New Version of Adaptive Speed Observer based on Neural Network for SPIM

This paper presents a new model reference adaptive system (MRAS) speed observer for high-performance field-oriented control induction motor drives based on adaptive linear neural networks. It is an evolution and an improvement of an MRAS observer presented in the literature. This new MRAS speed observer uses the current model as an adaptive model discretized with the modified Euler integration method. A linear neural network has been then designed and trained online by means of an ordinary least-squares (OLS) algorithm, differently from that in the literature which employs a nonlinear backpropagation network (BPN) algorithm. Moreover, the neural adaptive model is employed here in prediction mode, and not in simulation mode, as is usually the case in the literature, with a consequent quicker convergence of the speed estimation, no need of filtering the estimated speed, higher bandwidth of the speed loop, lower estimation errors both in transient and steady-state operation, better behavior in zero-speed operation at no load, and stable behavior in field weakening. A theoretical analysis of some stability issues of the proposed observer has also been developed. The OLS MRAS observer has been verified in numerical simulation and experimentally, and in comparison with the BPN MRAS one presented in the literature.

In this paper, a novel Stator Current Based Model Reference Adaptive System (SC_MRAS) speed estimation scheme using neural network (NN) and Sliding Mode (SM) is proposed to improve the performance of the MRAS speed observer for high-performance Six Phases Induction Motor (SPIM) drives, especially at low and zero speed region, where the poor performance of observers is still always a large challenge. In this novel SC_MRAS scheme, a two-layer linear NN, which has been trained online by means of the Total Least Squares (TLS) algorithm, is used as an adaptive model to estimate the stator current and this model is employed in prediction mode. These novel proposed can ensure that the whole drive system achieves faster satisfactory torque and speed control and strong robustness, the observer operate better accuracy and stability both in transient and steady-state operation. Especially, in this proposed observer, the rotor flux, which is needed for the stator current estimation of the adaptive model and providing to the controller, is identified based on adaptive SM technique. The improvement of Rotor Flux Estimation for SC_MRAS-Based Sensorless SPIM Drives help to eliminate the disadvantages in SC_MRAS based observer such as stator resistance sensitivity, and flux open loop integration which may cause dc drift and initial condition problems or instability in the regenerating mode of operation, therefore, enhancing the rotor flux estimation, speed estimation and control accuracy at very low and zero stator frequency operation help improve the overall observer and drive system performance. The indirect field oriented control (IFOC) for speed control of a sensorless SPIM drive using the proposed observer is built by MATLAB/ Simulink. The simulation results are presented under sensorless speed control performance to validate the effectiveness of the proposed estimation and control algorithms.

In this paper, a Stator Current Model Reference Adaptive System (SC_MRAS) based estimator using neural network for sensorless controlled Six Phase Induction Motor (SPIM) drive is presented. For this scheme the measured stator current components are used as the reference model of the MRAS estimator to avoid the use of a pure integrator and reduce influence of motor parameter variation. The adaptive model of proposal estimator use a two-layer Neural Network (NN) to estimate the stator current, which requires the rotor flux information that can be obtained from the voltage model. A back-propagation learning algorithm is used to minimise the error in current estimation and hence in generating the estimated speed. Speed estimation performance of the proposal MRAS scheme is verified and compared with the classical rotor flux MRAS when applied to sensorless vector control SPIM drive. Simulation results have demonstrated that the proposed MRAS estimator for sensorless control of SPIM drive has good dynamic response, high precision in both the transient and steady modes, the performance of the SC_MRAS using NN is significantly improved especially at low and near zero speed range.

In this paper, a novel model reference adaptive system (MRAS) observer based sensorless control of permanent magnet synchronous motor (PMSM) is proposed. This new speed observer uses the system model as the reference model, and the discrete system model with estimated rotor speed as the adaptive model to estimate the stator current, then, uses the gradient descent with an optimized proportional coefficient as the adaptive law to estimate the rotor speed. The adaptive model is a linear neural network which can be trained on line by means of back propagation. Moreover, the adaptive model is in prediction mode, which has a better performance compared with the simulation mode. The performance of the proposed method has been verified by Matlab/Simulink.

