Network Particle Swarm Optimization Algorithm Research Articles

Enhancing the reliability of CSP solder joints under thermal cycling conditions through particle swarm optimization of an improved BP neural network

The packaging of chip-scale (CSP) devices plays a pivotal role in enhancing the reliability of CSP under thermal cycling conditions, largely due to the intricate structure and the recurrent alterations in the actual operating environment. The objective of this study is to enhance the reliability of solder joints by optimising the structural parameters of CSP packaging in order to reduce the strain values at the solder joints. In order to enhance the design efficiency and accuracy of the computational model, the Anand model is adopted in order to define the solder joint parameters, and finite element simulations are conducted using Ansys software. This paper puts forward a novel intelligent algorithm that fuses response surface methodology with a neural network-particle swarm optimization algorithm, thereby enhancing the precision of the system. The method is capable of identifying the combination with the minimum strain value using limited data, thereby addressing the issue of insufficient generalisation ability in conventional methods. The implementation of this method into the model resulted in a minimum strain value of 4.071 × 10−3, representing a 37.6 % reduction compared to the original CSP structure. Furthermore, the optimized lifespan is approximately 3.24 times longer than that observed prior to optimization. The approach presented in this paper has the potential to significantly enhance design efficiency and increase the lifespan of components. It offers a novel perspective for optimising the structural parameters of CSP packaging.

A Hybrid Artificial Neural Network—Particle Swarm Optimization Algorithm Model for the Determination of Target Displacements in Mid-Rise Regular Reinforced-Concrete Buildings

The realistic determination of damage estimation and building performance depends on target displacements in performance-based earthquake engineering. In this study, target displacements were obtained by performing pushover analysis for a sample reinforced-concrete building model, taking into account 60 different peak ground accelerations for each of the five different stories. Three different target displacements were obtained for damage estimation, such as damage limitation (DL), significant damage (SD), and near collapse (NC), obtained for each peak ground acceleration for five different numbers of stories, respectively. It aims to develop an artificial neural network (ANN)-based sustainable model to predict target displacements under different seismic risks for mid-rise regular reinforced-concrete buildings, which make up a large part of the existing building stock, using all the data obtained. For this purpose, a hybrid structure was established with the particle swarm optimization algorithm (PSO), and the network structure’s hyper parameters were optimized. Three different hybrid models were created in order to predict the target displacements most successfully. It was found that the ANN established with particles with the best position revealed by the hybrid models produced successful results in the calculation of the performance score. The created hybrid models produced 99% successful results in DL estimation, 99% in SD estimation, and 99% in NC estimation in determining target displacements in mid-rise regular reinforced-concrete buildings. The hybrid model also revealed which parameters should be used in ANN for estimating target displacements under different seismic risks.

A hybrid approach of artificial neural network-particle swarm optimization algorithm for optimal load shedding strategy

This paper proposes an under-frequency load shedding (UFLS) method by using the optimization technique of artificial neural network (ANN) combined with particle swarm optimization (PSO) algorithm to determine the minimum load shedding capacity. The suggested technique using a hybrid algorithm ANN-PSO focuses on 2 main goals: determine whether process shedding plan or not and the distribution of the minimum of shedding power on each demand load bus in order to restore system’s frequency back to acceptable values. In the hybrid algorithm ANN-PSO, the PSO algorithm takes responsible for searching the optimal weights in the neural network structure, which can help to optimize the network training in terms of training speed and accuracy. The distribution of shedding power at each node considering the primary control and secondary control of the generators’ unit and the phase electrical distance between the outage generators and load nodes. The effectiveness of the proposed method is experimented with multiple generators outage cases at various load levels in the IEEE-37 Bus scheme where load shedding cases are considered compared with other traditional technique.

Optimization of Sour Water Stripping Unit Using Artificial Neural Network–Particle Swarm Optimization Algorithm

Sour water stripping can treat the sour water produced by crude oil processing, which has the effect of environmental protection, energy saving and emission reduction. This paper aims to reduce energy consumption of the unit by strengthening process parameter optimization. Firstly, the basic model is established by utilizing Aspen Plus, and the optimal model is determined by comparative analysis of back propagation neural network (BPNN), radial basis function neural network (RBFNN) and generalized regression neural network (GRNN) models. Then, the sensitivity analysis of Sobol is used to select the operating variables that have a significant influence on the energy consumption of the sour water stripping system. Finally, the particle swarm optimization (PSO) algorithm is used to optimize the operating conditions of the sour water stripping unit. The results show that the RBFNN model is more accurate than other models. Its network structure is 5-66-1, and the expected value has an approximately linear relationship with the output value. Through sensitivity analysis, it is found that each operating parameter has an impact on the sour water stripping process, which needs to be optimized by the PSO algorithm. After 210 iterations of the PSO algorithm, the optimal system energy consumption is obtained. In addition, the cold/hot feed ratio, sideline production position, tower bottom pressure, hot feed temperature, and cold feed temperature are 0.117, 18, 436 kPa, 146 °C, and 35 °C, respectively; the system energy consumption is 5.918 MW. Compared with value of 7.128 MW before optimization, the energy consumption of the system is greatly reduced by 16.97%, which shows that the energy-saving effect is very significant.

