LM Algorithm Research Articles

Enhancing performance of shipboard photovoltaic grid-connected inverter through CRNN-LM-BP control optimized by particle swarm optimization of LCL parameters

The global renewable energy portfolio is experiencing a growing proportion of shipboard photovoltaic (PV) energy. Grid-connected system has become the development trend of photovoltaic power generation technology in marine applications because of its high energy utilization efficiency. However, in the grid-connected process, harmonic pollution is easily caused by time-varying marine environment disturbance, which affects the quality of grid-connected power. The current recurrent neural network Levenberg-Marquardt backpropagation (CRNN-LM-BP) method, a control methodology, is introduced in this paper to ensure the robustness, stability, and consistency of the system. In addition, the inductance-capacitance-inductance (LCL) filter is optimized using particle swarm optimization (PSO) to minimize the presence of high frequency harmonics. The system that has been developed possesses the ability to improve the accuracy of steady-state control, operate with high efficiency and effectiveness, and maintain a total harmonic distortion of 1.73 %. Concurrently, streamlining the Jacobian matrix in the LM algorithm expedites convergence. It enables the system to improve its dynamic response, minimize excess, and quickly return to a stable condition when faced with load variations, a wide range of filter parameter adjustments, and voltage fluctuations. The suggested method’s effectiveness and applicability are evaluated by comparing it to the traditional proportional-resonant control, CRNN-BP control, using Matlab/Simulink and the real-time simulator OPAL-RT5700..

An integration method of a hybrid genetic algorithm and the Levenberg–Marquardt algorithm for ultrasonic testing

The accurate measurement of time-of-flight (TOF) is essential in ultrasonic testing. Further, noise interference is the key factor affecting the measurement accuracy. Therefore, to develop a reliable computational method of TOF for test pieces working in noisy environments, an integration method of a hybrid genetic algorithm and the Levenberg–Marquardt algorithm (GA–LM) for ultrasonic thickness measurement is proposed in the present research. A Gaussian model is first established for an echo signal. Further, the model-based parameter estimation is converted into a nonlinear optimization problem by applying the least square method. As the parameter estimation methods are easily affected by the initial value, an integrating innovation of the GA–LM algorithm is proposed. The initial values of the model parameters are selected by GA to obtain an approximate global optimal solution. Subsequently, this approximate solution is used as the initial value for the LM algorithm to perform iterations. The accurate global optimal solution of the Gaussian model is obtained through these iterations. Finally, the measuring accuracy and robustness of the GA–LM algorithm for TOF computation are verified by both numerical simulation and experiment data

Improved ellipsoid fitting aided geomagnetic sensor calibration algorithm

Aiming at the problems that the traditional single ellipsoid fitting fails to eliminate the errors of the geomagnetic sensor and does not consider the geomagnetic anomaly information, etc., this paper proposes a robust least squares combined with the Levenberg–Marquardt (LM-RLS) algorithm to fit an accurate ellipsoid to achieve accurate data calibration. According to the geomagnetic sensor error model, the total amount of geomagnetism is used as the constraint, the fitted ellipsoid coefficients of RLS are used as the initial values, and the LM algorithm is used to calculate the magnetic field dispersion points to the nearest point of the fitted ellipsoid surface. Then, the initial calibration value is solved by the criterion of the minimum weighted sum of squares of the closest point distance. Finally, the accurate ellipsoid is obtained to realize data calibration. The results of the dynamic experiment show that after the ellipsoid is precisely fitted by the LM-RLS algorithm, the sum of the absolute values of the distance from the magnetic field discrete data to the ellipsoid surface is reduced by 48.33% (RLS) and 43.10% (RWTLS). The relative errors of the magnetic field in the X, Y, and Z axes are reduced from 0.48%, 0.78%, 1.48% (RLS) and 0.30%, 0.58%, 1.14% (RWTLS) to 0.10%, 0.04% 0.16 %, respectively. The accuracy of geomagnetic calibration is improved by 68.06% (RLS) and 61.02% (RWTLS) in the dynamic experiment. The algorithm improves the accuracy of geomagnetic measurement, which provides a new idea for the calibration of three-axis geomagnetic sensors.