Digital image recognition of plant diseases could reduce the dependence of agricultural production on the professional and technical personnel in plant protection field and is conducive to the development of plant protection informatization. In order to find out a method to realize image recognition of plant diseases, four kinds of neural networks including backpropagation (BP) networks, radial basis function (RBF) neural networks, generalized regression networks (GRNNs) and probabilistic neural networks (PNNs) were used to distinguish wheat stripe rust from wheat leaf rust and to distinguish grape downy mildew from grape powdery mildew based on color features, shape features and texture features extracted from the disease images. The results showed that identification and diagnosis of the plant diseases could be effectively achieved using BP networks, RBF neural networks, GRNNs and PNNs based on image processing. For the two kinds of wheat diseases, the best prediction accuracy was 100% with the fitting accuracy equal to 100% while BP networks, GRNNs or PNNs were used, and the best prediction accuracy was 97.50% with the fitting accuracy equal to 100% while RBF neural networks were used. For the two kinds of grape diseases, the best prediction accuracy was 100% with the fitting accuracy equal to 100% while BP networks, GRNNs or PNNs were used, and the best prediction accuracy was 94.29% with the fitting accuracy equal to 100% while RBF neural networks were used.

Fast and accurate synthesis of frequency reconfigurable slot antenna using back propagation network

In order to forecast missile storage reliability better, the paper researched a forecasting method based on neural network which is with the ability of actualizing multi-nonlinear mapping from input to output, and discussed steps of forecasting based on back propagation (BP) network and radial basis function (RBF) network respectively. At last, the storage reliability of one type ship-to-ship missile is forecasted based on BP network and RBF network respectively. The results show that both of the BP and RBF are suitable for Missile storage reliability forecasting, and the precision of the train goal is better by using RBF network. RBF network is more suitable for dealing with this problem.

Malaysian experts have warned that several major cities, including the capital Kuala Lumpur, could face serious water shortages due to over-pollution of the country's rivers by problems brought about by over-development. As 97 per cent of Malaysia's water supply is sourced from surface water, the main aim of this study was to identify a relationship between water quality and land use attributes. The study which was conducted on Gombak River and its watershed in Malaysia, introduced data generation method for the prediction and forecast of LU/LC data within the watershed. The method used exponential model equation, Lagrange model equation third & fourth degree polynomial fit; saturation growth-rate model in order to generate the required data: and artificial neural network's back propagation network algorithm. The study also introduces the LA-WQI model. This model was developed by associating the appropriate loading factors to a set of sub-indices. The findings revealed that as the activities increased throughout the watershed, the values of WQI quality decreased accordingly. The accuracy of prediction of the proposed LA-WQI ranged from 94.3% to 99.3% between Actual DOE-WQI and LA-WQI for station 18 in Gombak River. The results of predicted WQI obtained using LA-WQI, showed a continuous decrease of water quality. Despite the high accuracy attained by the application of LA-WQI model on Gombak River; it has not yet been tested on other rivers. It is recommended that future studies should be able to further test the current model on a regional scale.

To improve the network performance of radial basis function (RBF) and back-propagation (BP) networks on complex nonlinear problems, an integrated neural network model with pre-RBF kernels is proposed. The proposed method is based on the framework of a single optimized BP network and an RBF network. By integrating and connecting the RBF kernel mapping layer and BP neural network, the local features of a sample set can be effectively extracted to improve separability; subsequently, the connected BP network can be used to perform learning and classification in the kernel space. Experiments on an artificial dataset and three benchmark datasets show that the proposed model combines the advantages of RBF and BP networks, as well as improves the performances of the two networks. Finally, the effectiveness of the proposed method is verified.