Research on Power Consumption Model of Permanent Magnet Direct Drive Belt Conveyor System Based on Fruit Fly Optimization Algorithm-Generalized Regression Neural Network-Particle Swarm Optimization

The permanent magnet direct drive electric drum is used to replace the driving device of the traditional belt conveyor, which simplifies the structure. However, the permanent magnet direct drive electric drum still operates in a high speed and stable state after starting, and cannot transport materials reasonably, belt speed cannot match the coal quantity reasonably; the problem of energy waste is still severe. Thus, based on fruit fly optimization algorithm-generalized regression neural network-particle swarm optimization algorithm, the power consumption network model of the permanent magnet direct drive belt conveyor system is established. The relationship among belt velocity, coal transport quantity and power is obtained, and the optimal belt velocity is found with the power consumption network model. Thus, the minimum power is obtained. The algorithm selects the optimal smoothing factor by fruit fly optimization algorithm, inputs the optimal smoothing factor into generalized regression neural network and establishes the optimal power consumption model. Then establishes the matching relationship of coal quantity, optimal power, and optimal belt speed in the power consumption model by the adaptive weight particle swarm optimization. The model is compared with the network model with smoothing factors of 0.7 and 0.4. The comparisons show that the optimized model performs better, which can be better applied to establish the energy consumption model of the system.

STUDY OF THE FLOW AND HEAT TRANSFER OF A VISCOELASTIC FLUID USING HYBRID NEURAL NETWORK-PARTICLE SWARM OPTIMIZATION (HNNPSO)

Fluid flow and heat transfer of a second-order viscoelastic fluid in an axisymmetric channel with a porous wall for turbine cooling applications are studied. The nonlinear differential equations of the fluid flow and heat transfer arising from similarity solutions are computed employing a Hybrid Neural Network-Particle Swarm Optimization algorithm (HNNPSO). A trial function, satisfying the boundary conditions, as a possible solution for the governing equations is introduced. The trial functions incorporate a multi-layer perceptron neural network with adjustable parameters (the weights and biases). The Particle Swarm Optimization algorithm (PSO) is applied to find the adjustable parameters of the trial solution to satisfy the governing equations. Finally, comparisons are made between the results of the present method (HNNPSO) and the results of the fourth order Runge–Kutta method, finite difference method, and Variational Iteration Method. The results indicate that HNNPSO method conveniently produces a polynomial analytic solution with remarkable accuracy, and the accuracy of the solution improves as the number of neurons of the neural network increases.

A predictive model on size of silver nanoparticles prepared by green synthesis method using hybrid artificial neural network-particle swarm optimization algorithm

The aim of this study is green synthesis of Ag nanoparticles (Ag NPs) and optimization of practical process by artificial neural network to prepare the minimum size of Ag. To decrease the issues of trapping in local minima, heuristic optimization methods, i.e., the particles swarm optimization (PSO) and modified PSO, were applied as training strategy for feed forward Neural Networks. To collect the experimental data, design of experiment with factorial D-optimal array was employed. Uv–visible, X-ray diffraction and field emission scanning electron microscopy were used to characterize the prepared samples. The results confirmed the possibility of Ag NPs preparation with the size about 4 nm at optimum condition. Based on the results of the optimal network sensetivity analysis, it was found that the feed rate, AgNO3 to opium ratio and agitation speed have the greatest impact on the particle size of the final product, respectively.

Hybrid neural network–particle swarm optimization algorithm and neural network–genetic algorithm for the optimization of quality characteristics during CO2 laser cutting of aluminium alloy

Aerospace and automobile industries employ aluminium alloys due to its lightweight and excellent resistance to corrosion. As aluminium is highly reflective and highly thermally conductive, it is difficult-to-cut material by laser processing. The quality characteristics of the cut predominantly depend upon the combination of laser processing parameters. The main quality indices for evaluating CO2 laser cutting were surface roughness, kerf width and kerf taper; and the machining parameters considered were laser power, speed and gas pressure. This work suggests hybrid artificial neural network (ANN)–particle swarm optimization (PSO) algorithm and artificial neural network (ANN)–genetic algorithm (GA) to optimize the associated multi-response characteristics during CO2 laser cutting of aluminium 6061 alloys. The results illustrate that the hybrid ANN–GA and ANN–PSO model is an efficient tool for the optimization of process parameters in CO2 laser cutting of difficult-to-cut material—aluminium. From the optimization results, it can be concluded that the proposed ANN–GA approach can be efficiently utilized to optimize the parameters for obtaining minimum roughness, kerf width and kerf taper.

