ADALINE Research Articles

Adaline-Based Control Schemes for Non-Sinusoidal Multiphase Drives—Part II: Torque Optimization for Faulty Mode

Fault tolerance has been known as one of the main advantages of multiphase drives. When an open-circuit fault happens, smooth torque can be obtained without any additional hardware. However, a reconfiguration strategy is required to determine new reference currents. Despite advantages of non-sinusoidal electromotive forces (NS-EMFs) such as high torque density, multi-harmonics existing in NS-EMFs cause more challenges for control, especially under faulty conditions. Therefore, to guarantee high-quality vector control of multiphase drives with multi-harmonic NS-EMFs, this two-part study proposes control schemes using adaptive linear neurons (Adalines) to adaptively eliminate torque ripples. The proposed simple Adalines are efficient because of taking advantage of the knowledge of rotor position and of torque harmonic rank induced by the NS-EMFs. The control scheme using an Adaline for healthy mode was described in part I of this study. In this second part, the control scheme using another Adaline for an open-circuit operation, under the impacts of multi-harmonics in NS-EMFs, is proposed. Notably, smooth torque and similar copper losses in the remaining healthy phases can be obtained. Experimental tests are carried out on a seven-phase permanent magnet synchronous machine (PMSM) with a high total harmonic distortion (THD = 38%) of NS-EMFs. A demonstration video is provided with this paper.

Adaline-Based Control Schemes for Non-Sinusoidal Multiphase Drives–Part I: Torque Optimization for Healthy Mode

More degrees of freedom not only enable multiphase drives to be fault-tolerant but also allow non-sinusoidal electromotive forces (NS-EMFs) in high-quality vector control. NS-EMFs lead to lower costs of design and manufacturing of electrical machines. However, the presence of multi-harmonics in NS-EMFs possibly generates pulsating torque in both healthy and faulty conditions of multiphase drives. To facilitate the use of NS-EMFs, this two-part study proposes control schemes to adaptively improve torque quality of multiphase drives in dealing with multi-harmonics of NS-EMFs. The proposed schemes are based on a simple but effective type of artificial intelligence, Adaptive Linear Neuron (Adaline). The knowledge of multiphase drives including the harmonic ranks of NS-EMFs and the rotor position is exploited to design the online-trained optimal Adalines. The first part of this study is to propose a control scheme using an Adaline for healthy mode with high-quality torque regardless of numerous harmonics in NS-EMFs. The second part of this study introduces a control scheme using another Adaline for open-circuit faults. The proposed schemes are numerically and experimentally validated on a seven-phase permanent magnet synchronous machine (PMSM) possessing a high total harmonic distortion (THD = 38%) of NS-EMFs.

Adaptive linear neuron control of three-phase shunt active power filter with anti-windup PI controller optimized by particle swarm optimization

In the present paper, particle swarm optimization technique (PSO) is used to tune anti-windup PI controller gains for DC bus voltage control of the shunt active power filter by offline simulation. The main advantage of the PSO tuning algorithm is to accurately estimate PI controller parameters and overcome the complexity of analytical calculation approaches as well as classical tuning methods. The sum of the total harmonic distortion of source currents and the DC bus's voltage error are used as a fitness function. The reference current is obtained using an artificial neural network named adaptive linear neuron (ADALINE). Three ADALINEs are trained online using the least mean square algorithm (LMS). The proposed system is designed by MATLAB/Simulink and validated experimentally through the dSPACE DS1104 platform. The analysis is done for various conditions, and the obtained results demonstrate the effectiveness of the controller in terms of performance, efficiency, and power quality enhancement.

Adaptive Control Structure with Neural Data Processing Applied for Electrical Drive with Elastic Shaft

This paper presents issues related to the adaptive control of the drive system with an elastic clutch connecting the main motor and the load machine. Firstly, the problems and the main algorithms often implemented for the mentioned object are analyzed. Then, the control concept based on the RNN (recurrent neural network) for the drive system with the flexible coupling is thoroughly described. For this purpose, an adaptive model inspired by the Elman model is selected, which is related to internal feedback in the neural network. The indicated feature improves the processing of dynamic signals. During the design process, for the selection of constant coefficients of the controller, the PSO (particle swarm optimizer) is applied. Moreover, in order to obtain better dynamic properties and improve work in real conditions, one model based on the ADALINE (adaptive linear neuron) is introduced into the structure. Details of the algorithm used for the weights’ adaptation are presented (including stability analysis) to perform the shaft torque signal filtering. The effectiveness of the proposed approach is examined through simulation and experimental studies.