This paper describes a speaker-dependent, isolated Chinese word recognition system based on neural networks. An improved neural network is applied to the recognition of speaker-dependent isolated Chinese words. The improved neural network is composed of several BP (back-propagation) networks. The isolated Chinese word sets are partitioned into a group of subsets based on a priori phonological knowledge. One of the BP networks identifies the subset to which the input word belongs; the others recognize the words in the subset. The improved neural network has the following advantages over a single BP network: training time is reduced; higher recognition accuracy is obtained with less training samples; new words can be easily added by adding new subsets. >

In our study, we use six nonlinear indexes and two linear ones, which are Complexity, Lyapunov Exponent, Predictability, Correlation Dimension, Pioncare Scattergraph, Approximate Entropy; Time and Frequency Domains, respectively, as input parameters of neural networks, i.e. Back-Propagation (BP) networks with only one hidden layer. According to all eight indexes of the subjects given by WuHan TongJi Hospital, we not only introduce a BP network, but improve the algorithm to avoid its proper weakness as well. Neural networks are always regarded as an accurate way to analyze and study the structure of a system. The simulation results reinforce the fact that it is more effective to clinically diagnose cardiovascular diseases by combining the eight indexes with the optimized algorithm in BP networks.

This paper proposes the design of sensorless induction motor drive based on direct power control (DPC) technique. It is shown that DPC technique enjoys all advantages of pervious methods such as fast dynamic and ease of implementation, without having their problems. To reduce the cost of drive and enhance the reliability, an effective sensorless strategy based on artificial neural network (ANN) is developed to estimate rotor’s position and speed of induction motor. Developed sensorless scheme is a new model reference adaptive system (MRAS) speed observer for direct power control induction motor drives. The proposed MRAS speed observer uses the current model as an adaptive model. The neural network has been then designed and trained online by employing a back propagation network (BPN) algorithm. The estimator was designed and simulated in Simulink. Some simulations are carried out for the closed-loop speed control systems under various load conditions to verify the proposed methods. Simulation results confirm the performance of ANN based sensorless DPC induction motor drive in various conditions.

In this paper, we described breast tissue image analyses using texture features from Haar wavelet transformed images to classify breast lesion of ductal organ Benign, DCIS and CA. The approach for creating a classifier is composed of 2 steps: feature extraction and classification. Therefore, in the feature extraction step, we extracted texture features from wavelet transformed images with 10/spl times/ magnification. In the classification step, we created three classifiers from each image of extracted features using statistical discriminant analysis, neural networks (back-propagation algorithm) and SVM (support vector machines). In this study, we conclude that the best classifier in histological sections of breast tissue in the texture features from second-level wavelet transformed images used in discriminant function.

In this paper, an approach for reducing the torque ripple in a two phase unsymmetrical induction motor is analyzed. A single phase induction motor is basically a two phase induction motor and generally, the windings on the stator are not identical. Hence, it is called unsymmetrical two phase induction motor. If an unsymmetrical motor is supplied by two phase balanced supply voltages then the ripples in the electromagnetic torque are unavoidable. However, if the supply voltages are made unbalance to counter the unbalance in the windings the torque ripple can be reduced. The mathematical model of the machine in stator attached reference frame is used to derive the torque expression in terms of stator currents. The torque expression consists of a constant term and a pulsating term. These terms used together unable finding the conditions on stator currents and hence stator voltages so that the ripples in the torque produced can be mitigated.

An adaptive predictive control algorithm of nonlinear non-minimum phase systems using neural network is proposed. The nonlinear system is separated into linear non-minimum phase system and nonlinear parts by Taylor series expansion. The resulting nonlinear part is identified by a neural network and compensated in the control algorithm such that feedback linearization can be achieved. A modified neural network composed of linear neural network (LNN) which represent the linearized model at the operating point and a multilayered feedforward neural network which approximate the nonlinear dynamics that cannot be modeled by the LNN is utilized in this investigation.