Seshadri Sivakumar - Academia.edu (original) (raw)
Papers by Seshadri Sivakumar
LESCOPE'02. 2002 Large Engineering Systems Conference on Power Engineering. Conference Proceedings
ABSTRACT This paper presents the optimized conditions for stepped wave inverters in order to mini... more ABSTRACT This paper presents the optimized conditions for stepped wave inverters in order to minimize the total harmonic distortion present in the output waveform. It is shown that one- and two-level waveforms do not have sufficient degrees of freedom to meet harmonic standards like the IEEE 519. However three-level waveforms can be constructed, which meet such power quality standards and as a result, do not require output filtering.
CCECE '97. Canadian Conference on Electrical and Computer Engineering. Engineering Innovation: Voyage of Discovery. Conference Proceedings
Cusp distortion in line current is present in power factor correcting boost converters for a dura... more Cusp distortion in line current is present in power factor correcting boost converters for a duration of time around the zero crossing of the current. A shunt active power electronic filter and its control to minimize the cusp distortion is considered in this paper. It is shown by ideal switch simulations that the topology considered and the control could be used to reduce the cusp distortion problem.
L'invention concerne un generateur de puissance a haute frequence (HF) qui inclut un premier ... more L'invention concerne un generateur de puissance a haute frequence (HF) qui inclut un premier amplificateur a decoupage, qui genere un premier signal HF conformement a un premier signal de commande, et un second amplificateur a decoupage, qui genere un second signal HF conformement a un second signal de commande. Les premier et second signaux de commande determinent une difference de phase entre les premier et second signaux HF. L'enveloppe du signal de sortie est fondee sur les premier et second signaux HF et sur la difference de phase. Les premier et second signaux de commande font alterner les phases des premier et second signaux HF.
2009 IEEE MTT-S International Microwave Symposium Digest, 2009
The complete design of phase controlled Class E amplifiers for pulsing applications is given. Cla... more The complete design of phase controlled Class E amplifiers for pulsing applications is given. Class E amplifier operating at high frequency (HF) range is designed, simulated, implemented and tested using the introduced method. Amplifier is pulsed up to 100 kHz at the rated power with duty cycles ranging from 10 percent to 90 percent. Experimental results are found to be
IEEE Transactions on Circuits and Systems I: Regular Papers, 2010
An analysis of the Class-E power amplifier is given using harmonic modeling that facilitates its ... more An analysis of the Class-E power amplifier is given using harmonic modeling that facilitates its characterization for a wide load-space regime expected in applications such as RF plasma generation for semiconductor manufacturing. Using a hybrid matrix-framework made up of the dominant harmonics of the circuit variables aided by the time-domain boundary conditions of the power switch voltage waveform, a quasi-linear algorithm is developed to accurately determine the voltage and the current levels expected on all components of the power amplifier for arbitrary loading and operating conditions. The application of the algorithm for design optimization is discussed, and its extension to phase-controlled Class-E power amplifier pairs is described.
OCEANS 2018 MTS/IEEE Charleston, 2018
This paper describes the hydrodynamic, pneumatic, and electrical modeling and simulation study pe... more This paper describes the hydrodynamic, pneumatic, and electrical modeling and simulation study performed in the development of the Healy Wave Energy Converter (HWEC). The HWEC, based on a two-body structure, consists of a flotation buoy rigidly connected to a long vertical inertial tube which is open at the top and bottom. A piston-rod assembly is enclosed and connected to the power take-off (PTO) mechanism. Due to inertia of water within the tube, relative motion between the piston and the buoy-inertia tube structure drives the PTO, which consists of an air compressor, a dynamic pressure regulator and a turbine connected to an induction generator supported by a power electronic conversion system. The induction generator is operated in both self-excitation mode using switched capacitors and external excitation mode using an inverter to generate flexible power forms and to maximize the energy yield over a wide range of wave conditions. A PLC based control system ensures balance between the generated and consumed power under all dynamic conditions. Nonlinear time-domain numerical models were developed for each subsystem. The resulting modeling framework was used to specify component requirements, select operational parameters, and evaluate system performance. The HWEC system is designed to output moderate power with high reliability. The present analysis demonstrates that the HWEC can produce and store energy over a wide range of sea states, making it highly suitable for off-grid and micro-grid applications.