A Method for Solving Ill-Conditioned Nonlinear Least Squares Problems and Its Application in Image Distortion Correction Using Self-Calibration

In this study, the ill-conditioning of the iterative method for nonlinear models is discussed. Due to the effectiveness of ridge estimation for ill-conditioned problems and the lack of a combination of the H-K formula with the iterative method, the improvement of the LM algorithm is studied in this paper. Considering the LM algorithm for ill-conditioned nonlinear least squares, an improved LM algorithm based on the H-K formula is proposed for image distortion correction using self-calibration. Three finite difference methods are used to approximate the Jacobian matrix, and the H-K formula is used to calculate the damping factor in each iteration. The Brown model, quadratic polynomial model and Fourier model are applied to the self-calibration, and the improved LM algorithm is used to solve the model parameters. In the simulation experiment of space resection of a single image, we evaluate the performance of the LM algorithm based on the gain ratio (LMh) and the improved LM algorithm based on the H-K formula (LMHK), and the accuracy of different models and algorithms is compared. A ridge trace analysis is carried out on the damping factor to illustrate the effects of the improved algorithm in handling ill-conditioning. In the second experiment, the improved algorithm is applied to measure the diameter of a coin using a single camera. The experimental results show that the improved LM algorithm can reach the same or higher accuracy as the LMh algorithm, and it can weaken the ill-conditioning to a certain extent and enhance the stability of the solution. Meanwhile, the applicability of the improved LM algorithm in self-calibration is verified.

Time Series Modeling of Global Average Absolute Sea Level Change

This study aimed to demonstrate the effectiveness of time series models in modeling long-term records of global average absolute sea level changes from 1880 to 2014. Following the Box–Jenkins methodology, the ARIMA(0,1,2) model with drift was identified as the best-fit model for the time series due to its lowest AIC value. Using the LM algorithm, the results revealed that the NARNN model with 7 neurons in the hidden layer and 7 time delays exhibited the best performance among the nonlinear autoregressive neural network models, as indicated by its lower MSE. While ARIMA models excel in modeling linear problems within time series data, NARNN models are better suited for nonlinear patterns. However, a HYBRID model was explored, which combines the strengths of both ARIMA and NARNN models, offering the capability to address both linear and nonlinear aspects of time series data. The comparative analysis of this study demonstrated that the HYBRID model, with 6 neurons in the hidden layer and 7 time delays, outperformed the NARNN model with 7 neurons in the hidden layer and 7 time delays, as well as the ARIMA(0,1,2) model, with the lowest MSE in this study. These findings represent a significant step in time series forecasting by leveraging the strengths of both statistical and machine learning methods.

IoT-Based Smart Energy Management in Hybrid Electric Vehicle Using Driving Pattern

Driving patterns involve multiple starts, stops, and goes, requiring both average and momentary demands on power. Batteries function better for the average power demand. An ultracapacitor is used as an auxiliary system with a longer life cycle and higher power density for the vehicle’s momentary power needs. The challenging part of hybridization is smart energy management among the sources. An energy management control strategy decides the energy share among the ESSs. Most of the researchers used standard driving patterns for analyzing vehicle performance. However, a real-time driving pattern is needed to investigate the vehicle’s performance. An IoT-based smart decision-making module can better regulate the energy across the multisource based on sensor unit data about vehicle driving speed. The experimental findings illustrate the effect of the suggested smart energy management control on the hybridization of ESSs to split the power among ESSs to meet the demand for power, maintain the charges in ESSs, and extend the lifespan of the battery. Based on the calculations, the LM approach has a minimal MAD of 0.046781 for training data and 0.043859 for testing data. It has been said that the LM algorithm has the lowest MSE, with values of 0.006397 and 0.005781 for training and testing, respectively. Even in RMSE, the training and testing values of 0.079355 and 0.075893 for the LM Algorithm are the lowest. When the LM algorithm is being trained, its MAPE score is 0.022489, but when it is being tested, it is 0.004846.

Turbo-shaft engine adaptive neural network control based on nonlinear state space equation

Intelligent Adaptive Control (AC) has remarkable advantages in the control system design of aero-engine which has strong nonlinearity and uncertainty. Inspired by the Nonlinear Autoregressive Moving Average (NARMA)-L2 adaptive control, a novel Nonlinear State Space Equation (NSSE) based Adaptive neural network Control (NSSE-AC) method is proposed for the turbo-shaft engine control system design. The proposed NSSE model is derived from a special neural network with an extra layer, and the rotor speed of the gas turbine is taken as the main state variable which makes the NSSE model be able to capture the system dynamic better than the NARMA-L2 model. A hybrid Recursive Least-Square and Levenberg-Marquardt (RLS-LM) algorithm is advanced to perform the online learning of the neural network, which further enhances both the accuracy of the NSSE model and the performance of the adaptive controller. The feedback correction is also utilized in the NSSE-AC system to eliminate the steady-state tracking error. Simulation results show that, compared with the NARMA-L2 model, the NSSE model of the turbo-shaft engine is more accurate. The maximum modeling error is decreased from 5.92% to 0.97% when the LM algorithm is introduced to optimize the neural network parameters. The NSSE-AC method can not only achieve a better main control loop performance than the traditional controller but also limit all the constraint parameters efficiently with quick and accurate switching responses even if component degradation exists. Thus, the effectiveness of the NSSE-AC method is validated.

