National Supercomputing Mission

PI(s): Dr. Kishalay Mitra / Dr. Soumya Jana

Scholars(s): Ms. Keerthi NagaSree Pujari, Mr. Subhasish Saikia, Dr. Pantula Devi Priyanka, Dr. Srinivas Soumitri, Dr. Inapakurthi Ravi Kiran, Dr. Kapil Gumte, Dr. Ankita Sharma, Dr. Anjali Dwivedi, Ms. Aswitha Tadepalli, Mr. Vishal N Mate, Ms. Pooja N Muley, Mr. Agathiyan Sushil R, Mr. Shivam Yadav, Mr. Samyak Bahuguna 

Objectives

• Big Data and analysis: Collection of wind speed data and preprocessing

• Machine Learning: Training machine learning models and optimizing them for accurate prediction overcoming overfitting

• Reduced order Wake modelling: Construction of LES based wake models and building ANN surrogates using the high-fidelity data

• MINLP reformulation of wind farm layout optimization using machine learning

• Realistic Windfarm design and control: Consider varying wind frequency distribution functions using time series models, and solving layout oprimization using Robust Optimization methods. Subsequently, perform real-time windfarm control using machine learning.

Summary

• Optimization of hyperparameters in NAR models, wavelet networks, and LSTM networks and VAEs using a multi-objective optimization framework through evolutionary Neural Architecture Search (NAS) strategy. 

• Contributing to Green Deep Learning through the NAS strategy by reducing the number of parameters and thereby computations. 

• Validating the prediction scope of optimal NAR models, wavelet networks, and LSTMs with real wind characteristics data. 

• Learning the distribution of wind characteristics using a Wind Frequency Map (WFM) which is a joint probability mass function of wind speed and wind direction. The hidden distribution in the data is modelled by machine learning with several WFMs which will be helpful in generating more such WFMs through the learnt distribution.

• Comparison among various leading time-series modeling techniques for handling nonlinearities in wind data. 

• Method demonstrating the effective usage of past and forecasted data for accurate modeling of wind for sustainable production and management of wind energy.

• A new and more accurate data-driven analytical model for wake characterization.  A nonlinear wake expansion is considered in analytical wake models.  To model the nonlinear expansion of wake, Artificial Neural Networks are used. Such a hybrid model possesses not only the knowledge of the process using physics involved in the problem but also practical validation using the data driven modeling aspect of it.  

• The combined effect of wind flow direction and yaw misalignment on wind farm performance through Large Eddy Simulations.

• An analytical model having the effect of wind flow direction and yaw misalignment on wind farm performance.

• Reformulation of the wind farm layout optimization problem, a NP-hard MINLP problem. 

• A robust design of wind farm under the wind state uncertainty.

• Application of ML technique such as RL for effective control of wind farm

Publications

Journals

• Pujari, K. N., Miriyala, S. S., Mittal, P., & Mitra, K. Better wind forecasting using evolutionary neural architecture search driven green deep learning. Expert Systems with Applications, 2023, 214, 119063. 

• Ravi Kiran, I., Soumitri M. S., Mitra, K., Deep Learning Based Dynamic Behaviour Modelling and Prediction of Particulate Matter in Air, Chemical Engineering Journal, 2021, 426, 131221.

• Krishnan, K. J., Mitra, K., A Modified Kohonen Map Algorithm for Clustering Time Series Data, Expert Systems With Applications, 2022, 201, 117249.

• Soumitri M. S., Pujari, K. N., Naik, S., Mitra, K., Evolutionary Neural Architecture Search for Surrogate models to Enable Optimization of Industrial Continuous Crystallization Process, Powder Technology, 2022, 405, 117527.

• Ravi Kiran, I., Mitra, K., Optimal Surrogate Building Using SVR for an Industrial Grinding Process, Materials and Manufacturing Processes, 2022.

Conferences

• Pujari, N. K., Soumitri, M. S., Mitra, K., A Generative Adversarial Networks based Modelling for Efficient Design of Wind Energy Conversion Systems, 9th IEEE Indian Control Conference, Vizag, Dec 2023.

