Kishore Bingi

This book suggests the development of single and multi-layer fractional-order neural networks that incorporate fractional-order activation functions derived using fractional-order derivatives. Activation functions are essential in neural networks as they introduce nonlinearity, enabling the models to learn complex patterns in data. However, traditional activation functions have limitations such as non-differentiability, vanishing gradient problems, and inactive neurons at negative inputs, which can affect the performance of neural networks, especially for tasks involving intricate nonlinear dynamics. To address these issues, fractional-order derivatives from fractional calculus have been proposed. These derivatives can model complex systems with non-local or non-Markovian behavior. The aim is to improve wind power prediction accuracy using datasets from the Texas wind turbine and Jeju Island wind farm under various scenarios. The book explores the advantages of fractional-order activation functions in terms of robustness, faster convergence, and greater flexibility in hyper-parameter tuning. It includes a comparative analysis of single and multi-layer fractional-order neural networks versus conventional neural networks, assessing their performance based on metrics such as mean square error and coefficient of determination. The impact of using machine learning models to impute missing data on the performance of networks is also discussed. This book demonstrates the potential of fractional-order activation functions to enhance neural network models, particularly in predicting chaotic time series. The findings suggest that fractional-order activation functions can significantly improve accuracy and performance, emphasizing the importance of advancing activation function design in neural network analysis. Additionally, the book is a valuable teaching and learning resource for undergraduate and postgraduate students conducting research in this field.

This book suggests the development of single and multi-layer fractional-order neural networks that incorporate fractional-order activation functions derived using fractional-order derivatives. Activation functions are essential in neural networks as they introduce nonlinearity, enabling the models to learn complex patterns in data. However, traditional activation functions have limitations such as non-differentiability, vanishing gradient problems, and inactive neurons at negative inputs, which can affect the performance of neural networks, especially for tasks involving intricate nonlinear dynamics. To address these issues, fractional-order derivatives from fractional calculus have been proposed. These derivatives can model complex systems with non-local or non-Markovian behavior. The aim is to improve wind power prediction accuracy using datasets from the Texas wind turbine and Jeju Island wind farm under various scenarios. The book explores the advantages of fractional-order activation functions in terms of robustness, faster convergence, and greater flexibility in hyper-parameter tuning. It includes a comparative analysis of single and multi-layer fractional-order neural networks versus conventional neural networks, assessing their performance based on metrics such as mean square error and coefficient of determination. The impact of using machine learning models to impute missing data on the performance of networks is also discussed. This book demonstrates the potential of fractional-order activation functions to enhance neural network models, particularly in predicting chaotic time series. The findings suggest that fractional-order activation functions can significantly improve accuracy and performance, emphasizing the importance of advancing activation function design in neural network analysis. Additionally, the book is a valuable teaching and learning resource for undergraduate and postgraduate students conducting research in this field.

This book presents the study to design, develop, and implement improved PI control techniques using dead-time compensation, structure enhancements, learning functions and fractional ordering parameters. Two fractional-order PI controllers are proposed and designed: fractional-order predictive PI and hybrid iterative learning based fractional-order predictive PI controller. Furthermore, the proposed fractional-order control strategies and filters are simulated over first- and second-order benchmark process models and further validated using the real-time experimentation of the pilot pressure process plant. In this book, five chapters are structured with a proper sequential flow of details to provide a better understanding for the readers. A general introduction to the controllers, filters and optimization techniques is presented in Chapter 1. Reviews of the PI controllers family and their modifications are shown in the initial part of Chapter 2, followed by the development of the proposed fractional-order predictive PI (FOPPI) controller with dead-time compensation ability. In the first part of chapter 3, a review of the PI based iterative learning controllers, modified structures of the ILC and their modifications are presented. Then, the design of the proposed hybrid iterative learning controller-based fractional-order predictive PI controller based on the current cyclic feedback structure is presented. Lastly, the results and discussion of the proposed controller on benchmark process models and the real-time experimentation of the pilot pressure process plant are given. Chapter 4 presents the development of the proposed filtering techniques and their performance comparison with the conventional methods. Chapter 5 proposes the improvement of the existing sine cosine algorithm (SCA) and arithmetic optimization algorithm (AOA) to form a novel arithmetic-trigonometric optimization algorithm (ATOA) to accelerate the rate of convergence in lesser iterations with mitigation towards getting caught in the same local position. The performance analysis of the optimization algorithm will be carried out on benchmark test functions and the real-time pressure process plant.

This book presents the study to design, develop, and implement improved PI control techniques using dead-time compensation, structure enhancements, learning functions and fractional ordering parameters. Two fractional-order PI controllers are proposed and designed: fractional-order predictive PI and hybrid iterative learning based fractional-order predictive PI controller. Furthermore, the proposed fractional-order control strategies and filters are simulated over first- and second-order benchmark process models and further validated using the real-time experimentation of the pilot pressure process plant. In this book, five chapters are structured with a proper sequential flow of details to provide a better understanding for the readers. A general introduction to the controllers, filters and optimization techniques is presented in Chapter 1. Reviews of the PI controllers family and their modifications are shown in the initial part of Chapter 2, followed by the development of the proposed fractional-order predictive PI (FOPPI) controller with dead-time compensation ability. In the first part of chapter 3, a review of the PI based iterative learning controllers, modified structures of the ILC and their modifications are presented. Then, the design of the proposed hybrid iterative learning controller-based fractional-order predictive PI controller based on the current cyclic feedback structure is presented. Lastly, the results and discussion of the proposed controller on benchmark process models and the real-time experimentation of the pilot pressure process plant are given. Chapter 4 presents the development of the proposed filtering techniques and their performance comparison with the conventional methods. Chapter 5 proposes the improvement of the existing sine cosine algorithm (SCA) and arithmetic optimization algorithm (AOA) to form a novel arithmetic-trigonometric optimization algorithm (ATOA) to accelerate the rate of convergence in lesser iterations with mitigation towards getting caught in the same local position. The performance analysis of the optimization algorithm will be carried out on benchmark test functions and the real-time pressure process plant.