Lithium battery model parameter identification based on the GA-LM algorithm

ABSTRACT The accuracy of lithium battery model parameters is the key to lithium battery state estimation. The offline parameter identification method for lithium batteries requires the nonlinear fitting of the voltage rebound curve of the hybrid pulse discharge experiment. The genetic algorithm has a strong global search ability, but it is easy to fall into local solutions. The Levenberg-Marquardt algorithm has a strong local optimization ability, but the algorithm cannot converge when the prior value is unknown. Given the above problems, this paper proposes a parameter identification method based on the Genetic-Levenberg-Marquardt (GA-LM) algorithm, which takes the sum of the squared model voltage errors as the objective function, and predicts the initial value of the parameter vector through the GA, providing the LM algorithm with prior value. In the case of unknown prior values, the GA-LM algorithm can achieve high-precision nonlinear optimization. Finally, the simulation test under the conditions of constant current discharge and hybrid pulse power discharge. The mean absolute error, mean relative error, and root mean square error of the model voltage in the two working conditions are 7.23 mV, 0.20%, 9.61 mV, and 13.37 mV, 0.37%, 15.44 mV, which shows that the algorithm has high accuracy.

Optimization of Levenberg Marquardt Algorithm Applied to Nonlinear Systems

As science and technology advance, industrial manufacturing processes get more complicated. Back Propagation Neural Network (BPNN) convergence is comparatively slower for processing nonlinear systems. The nonlinear system used in this study to evaluate the optimization of BPNN based on the LM algorithm proved the algorithm’s efficacy through a MATLAB simulation analysis. This paper examined the application impact of the enhanced approach using the Continuous stirred tank reactor (CSTR) control system as an example. The study’s findings demonstrate that the LM optimization algorithm’s identification error exceeds 10-5. The research’s suggested control approach for reactant concentration CA in CSTR systems provides a better tracking effect and a stronger anti-interference capacity. Compared to the PI control method, the overall control effect is superior. As a result, the optimization model for nonlinear systems has a greatly improved processing accuracy. With some data support for the accuracy study of neural network models and the application of nonlinear systems, the suggested LM-BP optimization algorithm is evidently more appropriate for nonlinear systems.

Pupil Detection for Automatic Diagnosis of Eye Diseases Using Optimized Color Mapping

Introduction: Pupil and iris disorders form an important category of eye diseases. Accurate segmentation of the pupil is the first and most important step in the automatic diagnosis of diseases related to the pupil and iris. Most of the existing methods do not have enough accuracy and are sensitive to the effects of noise and specular spot reflection. In addition, the images used in these methods usually have limitations, such as the viewing angle . Method: In the proposed algorithm, a stable method is offered to remove the effects of specular spot reflection in the pupil, and necessary preprocessing is done to detect the exact location of the pupil. An optimized color mapping algorithm is proposed and the mapping is calculated with the help of the LM algorithm to accurately determine the pupil boundary. This method does not impose any restrictions on the eye image and shape, and the angle of the pupil in the image can be in any shape and direction . Results: The proposed method does not assume any specific model as the final pupil boundary (circle or oval) and is robust to noise and specular reflection factors as well. This method has been able to accurately detect the pupil boundary with the accuracy of 98.8% using UBIRIS dataset and 98% using the collected data by authors . Conclusion: The method presented in this paper can be used to increase the accuracy in determining the internal and external border of the iris to diagnose diseases related to the pupil and iris, as well as identity identification based on iris tissue.

