Vadim Spiryaev

On the neural network approach for forecasting of nonstationary time series on the basis of the Hilbert-Huang transform

The two-stage adaptive approach for time series forecasting is proposed. The first stage involves the decomposition of the initial time series into basis functions and application to them of the Hilbert transform. At the second stage the obtained functions and their instantaneous amplitudes are used as input variables of neural network forecasting. The efficiency of the developed approach is displayed in real time series in the electric power problem of forecasting the sharply variable implementations of active power flows.

Optimal Training of Artificial Neural Networks to Forecast Power System State Variables

The problem of forecasting state variables of electric power system is studied. The paper suggests data-driven adaptive approach based on hybrid-genetic algorithm which combines the advantages of genetic algorithm and simulated annealing algorithm. The proposed method has two stages. At the first stage the input signal is decomposed into orthogonal basis functions based on the Hilbert-Huang transform. The genetic algorithm and simulated annealing algorithm are applied to optimal training of the artificial neural network and support vector machine at the second stage. The results of applying the developed approach for the short-term forecasts of active power flows in the electric networks are presented. The best efficiency of proposed approach is demonstrated on real retrospective data of active power flow forecast using the hybrid-genetic support vector machine algorithm.

Modeling of Nonlinear Dynamic Systems with Volterra Polynomials: Elements of Theory and Applications

The paper presents a review of the studies that were conducted at Energy Systems Institute ESI SB RAS in the field of mathematical modeling of nonlinear input-output dynamic systems with Volterra polynomials. The first part presents an original approach to identification of the Volterra kernels. The approach is based on setting special multi-parameter families of piecewise constant test input signals. It also includes a description of the respective software; presents illustrative calculations on the example of a reference dynamic system as well as results of computer modeling of real heat exchange processes. The second part of the review is devoted to the Volterra polynomial equations of the first kind. Studies of such equations were pioneered and have been carried out in the past decade by the laboratory of ill-posed problems at ESI SB RAS. A special focus in the paper is made on the importance of the Lambert function for the theory of these equations.

Application of two stages adaptive neural network approach for short-term forecast of electric power systems

The paper presents the two-stapes adaptive approach for short-term forecast of parameters of expected operating conditions. The first stage involves decomposition of the time series into intrinsic modal functions and subsequent application of the Hilbert transform. During the second stage the computed modal functions and amplitudes are employed as input functions for artificial neural networks. Their optimal combinations is constructed using methods of simulated annealing and neural-genetic input selection approach. The efficiency of developed approach is demonstrated on real time the problem of forecasting power flow and voltage level.

Electricity prices neural networks forecast using the Hilbert-Huang transform

The problem of forecasting of electicity prices is addressed in terms of joint approach employing the general regression artificial neural network and empirical mode decomposition approaches (EMD) which is part of Hilbert-Huang transform. The application of developed approach to day-ahead hourly time series has demonstrated the whole accuracy increase as well as peaks prediction.

Hybrid genetic algorithms for forecasting power systems state variables

A problem of forecasting state variables of electric power system is studied. The paper suggests data-driven adaptive approach based on hybrid-genetic algorithm which combines the advantages of genetic algorithm and simulated annealing algorithm. The input signal is decomposed into orthogonal basis functions using the Hilbert-Huang transform. The hybrid-genetic algorithm is applied to optimal training of the support vector machine and artificial neural network. The results of applying the developed approach for the short-term forecasts of active power flows in the electric networks are presented. The best efficiency of proposed approach is demonstrated on real retrospective data of active power flow forecast using the hybrid-genetic support vector machine algorithm empowered with the Hilbert-Huang transform.

Operating conditions forecasting for monitoring and control of electric power systems

Two approaches are proposed for short-term forecast of the parameters of expected operating conditions. The Kalman filter based algorithms and the modern technologies of an artificial intelligence and nonlinear optimization algorithms are employed for dynamical state estimation. The new approach combining the artificial neural networks and the Hilbert-Huang transform is designed in order to increase the accuracy of operating conditions forecasting. Numerical experiments on real time series have demonstrated the improvement of the prediction.

Forecasting nonstationary time series based on Hilbert-Huang transform and machine learning

We propose a modification of the adaptive approach to time series forecasting. On the first stage, the original signal is decomposed with respect to a special empirical adaptive orthogonal basis, and the Hilbert’s integral transform is applied. On the second stage, the resulting orthogonal functions and their instantaneous amplitudes are used as input variables for the machine learning unit that employs a hybrid genetic algorithm to train an artificial neural network and a regressive model based on support vector machines. The efficiency of the proposed approach is demonstrated on real data coming from Nord Pool Spot and Australian National Energy Market.

A hybrid wind speed forecasting strategy based on Hilbert-Huang transform and machine learning algorithms

Precise wind resource assessment is one of the more imminent challenges. In the present work, we develop an adaptive approach to wind speed forecasting. The approach is based on a combination of the efficient apparatus of non-stationary time series of wind speed retrospective data analysis based on the Hilbert-Huang transform and machine learning models. Models that are examined include neural networks, support vector machines, the regression trees approach: random forest and boosting trees. Evaluation results are presented for the Irish power system based on the Atlantic offshore buoy data.