• Pujari, N. K., Mitra, K., Wind Farm Layout Optimization Under Uncertainty using Bayesian Approach, 9th IEEE Indian Control Conference, Vizag, Dec 2023.

• Tiwari, U., D., Ghaisas, N. S., Mitra, K., Studying the combined effect of yaw misalignment and wind flow direction on wind farm wake losses, 14th Asian Computational Fluid Dynamics Conference, Bangalore, Nov, 2023.

• Pujari, N. K., Miriyala, S. S., Mitra, K., Jensen-ANN: A Machine Learning adaptation of Jensen Wake Model, 22nd IFAC World Congress 2023, IFAC-PapersOnLine, 56, 2, 2023, 4651-4656.

• Tiwari, U. D., Ghaisas N., Mitra K., Large Eddy Simulations of Yaw Misaligned Wind Farms and Evaluation of Wake Deflection and Velocity Deficit Models, Thirteenth International Symposium on Turbulence and Shear Flow Phenomena (TSFP13), June 25-28, 2024, Montréal, Canada (accepted).

• Ravi Kiran, I., Mitra, K., System Identification and Process Modelling of Dynamic Systems Using Machine Learning, 26th IEEE International Conference on System Theory, Control and Computing (ICSTCC), Sinaia, Romania, October, 2022.

• Ravi Kiran, I., Mitra, K., Artificial Intelligence Assisted Optimization Under Uncertainty for Robust Solutions, 26th IEEE International Conference on System Theory, Control and Computing (ICSTCC), Sinaia, Romania, October, 2022.

• Lakshmi S. C., Miriyala, S. S., Mitra, K., Statistical Inference and Analysis for Efficient Modeling of Environmental Pollution using Deep Neural Networks, 2022 Eighth IEEE Indian Control Conference (ICC), Chennai, India, 2022, pp. 385-390, doi: 10.1109/ICC56513.2022.10093411.

• Pantula, D. P., Miriyala, S. S., Mitra, K., A Deep Unsupervised Learning Algorithm for Clustering of Wind Frequency Maps, 2022 Eighth IEEE Indian Control Conference (ICC), Chennai, India, 2022, pp. 361-366, doi: 10.1109/ICC56513.2022.10093581.

• Ravi Kiran, I., Mitra, K., Data Based Time Series Modelling of Industrial Grinding Circuits, International Conference on Advances in Data-driven Computing and Intelligent Systems (ADCIS 2022), Goa, INDIA, September 2022.

• Ravi Kiran, I., Mitra, K., Machine Learning Based Surrogate Assisted Multi-Objective Optimization of Continuous Casting Process, Seventh IEEE Indian Control Conference, Bombay, India, Dec 2021.

• Krishnan, K. J., Mitra, K. Clustering Time Series Sensor Data Using Modified Kohonen Maps, Seventh IEEE Indian Control Conference, Bombay, India, Dec 2021.

• Pantula, D. P., Miriyala, S. S., Mitra, K., A Deep Unsupervised Learning Algorithm for Dynamic Data Clustering, Seventh IEEE Indian Control Conference, Bombay, India, Dec 2021.

• Pujari, K., Srivastava, V., Miriyala, S. S., Mitra, K., Comparison of Deep Reinforcement Learning Techniques with Gradient Based Approach in Cooperative Control of Wind Farm, Seventh IEEE Indian Control Conference, Bombay, India, Dec 2021.

Book Chapters

• Pujari, N. K., Miriyala, S. S., Mitra, K., Comparative Study of Automated Deep Learning Techniques for Wind Time Series Forecasting in “Statistical Modeling in Machine Learning”, Editors: Goswami, T. and Sinha, G. R., Elsevier, 2022, 327-356.

Book

• "Optimization, Uncertainty & Machine Learning in Wind Energy Conversion Systems” edited by Prof(s) K. Mitra, R. Everson and J. Fieldsend, Springer under the book series, “Engineering Optimization” (in press).