ANN에 기반한 3상 PWM 컨버터의 다중 스위치 개방고장 진단 방법

For single and double open-switch faults in three-phase PWM (Pulse Width Modulation) converters, there are 21 types of fault modes depending on faulty switches, causing secondary faults in peripherals. In this paper, for diagnosing those fault modes, two-step technique based on ANNs using dc and THDs (Total Harmonics Distortions) of phase currents is proposed. Those dc and THDs are obtained through an ADALINE (Adaptive Linear Neuron) in real-time. In the first step, the ANN categorizes fault modes into six sectors in the three-phase plane. In the second step, the ANN localizes fault modes in each sector. Especially, the fault modes of both switches in the same legs are localized by comparing the number of the sampled zero current of the fault currents. The proposed method allows a real-time diagnosis with a simple and an high accuracy for the multiple open-switch faults and operates online. Simulations and experiments for a 3.7kW three-phase PWM converter confirmed the validity of the proposed diagnostic method.

Ultra-Resolution Spectral Correction Based on Adaptive Linear Neuron for Biomedical Signal Processing.

Ultra-Resolution Spectral Correction Based on Adaptive Linear Neuron for Biomedical Signal Processing.

Uniform Demagnetization Diagnosis for Permanent-Magnet Synchronous Linear Motor Using a Sliding-Mode Velocity Controller and an ALN-MRAS Flux Observer

To realize the demagnetization fault diagnosis and degree estimation of permanent-magnet synchronous linear motors in linear motion applications with high precision and multiple working conditions, a novel double closed-loop control with an adaptive linear neuron (ALN) model reference adaptive system (MRAS) flux linkage observer is proposed in this article. First, a sliding-mode control (SMC) is introduced to design the velocity loop to obtain good speed adaptability. Then, the MRAS flux observer is established, and the ALN algorithm for modifying the MRAS adaptive law online is designed combining with MRAS flux observer. In addition, the finite element models of normal, weak (F1), mild (F2), moderate (F3), and severe (F4) demagnetization faults are established, and the accurate values of flux linkage under different demagnetization degrees are simulated. Furthermore, the demagnetization fault diagnosis algorithm is introduced with an SMC-ALN-MRAS flux observer. Finally, a prototype motor experimental platform is built, and comparative experiments are designed to confirm the correctness and effectiveness of the proposed method.

Power Quality Improvement for Grid-connected PV System Based on Distribution Static Compensator with Fuzzy Logic Controller and UVT/ADALINE-based Least Mean Square Controller

This paper presents a novel method of power quality enrichment in a grid-connected photovoltaic (PV) system using a distribution static compensator (DSTATCOM). The paper consists of two-step control processes. In the primary step, a fuzzy logic controller (FLC) is employed in the DC-DC converter to extract the peak power point from the PV panel, where the FLC produces a switching signal for the DC-DC converter. In the secondary step, a unit vector template (UVT)/adaptive linear neuron (ADALINE)-based least mean square (LMS) controller is adopted in the DC-AC converter, i.e., voltage source converter (VSC). The input to this VSC is the boosted DC voltage, which originates from the PV panel as a result of DC-DC conversion. The VSC shunted with the power grid is known as a DSTATCOM, which can maintain the power quality in the distribution system. The UVT controller generates reference source currents from the grid voltages and DC-link voltages. The ADALINE-based LMS controller calculates the online weight according to the previous weights by the sensed load current. The UVT/ADALINE-based LMS controller of a DSTAT-COM performs several tasks such as maintaining the sinusoidal source current, achieving a unity power factor, and performing reactive power compensation. The reference current extracted from the UVT/ADALINE-based LMS controller is fed to the hysteresis current controller to obtain the desired switching signal for the VSC. A 100 kW solar PV system integrated into a three-phase four-wire distribution system through a four-leg VSC is designed in MATLAB/Simulink. The performances of the FLC and UVT/ADALINE-based LMS controllers are demonstrated under various irradiances as well as constant temperature and nonlinear loading conditions.