Die design and process optimization of plastic gear extrusion

The flow velocity of the melt in the extruder was simulated by using software Polyflow, and the size of the die channel with the best flow uniformity was obtained. The die profile shape is obtained by reverse design. The length of the shaping section is determined by Ansys transient thermal analysis. According to the simulation results, the design and manufacture of extrusion die of plastic gear and vacuum cooling setting were obtained. The influence of the five process parameters on the precision of the plastic gear were studied by the single factor analysis method, such as the die temperature T, the screw speed R, the die spacing S, the vacuum degree M and the hauling speed V. The optimal combination of process parameters was obtained by using the neural network particle swarm optimization algorithm(T = 197.05 °C, R = 9.04rpm, S = 67mm, M = −0.0194MPa). The tooth profile deviation of the extruded plastic gear can reach 9 level of accuracy.

Application of Artificial Neural Network–Particle Swarm Optimization Algorithm for Prediction of Gas Condensate Dew Point Pressure and Comparison With Gaussian Processes Regression–Particle Swarm Optimization Algorithm

For gas condensate reservoirs, as the reservoir pressure drops below the dew point pressure (DPP), a large amount of valuable condensate drops out and remains in the reservoir. Thus, prediction of accurate values for DPP is important and leads to successful development of gas condensate reservoirs. There are some experimental methods such as constant composition expansion (CCE) and constant volume depletion (CVD) for DPP measurement but difficulties in experimental measurement especially for lean retrograde gas condensate causes to develop of different empirical correlations and equations of state for DPP calculation. Equations of state and empirical correlations are developed for special and limited data sets and for unseen data sets they are not generalizable. To mitigate this problem, in this paper we developed new artificial neural network optimized by particle swarm optimization (ANN-PSO) for DPP prediction. Reservoir fluid composition, temperature and characteristics of the C7+ considered as input parameters to neural network and DPP as target parameter. Comparing results of the developed model in this research with Gaussian processes regression by particle swarm optimization (GPR-PSO), previous models and correlations shows that the predictive model is accurate and is generalizable to new unseen data sets.

Application of Artificial Neural Network-Particle Swarm Optimization Algorithm for Prediction of Asphaltene Precipitation During Gas Injection Process and Comparison With Gaussian Process Algorithm

Asphaltene precipitation is a major problem in the oil production and transportation of oil. Changes in pressure, temperature, and composition of oil can lead to asphaltene precipitation. In the case of gas injection into oil reservoirs, the injected gas causes a change in oil composition and may lead to asphaltene precipitation. Accurate determination and prediction of the precipitated amount are vital, for this purpose there are several approaches such as experimental method, scaling equation, thermodynamics models, and neural network as the most recent ones. In this paper, we propose a new artificial neural network (ANN) optimized by particle swarm optimization (PSO) to predict the amount of asphaltene precipitation. This is conducted during the process of gas injection into oil reservoirs for enhanced oil recovery purposes. In the developed models, (1) oil composition, (2) temperature, (3) pressure, (4) oil specific gravity, (5) solvent mole percent, (6) solvent molecular weight, and (7) asphaltene content are considered as input parameters to the neural network. The weight of asphaltene and asphaltene content are considered as input parameters to the neural network and the weight of asphaltene precipitation as an output parameter. A comparison between the results of the proposed new model with Gaussian Process algorithm and previous research shows that the predictive model is more accurate.

Research on Particle Swarm Optimsiation and its Application in Neural Network

This paper based on the PSO algorithm is a neural network model, and with other learning algorithm, and the results show that the performance comparisons are based on improved PSO algorithm two perceptron networks have higher classification accuracy and strong generalization ability. Particle Swarm Optimization (PSO) as an emerging evolutionary algorithm fast convergence speed, robustness, global search ability, and does not need the help of the characteristics of the problem itself (such as gradient). Combination of PSO and neural network PSO algorithm to optimize the connection weights of the neural network can be used to overcome the problem of BP neural network can not only play the generalization ability of the neural network, but also can improve the convergence rate of the neural network and learning capacity.