IFAC Proceedings Volumes, 1983
This paper presents a digital simulation study of the use of a static thyristor-controlled phase ... more This paper presents a digital simulation study of the use of a static thyristor-controlled phase shifter as an effective power controller to enhance system stability and to damp the line electromechanical oscillations, hence to improve AC tie line capacity. A dynamic model of the static phase shifter is incorporated in the integrated power system dynamics; the unified state space form is used to investigate digital and analog control laws to transform static phase shifters into viable tools in enhancing interconnected power system steady state stability.
PESC '91 Record 22nd Annual IEEE Power Electronics Specialists Conference
The design of a controller for the series resonant converter (SRC) is presented with an objective... more The design of a controller for the series resonant converter (SRC) is presented with an objective to force the resonant state variables (the inductor current and the capacitor voltage) to track their optimal trajectory. The inherently nonlinear optimal trajectory control strategy is analytically shown equivalent to a modified capacitor voltage control with a pseudo-linear form. The resulting structure is shown
Proceedings of 1994 IEEE Workshop on Computers in Power Electronics
This paper describes an optimization procedure for the design of fixed-point digital filters for ... more This paper describes an optimization procedure for the design of fixed-point digital filters for the high speed measurement of low frequency AC signals encountered in uninterruptible power supplies (UPS). The proposed procedure addresses the minimization of the noise effects due to roundoff, analog to digital conversion quantization and filter coefficient quantization together with filter performance requirements such as speed of response.
Proceedings of 1994 IEEE Workshop on Computers in Power Electronics
This paper describes a digital control scheme for a three-phase power factor correcting rectifier... more This paper describes a digital control scheme for a three-phase power factor correcting rectifier using template modulated PWM signals. The control strategy is based on the notion that the three-phase line currents can be forced to trace a deterministic waveform such as a sinusoid by considering its implicit model in the controller structure. The modulating template waveform is shown to be based on segments of a tangent waveform and is generated using the line voltage, the rectifier output voltage and the load current. The paper provides the analysis and the design of the controller and presents simulation and preliminary implementation results to demonstrate its effectiveness
Proceedings of IEEE Power Electronics Specialist Conference - PESC '93
This paper proposes a new scheme for high speed measurement and processing of AC voltages encount... more This paper proposes a new scheme for high speed measurement and processing of AC voltages encountered in off-line uninterruptible power supplies (UPS) based on the design of a Kalman fiiter with an internal model of the sinusoidal waveform. The proposed scheme provides a common framework to monitor the UPS mains input supply, to sense its out-of-specifuation voltage, frequency conditions and to measure the UPS inverter output voltage for its feedback control. Issues related to the accuracy of measurement with finite wordlength implementation are addressed. The superior performance of the scheme is demonstrated through a practical implementation of a high power quality UPS design used onboard ships to supply critical loads.
Proceedings of IEEE Power Electronics Specialist Conference - PESC '93
An optimal trajectory control (OTC) scheme is proposed for the constant frequency series resonant... more An optimal trajectory control (OTC) scheme is proposed for the constant frequency series resonant power converter (CF-SRC). The design of the control scheme is based on the graphical state-plane analysis of the CF-SRC and is aimed at determining the switching instants which force the tank circuit voltage and current to track their optimal trajectory for a given switching cycle. This approach provides a suitable framework to achieve the lowest possible response time with guaranteed stability. Using a near-linear form of the control law, the OTC is implemented to control a CF-SRC. Results are presented to demonstrate the superior performance of the proposed control scheme
IEEE Transactions on Power Electronics, 1995
This paper proposes a new control method for the constant-frequency control of power factor corre... more This paper proposes a new control method for the constant-frequency control of power factor correcting boost power converter using a sinewave template modulated PWM signal which eliminates the need for instantaneous measurement of the line current for the switching control of the boost converter. The control strategy is based on the notion that the line current can be forced to
IEEE Transactions on Energy Conversion, 1987
power flow andi simultaneously keep the lagging reactive power burden on the generator This paper... more power flow andi simultaneously keep the lagging reactive power burden on the generator This paper deals with dynamic modelling of to a minimum. self-excited induction generator connected to a supply system through a DC link (converter-A nonlinear model is developed by using line commutated inverter) and digital control the d-q model of induction generator and indesign for regulating the power transfer terfacisng to it the models of the subsystem4 through the link. Using different sampling viz. capacitors, converter, d.c link and inrates for the converter and inverter control verter appropriately. The objective of the loops, the proposed design ensures reduced modelling excludes the self-excitation proreactive power burden on the induction genercess of the induction generator and is aimed ator. Results of implementation on a labora-only at the bahaviour after self-excitation4 tory system are given. The validity of the developed model is experimentally tested using a prototype system.