A Novel Adaptive Linear Neuron Based on DNA Strand Displacement Reaction Network.

Analog DNA strand displacement circuits can be used to build artificial neural network due to the continuity of dynamic behavior. In this study, DNA implementations of novel catalysis, novel degradation and adjustment reaction modules are designed and used to build an analog DNA strand displacement reaction network. A novel adaptive linear neuron (ADALINE) is constructed by the ordinary differential equations of an ideal formal chemical reaction network, which is built by reaction modules. When reaction network approaches equilibrium, the weights of the ADALINE are updated without learning algorithm. Simulation results indicate that, ADALINE based on the analog DNA strand displacement circuit has ability to implement the learning function of the ADALINE based on the ideal formal chemical reaction networks, and fit a class of linear function.

ANN design of multiple open‐switch fault diagnosis for three‐phase PWM converters

In this paper, Artificial Neural Networks (ANNs) for diagnosing multiple open-switch faults in three-phase PWM (Pulse Width Modulation) converters are designed. For single and double open-switch faults in the converters, there are 21 types of fault modes, causing distorted phase currents. Since these abnormal currents can induce secondary faults in peripherals, the open-fault diagnosis is essentially required. In this paper, a two-step technique based on ANNs is utilized for the diagnosis of the multiple open-switch fault. First, dc and harmonic components of the currents are extracted by using an ADALINE (Adaptive Linear Neuron). In the first step, the ANN categorizes fault modes into six sectors by using dc components in the three-phase plane. In the second step, the ANN localizes fault modes by using dc components and the ratio of d–q axes currents THDs (Total Harmonics Distortions) in each sector. Especially, both switches open-faults in the same legs are localised by counting the sampled zero current of the fault currents. The proposed two-step technique allows a simple design of the ANNs for the diagnosis, and a short execution time about 22 s. Simulations and experiments for a 3.7 kW three-phase PWM converter confirmed the validity of the proposed diagnostic method.

Performance of PV integrated multilevel inverter for PQ enhancement

ABSTRACT Three phase shunt active power filter (SAPF) using multilevel inverter (MLI) is implemented in this paper for a grid integrated solar photovoltaic (PV) system for eliminating the harmonics. The objective in this work is to develop and regulate the grid integrated PV system for mitigation of harmonics with incorporation of SAPF. The PV integrated SAPF delivers the active/reactive power to the grid to reduce the harmonic contents. Harmonic compensation ability of SAPF is mainly achieved by employing suitable control technique to estimate the reference current signal. Here, adaptive artificial neural networks (AANNs) like adaptive neuro-fuzzy interference system (ANFIS), Widrow–Hoff adaptive linear neuron (WH ADALINE) and Modified Widrow–Hoff adaptive linear neuron (MWH ADALINE) are employed for generating the reference current of VSI in the MLI. The control techniques are tested in MATLAB/SIMULINK as well as real-time OPAL-RT environments. It is noticed that proposed MWH ADALINE-based SAPF design shows enhanced performance to maintain the power quality (PQ) by restricting the distortion level below 5% as per the IEEE/IEC standard.

Performance Analysis of BLDC Motor Drive using Enhanced Neural Based Speed Controller for Electric Vehicle Applications

This paper deals with Brushless DC (BLDC) motor drive mathematical model with hysteresis current controlled based Voltage Source Inverter (VSI) which operates using Neural Network (NN). The developed simulation model is applied with the artificial neuro speed controller which is based on Least Mean Square (LMS) and Adaptive Linear Neuron (ADALINE) to improve the performance of BLDC motor drive. Under dynamic operating conditions, the NN speed controller is trained by the data obtained from closed loop speed controller of BLDC motor drive system. The developed conventional and proposed simulation models are simulated using MATLAB/Simulation software. The proposed neural based speed controller is important for tracking of motor speed as well as torque with their reference values with minimum transient time. In this paper ADALINE network model is developed for update of weights in NN. Here, a comparative study has been done for PI and NN controller for BLDC drive. The simulation results show that the NN based speed controller is more effective controller compared to classical PI based speed controller during most of the dynamic operating conditions.