Stock Market Prediction, COVID Pandemic, and Neural Networks: An Levenberg Marquardt Algorithm Application

Stock market forecasting has always piqued the interest of a wide range of investors, practitioners, and researchers. Stock prediction is a complex process due to the presence of an inherent noisy and volatile environment. The stock market’s movement is influenced by a variety of factors. The study of ANN models began in 1969, “when Minsky and Papert discovered two critical flaws in the Artificial Neural Network technique. The first was the machine’s ability to solve complex problems, and the second was the computers’ inability to run large ANN models efficiently”. The study aims to forecast the Nifty 50 using macroeconomic factors as input variables in the two sub-periods, that is, pre-COVID (February 2018–February 2020) and during COVID (March 2020–December 2021). A model trained using the LM algorithm was used for predicting the NSE’s flagship index Nifty 50. The findings reveal that the LM algorithm achieved 95.18% accuracy in predicting the Nifty 50 in the pre-COVID scenario. Whereas during COVID period, the proposed ANN model achieved 94.21% accuracy. The empirical results have important implications for every class of investors, such as FIIs, DIIs, retail investors, and so on.

Development of safety method for a 3-DOF industrial robot based on recurrent neural network

In this paper, a safety method for a 3-DOF industrial robot is developed based on recurrent neural network (RNN). Safety standards for human robot interaction (HRI) are taken into accounts. The main objective is to detect the undesired collisions on any of robot links. Since most of industrial robots are not collaborative, the dependence of the method on torque sensors to detect collisions makes its ability to use very restricted. Therefore, only the position data of joints are collected to be the data inputs of the proposed method in order to detect the undesired collisions. These data are aggregated from KUKA LWR IV robot while no collisions and in another time when applying collisions. These data are used to train the proposed RNN using Levenberg-Marquardt LM algorithm. KUKA robot is configured to act as a 3-DOF manipulator that moves in space and under the effect of gravity.The results show that the modelled and trained RNN is sensitive and efficient in detecting collisions on each link of robot separately. Studying the resulted error from the developed model reveals clearly that the method is reliable.

Research on Propellant Performance Prediction Based on Neural Network or Algorithm Optimization

In order to analyze the four main properties of propellant, the performance prediction model based on BP neural network is established with LM algorithm and CG algorithm as the core algorithm, and the simulation calculation is carried out with the actual production process data and the performance of the model is analyzed. The results show that the mass prediction model based on BP neural network algorithm can quickly predict the properties of propellant. Compared with the prediction model based on CG algorithm, the prediction model based on LM algorithm is more suitable for propellant performance prediction and analysis.

Optimal Control of Rural Water Supply Network Based on Intelligent Algorithm

Optimizing Rural Water Supply Network (RWSN) is the basis for improving rural people’s lives and improving people’s health. Currently, the RWSN in China is relatively backward and can no longer meet the needs of the unified management of rural water resources. To optimize the RWSN, this study innovatively established a Multi-Objective Optimization Mathematical Model (MOMM) of RWSN, combining economic factors and reliability. This experiment first analyzes the characteristics of the RWSN system and then establishes a MOMM of a water supply network. NSGA-II algorithm and LM algorithm are introduced to handle the multi-objective model. The research results show that compared to Web decision tools, the RWSN based on the LM-NSGA-II algorithm can save 5.4% of the total annual cost of water supply pipelines. Therefore, the MOMM of the rural water supply pipeline based on the LM-NSGA-II algorithm has better economy and reliability. The experiment aims to provide certain reference values for the optimal control of RWSN through this study.

Comparison of Various Prediction Model for Biodiesel Cetane Number using Cascade-Forward Neural Network

Cetane number (CN) is one of the important fuel properties of diesel fuels. It is a measurement of the ignition quality of diesel fuel. Numerous studies have been published to predict the CN of biodiesels. More recently, the utilization of soft computing methods such as artificial neural networks (ANN) has received considerable attention as a prediction tool. However, most studies in the use of ANN for estimating the CN of biodiesels have only used one algorithm to train a small number of datasets. This study aims to predict the CN of 63 biodiesels based on the fatty acid methyl esters (FAME) composition by developing an ANN model that was trained with 10 different algorithms. To the best of our knowledge, this is the first study to predict the CN of biodiesels using numerous ANN training algorithms utilizing sizeable datasets. Results revealed that the ANN model trained with Levenberg-Marquardt gave the highest prediction accuracy. LM algorithm successfully predicted the CN of biodiesels with the highest correlation and determination coefficient (R = 0.9615, R2 = 0.9245) as well as the lowest errors (MAD = 2.0804, RMSE = 3.1541, and MAPE = 4.2971). Hence, the Cascade neural network trained with the LM algorithm could be considered a promising alternative to the empirical correlations for predicting biodiesel’s CN.