IEEE Access
Autonomous generation of time series is challenging because the network must capture short-term f... more Autonomous generation of time series is challenging because the network must capture short-term features while tracking long-term time dependencies. This paper introduces the modulation of the activation function slopes of the upper-lower triangular recurrent neural networks (ULTRNNs) for dynamic variation of memory through a secondary recurrent network with its own independent states. A zigzag propagation algorithm for weight updates is proposed that accounts for the dynamic interaction of the states between the ULTRNN and the secondary network. A novel training method is proposed that distributes the eigenvalues of the closed-loop system around the unit circle in the complex z-plane to ensure that the network behaves as a nonlinear oscillator with an output that neither collapses nor saturates but continues to emulate the target. Examples encompassing the Lorenz series, Santa Fe laser data, kolam patterns, electrocardiogram (ECG) signals, stock pricing data, and smart grid data are presented to demonstrate that the proposed approach is highly effective in the generative modeling of complex periodic, chaotic, and nonstationary time series. The qualitative and quantitative performance of the ULTRNN obtained with the proposed activation-function modulation technique is comparable to that of state-of-the-art techniques including feedforward networks and generative adversarial networks, but with far fewer trainable parameters and shorter computation times.
2019 International Joint Conference on Neural Networks (IJCNN), 2019
This paper introduces a novel method to dynamically vary the activation function slopes in recent... more This paper introduces a novel method to dynamically vary the activation function slopes in recently developed upper and lower triangular recurrent neural networks (ULTRNN) to enhance their modeling capability. The ULTRNN employs a pair of triangular feedback weight matrices with block diagonal elements whose eigenvalues are constrained to lie on or close to the unit circle in the complex z-plane to maintain network and training stability. The activation function slopes of the ULTRNN state variables are dynamically varied by a second modulating network. The inputs to the modulating network are the state variables of the principal ULTRNNs and their inputs. The modulating network is trained simultaneously with the principal ULTRNN to compute the activation function slope for the latter’s each state variable at each time step. Such dynamic variation of the activation function slopes selectively enhances the contribution of certain states while suppressing that of the others. A larger slope results in a longer time contribution of the corresponding state and helps model long-term dependencies. Conversely, a smaller slope results in a shorter time contribution and may be used to model controlled "forgetting". The proposed modulation technique enhances the ULTRNN’s ability to effectively incorporate short-term memory and long-term dependencies. Simulation results show that with activation function modulation the ULTRNNs are able to autonomously replicate the outputs of sample chaotic dynamic system with good accuracy. This capability can be highly effective in modeling or characterizing the inherent process that generates the time series.
IEEE Transactions on Cybernetics, 2017
This paper introduces a discrete-time recurrent neural network architecture using triangular feed... more This paper introduces a discrete-time recurrent neural network architecture using triangular feedback weight matrices that allows a simplified approach to ensuring network and training stability. The triangular structure of the weight matrices is exploited to readily ensure that the eigenvalues of the feedback weight matrix represented by the block diagonal elements lie on the unit circle in the complex z-plane by updating these weights based on the differential of the angular error variable. Such placement of the eigenvalues together with the extended close interaction between state variables facilitated by the nondiagonal triangular elements, enhances the learning ability of the proposed architecture. Simulation results show that the proposed architecture is highly effective in time-series prediction tasks associated with nonlinear and chaotic dynamic systems with underlying oscillatory modes. This modular architecture with dual upper and lower triangular feedback weight matrices mimics fully recurrent network architectures, while maintaining learning stability with a simplified training process. While training, the block-diagonal weights (hence the eigenvalues) of the dual triangular matrices are constrained to the same values during weight updates aimed at minimizing the possibility of overfitting. The dual triangular architecture also exploits the benefit of parsing the input and selectively applying the parsed inputs to the two subnetworks to facilitate enhanced learning performance.