Experimental Validation of ADALINE Least Mean Square Algorithm in a Three-Phase Four-Wire DSTATCOM to Enhance Power Quality

The requirement of a robust and adaptive control algorithm in active filters is inevitable nowadays. In this paper, an ADAptive LINEar neuron (ADALINE) based Least Mean Square (LMS) control algorithm is proposed for a three-phase-four-wire (3P4W) distribution power system under unbalanced loading conditions. The proposed controller has a unique inbuilt characteristic to extract the fundamental active positive sequence component from an unbalanced load current. Particularly, the control algorithm is intended to calculate an adaptive and suitable amplitude for the three-phase reference source currents which are used for the switching pulse generation in the subsequent steps. Moreover, the Distribution STATic COMpensator (DSTATCOM) is performed to eliminate harmonic contents of supply current and harmful excessive neutral current in the system. Further, the DC-link voltage is regulated in order to reduce the power loss across the DSTATCOM, and to attain a balanced, high-quality power availability at the Point of Common Coupling (PCC). The proposed system is accomplished in the MATLAB2014a/Simulink, and a low power rated 3P4W DSTATCOM prototype incorporating with the dSPACE1104 real-time environment to verify the proposed control algorithm.

A Commutation Error Compensation Strategy for High-Speed Brushless DC Drive Based on Adaline Filter

The commutation errors undermine the control performance of a brushless direct current (BLdc) drive especially at high-speed operation. In this article, a novel commutation error compensation strategy for BLdc drive that eliminates the phase deviation between the current and the back electromotive force is proposed. The d- axis current is introduced and used as the criterion to determine the commutation errors. Since the phase currents of BLdc drive contain large amounts of harmonics, which makes the d- axis current hard to detect, a least-mean-square-algorithm-based adaptive linear neuron filter is proposed. The optimal phase advance angle is adaptively generated by reducing the filtered d- axis current to zero. The proposed method does not need the motor parameter information or additional hardware, and can be implemented in general inverter controllers. The effectiveness of the proposed strategy is verified by both simulation and experimental results.

Unity Efficiency and Zero-Oscillations Based MPPT for Photovoltaic Systems

Maximum power point tracking (MPPT) is essential for photovoltaic systems to ensure a maximum power extraction from PV panels. However, some issues such as oscillations, power loss and other technical aspects still unsolved. This paper presents and discusses a new MPPT algorithm with zero-oscillations and unity efficiency in transient and steady-states. This algorithm leads to track the maximum power point under extreme operating conditions. The proposed MPPT method is based on the simple adaptive linear neuron. In addition, its implementation is achieved without any additional control loop, which resulted in a simple control. In order to validate the proposal effectiveness, both simulation and experiment tests are carried out under variable irradiance and load. Comparison between the developed MPPT and the conventional perturb and observe algorithm is also performed. Obtained results show that with the proposed method, unity efficiency is reached and oscillations are fully removed in the transient and steady-states. The originality of this work is the design of a simple and efficient MPPT algorithm based on the ADALINE with unity efficiency and zero-oscillations. Moreover, the proposal is verified using a real PV system under irradiance and load changes.