Application of Combined Prediction Model in Surface Roughness Prediction

In order to improve the stability work piece surface roughness prediction model after machining, LM algorithm, least squares algorithm and proportional conjugate gradient algorithm are used to build the prediction model of the cutting process of AL-7075 aluminum alloy. Two error analysis methods are used to compare the combination model with other single models. The study found that the forecasting model is more accurate and stable than the single forecasting model, and closer to the actual measurement results. The combination model provides a new way to predict the surface roughness of the work piece after machining and provides a theoretical basis for selecting the three main factors of machining.

Performance Prediction of solar still using Artificial neural network

The development of various solar energy systems has emerged as one of the most important answers to the problem of rising energy demand. By utilising the notion of Intelligent system-based methodologies to forecast the performance of solar stills working in an air-dry environment. The artificial neural network concept was applied to forecast the performance of single slope single basin solar stills in this study.Data gathered from previously published articles and used it to forecast the performance of conventional solar still. The developed model can be used to forecast productivity with accuracy of 0.99459 by using LM algorithm.

Fault classification of power plants using artificial neural network

ABSTRACT In order to run power plant operations smoothly, power plant faults need to be detected, located, and classified quickly. For this, artificial neural network approaches are considered significant tool in related application of power system. This research paper focuses on power plant faults detection and classification using intelligent approach of artificial neural network. There is frequent nature of power plant faults, in particular transmission line faults. The key intention of this research is to use ANN approach to detect and classify power plant faults, and in the end to compare their performance and accuracy. In this proposal, the three-phase current and voltage are use as input. Backpropagation (BPNN) technique is used in developing the learning algorithm for multi-layer neural network that can accurately and reliably classify faults. The multi-layer neural network such as Levenberg-Marquardt and Bayesian regularization are used for offline training of the data. Findings from this research show that both Levenberg-Marquardt and Bayesian regularization show good accuracy to detect and classify faults in transmission line. However, LM algorithm trained data much faster than BR. LM shows fastest detection of faults with overall mean square error value of 6.13785e-3 at the 91 epoch. Whereas, BR trained network at the 717 epochs with overall mse of 4.67123e-3. It is to be noted that the accuracy of trained network to classify all the fault types is found to be 81.1%, which is higher than previous studies using Backpropagation neural network technique. Training and test results show that these neural network-based methods can effectively detect and classify faults and faults types and have adequate performance. Matlab is used in this research because it has toolbox, which has effective features to train required data and achieve effective result.

Prediction of Groundwater Quality Index in the Selected Divisions of Srikakulam Using Artificial Neural Networks Approach

Applicability of artificial neural network (ANN) modelling in predicting the water quality index (WQI) and in turn to ascertain the suitability of the water for human consumption has been presented in the paper. In the light of the present study, seventy-nine (79) groundwater samples were collected from two mandals (divisions) Veeraghattam (VGT) and Palakonda (PLKD) and analyzed for physicochemical parameters during the pre-monsoon and post-monsoon seasons of 2015 and 2016. In computing the WQI, physicochemical parameters such as pH, EC, TDS, TH, Ca, Mg, chlorine, fluoride, nitrite, DO and TA have been considered. From the results it was found that the WQI varies from 43.9 to 46.5 and 31.4 to 34.7 in VGT and PLKD divisions respectively. ANN tool in MATLAB has been used to predict the WQI. Back propagation methodology and LM algorithm has been chosen for the study. To train the network, physicochemical parameters have been given as inputs and the already computed WQI values as output. A particular season has been chosen for testing the network. After simulating the network, the results obtained were compared with the experimental value and found to have an error of 0.6%. It is inferred that the chosen model fits apt for the prediction of WQI in the present study.

Construction of Remote Sensing Model of Fresh Corn Biomass Based on Neural Network.

Corn has a high yield and is widely used. Therefore, developing corn production and accurately estimating corn biomass yield are of great significance to improving people's lives, developing rural economy and climate issues. In this paper, a 3-layer BP neural network model is constructed by using the LM algorithm as the training algorithm of the corn biomass BP network model. From the three aspects of elevation, slope, and aspect, combined with the BP neural network model of corn biomass, the spatial distribution of corn biomass in the study area is analyzed. The results showed that the average biomass per unit area of maize increased with the increase in altitude below 1000 m. There are relatively more human activities in low altitude areas, which are more active in forestry production. The best planting altitude of corn is 0 ∼ 1000 m. When the altitude is higher than 1000 m, the corn biomass gradually decreases. In terms of slope, if the slope is lower than 15°, the biomass of maize increases with the increase in slope. If the slope is lower than 15°, the biomass of maize decreases gradually with the increase in slope. The biomass of maize on sunny slope was higher than that on shady slope.