LESCOPE'02. 2002 Large Engineering Systems Conference on Power Engineering. Conference Proceedings
ABSTRACT This paper presents the optimized conditions for stepped wave inverters in order to mini... more ABSTRACT This paper presents the optimized conditions for stepped wave inverters in order to minimize the total harmonic distortion present in the output waveform. It is shown that one- and two-level waveforms do not have sufficient degrees of freedom to meet harmonic standards like the IEEE 519. However three-level waveforms can be constructed, which meet such power quality standards and as a result, do not require output filtering.
CCECE '97. Canadian Conference on Electrical and Computer Engineering. Engineering Innovation: Voyage of Discovery. Conference Proceedings
Cusp distortion in line current is present in power factor correcting boost converters for a dura... more Cusp distortion in line current is present in power factor correcting boost converters for a duration of time around the zero crossing of the current. A shunt active power electronic filter and its control to minimize the cusp distortion is considered in this paper. It is shown by ideal switch simulations that the topology considered and the control could be used to reduce the cusp distortion problem.
L'invention concerne un generateur de puissance a haute frequence (HF) qui inclut un premier ... more L'invention concerne un generateur de puissance a haute frequence (HF) qui inclut un premier amplificateur a decoupage, qui genere un premier signal HF conformement a un premier signal de commande, et un second amplificateur a decoupage, qui genere un second signal HF conformement a un second signal de commande. Les premier et second signaux de commande determinent une difference de phase entre les premier et second signaux HF. L'enveloppe du signal de sortie est fondee sur les premier et second signaux HF et sur la difference de phase. Les premier et second signaux de commande font alterner les phases des premier et second signaux HF.
2009 IEEE MTT-S International Microwave Symposium Digest, 2009
The complete design of phase controlled Class E amplifiers for pulsing applications is given. Cla... more The complete design of phase controlled Class E amplifiers for pulsing applications is given. Class E amplifier operating at high frequency (HF) range is designed, simulated, implemented and tested using the introduced method. Amplifier is pulsed up to 100 kHz at the rated power with duty cycles ranging from 10 percent to 90 percent. Experimental results are found to be
IEEE Transactions on Circuits and Systems I: Regular Papers, 2010
An analysis of the Class-E power amplifier is given using harmonic modeling that facilitates its ... more An analysis of the Class-E power amplifier is given using harmonic modeling that facilitates its characterization for a wide load-space regime expected in applications such as RF plasma generation for semiconductor manufacturing. Using a hybrid matrix-framework made up of the dominant harmonics of the circuit variables aided by the time-domain boundary conditions of the power switch voltage waveform, a quasi-linear algorithm is developed to accurately determine the voltage and the current levels expected on all components of the power amplifier for arbitrary loading and operating conditions. The application of the algorithm for design optimization is discussed, and its extension to phase-controlled Class-E power amplifier pairs is described.
OCEANS 2018 MTS/IEEE Charleston, 2018
This paper describes the hydrodynamic, pneumatic, and electrical modeling and simulation study pe... more This paper describes the hydrodynamic, pneumatic, and electrical modeling and simulation study performed in the development of the Healy Wave Energy Converter (HWEC). The HWEC, based on a two-body structure, consists of a flotation buoy rigidly connected to a long vertical inertial tube which is open at the top and bottom. A piston-rod assembly is enclosed and connected to the power take-off (PTO) mechanism. Due to inertia of water within the tube, relative motion between the piston and the buoy-inertia tube structure drives the PTO, which consists of an air compressor, a dynamic pressure regulator and a turbine connected to an induction generator supported by a power electronic conversion system. The induction generator is operated in both self-excitation mode using switched capacitors and external excitation mode using an inverter to generate flexible power forms and to maximize the energy yield over a wide range of wave conditions. A PLC based control system ensures balance between the generated and consumed power under all dynamic conditions. Nonlinear time-domain numerical models were developed for each subsystem. The resulting modeling framework was used to specify component requirements, select operational parameters, and evaluate system performance. The HWEC system is designed to output moderate power with high reliability. The present analysis demonstrates that the HWEC can produce and store energy over a wide range of sea states, making it highly suitable for off-grid and micro-grid applications.
