Data-driven demand forecasting method for fused magnesium furnaces

The demand of fused magnesium furnaces (FMFs) refers to the average value of the power of the FMFs over a fixed period of time before the current time. The demand is an indicator of the electricity consumption of high energy-consuming FMFs. When the demand exceeds the limit of the Peak Demand (a predetermined maximum demand), the power supply of some FMF will be cut off to ensure that the demand is no more than Peak Demand. But the power cutoff will destroy the heat balance, reduce the quality and yield of the product. The composition change of magnesite in FMFs will cause demand spike occasionally, which a sudden increase in demand exceeds the limit and then drops below the limit. As a result, demand spike cause the power cutoff. In order to avoid the power cutoff at the moment of demand spike, the demand of FMFs needs to be forecasted. This paper analyzes the dynamic model of the demand of FMFs, using the power data, presents a data-driven demand forecasting method. This method consists of the following: PACF based decision module for the number of the input variables of the forecasting model, RBF neural network (RBFNN) based power variation rate forecasting model and demand forecasting model. Simulations based on actual data and industrial experiments at a fused magnesia plant show the effectiveness of the proposed method.

[1]  O. Ozgun,et al.  Flicker Study Using a Novel ARC Furnace Model , 2002, IEEE Power Engineering Review.

[2]  Theofanis Sapatinas,et al.  Short-Term Load Forecasting: The Similar Shape Functional Time-Series Predictor , 2012, IEEE Transactions on Power Systems.

[3]  Dingli Yu,et al.  Selecting radial basis function network centers with recursive orthogonal least squares training , 2000, IEEE Trans. Neural Networks Learn. Syst..

[4]  H. Saunders,et al.  Mechanical Engineering Design (4th Ed.) , 1985 .

[5]  Liu Ning.  Application of Time Series Analysis in Securities Analysis , 2008 .

[6]  Jooyoung Park,et al.  Universal Approximation Using Radial-Basis-Function Networks , 1991, Neural Computation.

[7]  B. Araabi,et al.  Short-Term Load Forecasting With a New Nonsymmetric Penalty Function , 2011, IEEE Transactions on Power Systems.

[8]  Joseph Edward Shigley,et al.  Mechanical engineering design , 1972 .

[9]  Tianyou Chai,et al.  Data-Driven Abnormal Condition Identification and Self-Healing Control System for Fused Magnesium Furnace , 2015, IEEE Transactions on Industrial Electronics.

[10]  Tianyou Chai,et al.  Data-Driven Optimization Control for Safety Operation of Hematite Grinding Process , 2015, IEEE Transactions on Industrial Electronics.

[11]  Hao Yu,et al.  Fast and Efficient Second-Order Method for Training Radial Basis Function Networks , 2012, IEEE Transactions on Neural Networks and Learning Systems.

[12]  Vladimir Ceperic,et al.  A Strategy for Short-Term Load Forecasting by Support Vector Regression Machines , 2013, IEEE Transactions on Power Systems.

[13]  Abbas Khosravi,et al.  Short-Term Load and Wind Power Forecasting Using Neural Network-Based Prediction Intervals , 2014, IEEE Transactions on Neural Networks and Learning Systems.

[14]  M.N.S. Swamy,et al.  Radial Basis Function Networks , 2014 .

[15]  Pierluigi Siano,et al.  A Novel RBF Training Algorithm for Short-Term Electric Load Forecasting and Comparative Studies , 2015, IEEE Transactions on Industrial Electronics.

[16]  Shang-Liang Chen,et al.  Orthogonal least squares learning algorithm for radial basis function networks , 1991, IEEE Trans. Neural Networks.

[17]  Hao Yu,et al.  Advantages of Radial Basis Function Networks for Dynamic System Design , 2011, IEEE Transactions on Industrial Electronics.