Recent advances in wireless technology have led to the emergence of industry standards such as WirelessHART. These strategies minimise the need for cumbersome cabling, thereby reducing costs. However, applying them involves the challenge of handling stochastic network delays, which can degrade control performance. To address this problem, commonly used simple PID could be employed. However, PID suffers from gain range limitations when used in a delayed environment. Furthermore, model-based controllers are complex and require exact models of the process and systematic system identification for implementation. Therefore, to address these issues, the book proposes control strategies that retain the simplicity of PID in terms of ease of tuning and structure, while improving on the performance of the closed-loop system with regard to stochastic network delays and mismatches. Concretely, it proposes and discusses three strategies, namely: Setpoint Weighting (SW), Filtered Predictive PI (FPPI) and Optimal Fuzzy PID. In order to optimise some of these controllers, two novel hybrid optimisation algorithms combining the dynamism of the Bacterial Foraging Algorithm (BFA) and advantages of both the Spiral Dynamic Algorithm (SDA) and the Accelerated Particle Swarm Optimisation (APSO) have been used. The strategies proposed here can also be applied in stochastic control scenarios (not necessarily wireless) characterised by uncertainties. This book will be useful to engineers and researchers in both industry and academia. In industry, it will be particularly useful to research and development efforts where PID controllers and wireless sensor networks (WSNs) involving both short and long term stochastic network delay are employed. Thus, it can be used for real-time control design in these areas. In the academic setting, the book will be useful for researchers, undergraduate and graduate students of instrumentation and control. It can also be used as referencematerial for teaching courses on predictive and adaptive controls and their application.

This book presents a detailed study on fractional-order, set-point, weighted PID control strategies and the development of curve-fitting-based approximation techniques for fractional-order parameters. Furthermore, in all the cases, it includes the Scilab-based commands and functions for easy implementation and better understanding, and to appeal to a wide range of readers working with the software. The presented Scilab-based toolbox is the first toolbox for fractional-order systems developed in open-source software. The toolboxes allow time and frequency domains as well as stability analysis of the fractional-order systems and controllers. The book also provides real-time examples of the control of process plants using the developed fractional-order based PID control strategies and the approximation techniques. The book is of interest to readers in the areas of fractional-order controllers, approximation techniques, process modeling, control, and optimization, both in industry and academia. In industry, the book is particularly valuable in the areas of research and development (R&D) as well as areas where PID controllers suffice – and it should be noted that around 80% of low-level controllers in industry are PID based. The book is also useful where conventional PIDs are constrained, such as in industries where long-term delay and non-linearity are present. Here it can be used for the design of controllers for real-time processes. The book is also a valuable teaching and learning resource for undergraduate and postgraduate students.

Recent advances in wireless technology have led to the emergence of industry standards such as WirelessHART. These strategies minimise the need for cumbersome cabling, thereby reducing costs. However, applying them involves the challenge of handling stochastic network delays, which can degrade control performance. To address this problem, commonly used simple PID could be employed. However, PID suffers from gain range limitations when used in a delayed environment. Furthermore, model-based controllers are complex and require exact models of the process and systematic system identification for implementation. Therefore, to address these issues, the book proposes control strategies that retain the simplicity of PID in terms of ease of tuning and structure, while improving on the performance of the closed-loop system with regard to stochastic network delays and mismatches. Concretely, it proposes and discusses three strategies, namely: Setpoint Weighting (SW), Filtered Predictive PI (FPPI) and Optimal Fuzzy PID. In order to optimise some of these controllers, two novel hybrid optimisation algorithms combining the dynamism of the Bacterial Foraging Algorithm (BFA) and advantages of both the Spiral Dynamic Algorithm (SDA) and the Accelerated Particle Swarm Optimisation (APSO) have been used. The strategies proposed here can also be applied in stochastic control scenarios (not necessarily wireless) characterised by uncertainties. This book will be useful to engineers and researchers in both industry and academia. In industry, it will be particularly useful to research and development efforts where PID controllers and wireless sensor networks (WSNs) involving both short and long term stochastic network delay are employed. Thus, it can be used for real-time control design in these areas. In the academic setting, the book will be useful for researchers, undergraduate and graduate students of instrumentation and control. It can also be used as referencematerial for teaching courses on predictive and adaptive controls and their application.

This book presents a detailed study on fractional-order, set-point, weighted PID control strategies and the development of curve-fitting-based approximation techniques for fractional-order parameters. Furthermore, in all the cases, it includes the Scilab-based commands and functions for easy implementation and better understanding, and to appeal to a wide range of readers working with the software. The presented Scilab-based toolbox is the first toolbox for fractional-order systems developed in open-source software. The toolboxes allow time and frequency domains as well as stability analysis of the fractional-order systems and controllers. The book also provides real-time examples of the control of process plants using the developed fractional-order based PID control strategies and the approximation techniques. The book is of interest to readers in the areas of fractional-order controllers, approximation techniques, process modeling, control, and optimization, both in industry and academia. In industry, the book is particularly valuable in the areas of research and development (R&D) as well as areas where PID controllers suffice – and it should be noted that around 80% of low-level controllers in industry are PID based. The book is also useful where conventional PIDs are constrained, such as in industries where long-term delay and non-linearity are present. Here it can be used for the design of controllers for real-time processes. The book is also a valuable teaching and learning resource for undergraduate and postgraduate students.