IFAC Proceedings Volumes, 1983
This paper presents a digital simulation study of the use of a static thyristor-controlled phase ... more This paper presents a digital simulation study of the use of a static thyristor-controlled phase shifter as an effective power controller to enhance system stability and to damp the line electromechanical oscillations, hence to improve AC tie line capacity. A dynamic model of the static phase shifter is incorporated in the integrated power system dynamics; the unified state space form is used to investigate digital and analog control laws to transform static phase shifters into viable tools in enhancing interconnected power system steady state stability.
PESC '91 Record 22nd Annual IEEE Power Electronics Specialists Conference
The design of a controller for the series resonant converter (SRC) is presented with an objective... more The design of a controller for the series resonant converter (SRC) is presented with an objective to force the resonant state variables (the inductor current and the capacitor voltage) to track their optimal trajectory. The inherently nonlinear optimal trajectory control strategy is analytically shown equivalent to a modified capacitor voltage control with a pseudo-linear form. The resulting structure is shown
Proceedings of 1994 IEEE Workshop on Computers in Power Electronics
This paper describes an optimization procedure for the design of fixed-point digital filters for ... more This paper describes an optimization procedure for the design of fixed-point digital filters for the high speed measurement of low frequency AC signals encountered in uninterruptible power supplies (UPS). The proposed procedure addresses the minimization of the noise effects due to roundoff, analog to digital conversion quantization and filter coefficient quantization together with filter performance requirements such as speed of response.
Proceedings of 1994 IEEE Workshop on Computers in Power Electronics
This paper describes a digital control scheme for a three-phase power factor correcting rectifier... more This paper describes a digital control scheme for a three-phase power factor correcting rectifier using template modulated PWM signals. The control strategy is based on the notion that the three-phase line currents can be forced to trace a deterministic waveform such as a sinusoid by considering its implicit model in the controller structure. The modulating template waveform is shown to be based on segments of a tangent waveform and is generated using the line voltage, the rectifier output voltage and the load current. The paper provides the analysis and the design of the controller and presents simulation and preliminary implementation results to demonstrate its effectiveness
Proceedings of IEEE Power Electronics Specialist Conference - PESC '93
This paper proposes a new scheme for high speed measurement and processing of AC voltages encount... more This paper proposes a new scheme for high speed measurement and processing of AC voltages encountered in off-line uninterruptible power supplies (UPS) based on the design of a Kalman fiiter with an internal model of the sinusoidal waveform. The proposed scheme provides a common framework to monitor the UPS mains input supply, to sense its out-of-specifuation voltage, frequency conditions and to measure the UPS inverter output voltage for its feedback control. Issues related to the accuracy of measurement with finite wordlength implementation are addressed. The superior performance of the scheme is demonstrated through a practical implementation of a high power quality UPS design used onboard ships to supply critical loads.
Proceedings of IEEE Power Electronics Specialist Conference - PESC '93
An optimal trajectory control (OTC) scheme is proposed for the constant frequency series resonant... more An optimal trajectory control (OTC) scheme is proposed for the constant frequency series resonant power converter (CF-SRC). The design of the control scheme is based on the graphical state-plane analysis of the CF-SRC and is aimed at determining the switching instants which force the tank circuit voltage and current to track their optimal trajectory for a given switching cycle. This approach provides a suitable framework to achieve the lowest possible response time with guaranteed stability. Using a near-linear form of the control law, the OTC is implemented to control a CF-SRC. Results are presented to demonstrate the superior performance of the proposed control scheme
IEEE Transactions on Power Electronics, 1995
This paper proposes a new control method for the constant-frequency control of power factor corre... more This paper proposes a new control method for the constant-frequency control of power factor correcting boost power converter using a sinewave template modulated PWM signal which eliminates the need for instantaneous measurement of the line current for the switching control of the boost converter. The control strategy is based on the notion that the line current can be forced to
IEEE Transactions on Energy Conversion, 1987
power flow andi simultaneously keep the lagging reactive power burden on the generator This paper... more power flow andi simultaneously keep the lagging reactive power burden on the generator This paper deals with dynamic modelling of to a minimum. self-excited induction generator connected to a supply system through a DC link (converter-A nonlinear model is developed by using line commutated inverter) and digital control the d-q model of induction generator and indesign for regulating the power transfer terfacisng to it the models of the subsystem4 through the link. Using different sampling viz. capacitors, converter, d.c link and inrates for the converter and inverter control verter appropriately. The objective of the loops, the proposed design ensures reduced modelling excludes the self-excitation proreactive power burden on the induction genercess of the induction generator and is aimed ator. Results of implementation on a labora-only at the bahaviour after self-excitation4 tory system are given. The validity of the developed model is experimentally tested using a prototype system.
