to help this session run smoothly, your phones are muted. To make the presentation
More informationNeural Network to Control Output of Hidden Node According to Input Patterns
American Journal of Intelligent Systems 24, 4(5): 96-23 DOI:.5923/j.ajis.2445.2 Neural Network to Control Output of Hidden Node According to Input Patterns Takafumi Sasakawa, Jun Sawamoto 2,*, Hidekazu
More informationMachine Learning with Neural Networks. J. Stuart McMenamin, David Simons, Andy Sukenik Itron, Inc.
Machine Learning with Neural Networks J. Stuart McMenamin, David Simons, Andy Sukenik Itron, Inc. Please Remember» Phones are Muted: In order to help this session run smoothly, your phones are muted.»
More informationNeural-wavelet Methodology for Load Forecasting
Journal of Intelligent and Robotic Systems 31: 149 157, 2001. 2001 Kluwer Academic Publishers. Printed in the Netherlands. 149 Neural-wavelet Methodology for Load Forecasting RONG GAO and LEFTERI H. TSOUKALAS
More informationWIND SPEED ESTIMATION IN SAUDI ARABIA USING THE PARTICLE SWARM OPTIMIZATION (PSO)
WIND SPEED ESTIMATION IN SAUDI ARABIA USING THE PARTICLE SWARM OPTIMIZATION (PSO) Mohamed Ahmed Mohandes Shafique Rehman King Fahd University of Petroleum & Minerals Saeed Badran Electrical Engineering
More informationHighly-accurate Short-term Forecasting Photovoltaic Output Power Architecture without Meteorological Observations in Smart Grid
Highly-accurate Short-term Forecasting Photovoltaic Output Power Architecture without Meteorological Observations in Smart Grid Jun Matsumoto, Daisuke Ishii, Satoru Okamoto, Eiji Oki and Naoaki Yamanaka
More informationANN based techniques for prediction of wind speed of 67 sites of India
ANN based techniques for prediction of wind speed of 67 sites of India Paper presentation in Conference on Large Scale Grid Integration of Renewable Energy in India Authors: Parul Arora Prof. B.K Panigrahi
More informationShort-term water demand forecast based on deep neural network ABSTRACT
Short-term water demand forecast based on deep neural network Guancheng Guo 1, Shuming Liu 2 1,2 School of Environment, Tsinghua University, 100084, Beijing, China 2 shumingliu@tsinghua.edu.cn ABSTRACT
More informationWind Energy Predictions of Small-Scale Turbine Output Using Exponential Smoothing and Feed- Forward Neural Network
Wind Energy Predictions of Small-Scale Turbine Output Using Exponential Smoothing and Feed- Forward Neural Network Zaccheus O. Olaofe 1, 2 1 ZakkWealth Energy 2 Faculty of Engineering and Built Environment,
More informationHousehold Energy Disaggregation based on Difference Hidden Markov Model
1 Household Energy Disaggregation based on Difference Hidden Markov Model Ali Hemmatifar (ahemmati@stanford.edu) Manohar Mogadali (manoharm@stanford.edu) Abstract We aim to address the problem of energy
More informationDevelopment of System for Supporting Lock Position Adjustment Work for Electric Point Machine
PAPER Development of System for Supporting Lock Position Adjustment Work for Electric Point Machine Nagateru IWASAWA Satoko RYUO Kunihiro KAWASAKI Akio HADA Telecommunications and Networking Laboratory,
More informationEffect of Weather on Uber Ridership
SAS and all other SAS Institute Inc. product or service names are registered trademarks or trademarks of SAS Institute Inc. in the USA and other countries. indicates USA registration. Other brand and product
More informationMonthly Long Range Weather Commentary Issued: APRIL 1, 2015 Steven A. Root, CCM, President/CEO
Monthly Long Range Weather Commentary Issued: APRIL 1, 2015 Steven A. Root, CCM, President/CEO sroot@weatherbank.com FEBRUARY 2015 Climate Highlights The Month in Review The February contiguous U.S. temperature
More informationShort Term Load Forecasting by Using ESN Neural Network Hamedan Province Case Study
119 International Journal of Smart Electrical Engineering, Vol.5, No.2,Spring 216 ISSN: 2251-9246 pp. 119:123 Short Term Load Forecasting by Using ESN Neural Network Hamedan Province Case Study Milad Sasani
More informationMulti-Plant Photovoltaic Energy Forecasting Challenge with Regression Tree Ensembles and Hourly Average Forecasts
Multi-Plant Photovoltaic Energy Forecasting Challenge with Regression Tree Ensembles and Hourly Average Forecasts Kathrin Bujna 1 and Martin Wistuba 2 1 Paderborn University 2 IBM Research Ireland Abstract.
More informationShort Term Solar Radiation Forecast from Meteorological Data using Artificial Neural Network for Yola, Nigeria
American Journal of Engineering Research (AJER) 017 American Journal of Engineering Research (AJER) eiss: 300847 piss : 300936 Volume6, Issue8, pp8389 www.ajer.org Research Paper Open Access Short Term
More information