A New Robust Identification Method for Transmission Line Parameters Based on ADALINE and IGG Method

Accurate transmission line parameters are the basis of power system calculations. Aiming at obtaining accuracy online transmission line parameters in the case of large random noises even bad data in Phasor Measurement Unit (PMU) measurements, which occurs frequently in the practice, a new adaptive robust identification method combining adaptive linear neuron (ADALINE) and traditional robust IGG (Institute of Geodesy & Geophysics, Chinese Academy of Sciences) method is proposed. In detail, first, the identification model of transmission line parameters is presented based on the multi-period PMU measurements at both ends of the transmission line. Then, a parameter solving model based on ADALINE is established. Furthermore, to fully use measurement information, the adaptive robust ADALINE (ARA) are proposed, which applies the robust IGG weight function (Scheme I) to ADALINE to realize “three segments” robust identification. In addition, to improve the robustness, the expectation and variance of the equation residual sequence are estimated adaptively with the median principle to adjust the threshold for the IGG function to assure robustness (TAR), which is independent of the known information for the error of the measurement equipment. The cases based on PSCAD simulated data and measured data show the effectiveness and engineering practicality of the proposed method.

Adaptive parameters identification of lithium-ion batteries with adaptive linear neuron and state-of-charge estimation based on open circuit voltage

The state of charge (SOC) is a critical parameter of a Li-ion battery. An accurate online estimation of the SOC is important for forecasting the electric vehicle driving range. A good estimation of the SOC results from a good identification of the battery parameters. Reducing the algorithm complexity is important to improve the accuracy of SOC estimation results. We propose in this work an original structure of an adaptive linear neuron (ADALINE) to estimate the SOC. The ADALINE provides the weighted sum of the inputs based on an online identification of the open-circuit voltage (OCV). The advantage of this approach is its adaptable capability and the speed of execution (fast training) as well as the possibility of interpreting these weights. The simulation results indicate that the proposed method can ensure an acceptable accuracy of SOC estimation for online application with a maximum error being less than 5%.

Adaptive parameters identification of lithium-ion batteries with adaptive linear neuron and state-of-charge estimation based on open circuit voltage

The state of charge (SOC) is a critical parameter of a Li-ion battery. An accurate online estimation of the SOC is important for forecasting the electric vehicle driving range. A good estimation of the SOC results from a good identification of the battery parameters. Reducing the algorithm complexity is important to improve the accuracy of SOC estimation results. We propose in this work an original structure of an adaptive linear neuron (ADALINE) to estimate the SOC. The ADALINE provides the weighted sum of the inputs based on an online identification of the open-circuit voltage (OCV). The advantage of this approach is its adaptable capability and the speed of execution (fast training) as well as the possibility of interpreting these weights. The simulation results indicate that the proposed method can ensure an acceptable accuracy of SOC estimation for online application with a maximum error being less than 5%.

Different artificial intelligence strategies to control an inverter of an active power filter for power quality improvement

In this paper, performance of active power filter (APF) is analysed for various types of artificial intelligent current controllers which are used to generate switching signals of the voltage source inverter (VSI). Three efficient and reliable intelligent approaches were adopted. The first two are based on a PI-neural regulator and the direct control using a multi-layer perceptron neural network (MLP) respectively. The third one is based on the fuzzy logic regulator. The objective is to take advantages of these intelligent techniques to improve the compensation performance of the conventional APF to minimise harmonic contamination drawn from nonlinear loads, and enhance the power quality. Further, we propose to extend the use of the adaptive linear neuron (ADALINE) in all control schemes to build a homogeneous computing structure. The proposed compensator APF-neural, based on a complete neuromimetic strategy shows its effectiveness and robustness compared to the fuzzy and conventional current controllers.

Different artificial intelligence strategies to control an inverter of an active power filter for power quality improvement

In this paper, performance of active power filter (APF) is analysed for various types of artificial intelligent current controllers which are used to generate switching signals of the voltage source inverter (VSI). Three efficient and reliable intelligent approaches were adopted. The first two are based on a PI-neural regulator and the direct control using a multi-layer perceptron neural network (MLP) respectively. The third one is based on the fuzzy logic regulator. The objective is to take advantages of these intelligent techniques to improve the compensation performance of the conventional APF to minimise harmonic contamination drawn from nonlinear loads, and enhance the power quality. Further, we propose to extend the use of the adaptive linear neuron (ADALINE) in all control schemes to build a homogeneous computing structure. The proposed compensator APF-neural, based on a complete neuromimetic strategy shows its effectiveness and robustness compared to the fuzzy and conventional current controllers.