IEEE Access
Autonomous generation of time series is challenging because the network must capture short-term f... more Autonomous generation of time series is challenging because the network must capture short-term features while tracking long-term time dependencies. This paper introduces the modulation of the activation function slopes of the upper-lower triangular recurrent neural networks (ULTRNNs) for dynamic variation of memory through a secondary recurrent network with its own independent states. A zigzag propagation algorithm for weight updates is proposed that accounts for the dynamic interaction of the states between the ULTRNN and the secondary network. A novel training method is proposed that distributes the eigenvalues of the closed-loop system around the unit circle in the complex z-plane to ensure that the network behaves as a nonlinear oscillator with an output that neither collapses nor saturates but continues to emulate the target. Examples encompassing the Lorenz series, Santa Fe laser data, kolam patterns, electrocardiogram (ECG) signals, stock pricing data, and smart grid data are presented to demonstrate that the proposed approach is highly effective in the generative modeling of complex periodic, chaotic, and nonstationary time series. The qualitative and quantitative performance of the ULTRNN obtained with the proposed activation-function modulation technique is comparable to that of state-of-the-art techniques including feedforward networks and generative adversarial networks, but with far fewer trainable parameters and shorter computation times.
2019 International Joint Conference on Neural Networks (IJCNN), 2019
This paper introduces a novel method to dynamically vary the activation function slopes in recent... more This paper introduces a novel method to dynamically vary the activation function slopes in recently developed upper and lower triangular recurrent neural networks (ULTRNN) to enhance their modeling capability. The ULTRNN employs a pair of triangular feedback weight matrices with block diagonal elements whose eigenvalues are constrained to lie on or close to the unit circle in the complex z-plane to maintain network and training stability. The activation function slopes of the ULTRNN state variables are dynamically varied by a second modulating network. The inputs to the modulating network are the state variables of the principal ULTRNNs and their inputs. The modulating network is trained simultaneously with the principal ULTRNN to compute the activation function slope for the latter’s each state variable at each time step. Such dynamic variation of the activation function slopes selectively enhances the contribution of certain states while suppressing that of the others. A larger slope results in a longer time contribution of the corresponding state and helps model long-term dependencies. Conversely, a smaller slope results in a shorter time contribution and may be used to model controlled "forgetting". The proposed modulation technique enhances the ULTRNN’s ability to effectively incorporate short-term memory and long-term dependencies. Simulation results show that with activation function modulation the ULTRNNs are able to autonomously replicate the outputs of sample chaotic dynamic system with good accuracy. This capability can be highly effective in modeling or characterizing the inherent process that generates the time series.
IEEE Transactions on Cybernetics, 2017
This paper introduces a discrete-time recurrent neural network architecture using triangular feed... more This paper introduces a discrete-time recurrent neural network architecture using triangular feedback weight matrices that allows a simplified approach to ensuring network and training stability. The triangular structure of the weight matrices is exploited to readily ensure that the eigenvalues of the feedback weight matrix represented by the block diagonal elements lie on the unit circle in the complex z-plane by updating these weights based on the differential of the angular error variable. Such placement of the eigenvalues together with the extended close interaction between state variables facilitated by the nondiagonal triangular elements, enhances the learning ability of the proposed architecture. Simulation results show that the proposed architecture is highly effective in time-series prediction tasks associated with nonlinear and chaotic dynamic systems with underlying oscillatory modes. This modular architecture with dual upper and lower triangular feedback weight matrices mimics fully recurrent network architectures, while maintaining learning stability with a simplified training process. While training, the block-diagonal weights (hence the eigenvalues) of the dual triangular matrices are constrained to the same values during weight updates aimed at minimizing the possibility of overfitting. The dual triangular architecture also exploits the benefit of parsing the input and selectively applying the parsed inputs to the two subnetworks to facilitate enhanced learning performance.